Talk:Comet (programming)/Archive 2

This page is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Disamb

Hihi

I saw you reverted my disamb edit...I'm trying to learn how to do this right...could you tell me why its correct to link to the disamb page? I've been working from this guideline: "Wikipedia articles should not link to disambiguation pages (with rare exceptions where the ambiguity of a term is being discussed); instead, links should go directly to the appropriate article."

Just trying to figure out whats up...thank you!! Legotech (talk) 06:22, 10 January 2008 (UTC)

I reverted it because “computer virus” is not the proper meaning of the term payload as used in this article. It looks like a new stub article has been created for Payload (communication and information technology), which would I suppose be appropriate. But a) the parenthetical in that title is absurd, b) the stub of an article there doesn't say anything useful, and c) the definition as given at the top of the payload page is quite sufficient in my opinion. I wouldn't mind terribly if the link was pointed at the Payload (communication and information technology) page, but that change wouldn't make much substantive difference I don't think. The purpose of the wikipedia guideline is to prevent links from going to disambiguation pages with nothing to do with the intention of the link. Because in this case the intent of the link is merely to provide a definition, as the concept of a payload doesn't frankly have much depth, I think the current link is fine. Does that make sense? --jacobolus (t) 12:35, 10 January 2008 (UTC)

Thanks! I'm pretty new to all this and appreciate the help! Legotech (talk) 16:10, 10 January 2008 (UTC)

I was looking at things again and it looks like someone tried to start Payload (Software) but never got anywhere...I think that would work better than that mouthful of Payload (communication and information technology). What would you think about me moving the info from C&IT to Software and then redirecting C&IT to Software and then using that in this case?

I'm not trying to be a jerk, just trying to figure things out so that they make sense.

Thanks again. Legotech (talk) 23:37, 10 January 2008 (UTC)

I certainly don't think you're trying to be a jerk. :) As for payload (software), that's not quite right either. In that case the implication is that the payload is an executable program, carried by some other container, such as a virus in an email message. In this case, the “payload” is just an arbitrary message. Rewriting the page about payload (communication and information technology) to be clear and understandable, and optionally moving it to a better title, is not worth it to me personally. The benefit to this article of having a link go to such a page instead of the payload page is negligible. If you want to do that though, go for it! --jacobolus (t) 00:16, 11 January 2008 (UTC)

Given that the (recently moved to) payload (software) article is confusing and uninformative (also, “software” is not a good description for the field), I disagree with pointing the “payload” link in this article at it, until it has been rewritten. I might point “payload” back at payload in a few days, if it stays that way. --jacobolus (t) 20:02, 12 January 2008 (UTC)

No worries, can you take a look now? Think it should be Payload (network) or something else? I can't come up with any way to describe it. Thanks again for your help! —Preceding unsigned comment added by Legotech (talk • contribs) 00:33, 13 January 2008 (UTC)

Bayeux

Any objections to adding more detail about the Bayeux protocol? —Preceding unsigned comment added by RealWorldExperience (talk • contribs) 21:33, 2 May 2008 (UTC)

What kind of more detail? It's quite possible I'll object, if the addition seems excessive for this article. What do you think readers of this page would need to know about Bayeux that they can't currently figure out from a combination of this article and the linked Bayeux spec? My suggestion would be instead to add an article about Bayeux to Wikipedia, and put detailed information about it there. But this article could potentially have a bit of detail added. Depends what you have in mind. —jacobolus (t) 02:10, 3 May 2008 (UTC)

I would be interested to hear more, is this a related technology? - 83.254.208.192 (talk) 21:28, 12 June 2008 (UTC)

Comet means pushing, polling, streaming?

What exactly is the relationship between Comet and push technology, could someone familiar with the topic explain the technical differences? Is comet a server side technology, a client side technology (specific form of HTTP requests) or both? Here a quote from the push technology article: "The Comet technique tries to emulate server-push with a lot of overhead in JavaScript programming". Thanks! - 83.254.208.192 (talk) 01:23, 1 June 2008 (UTC)

As far as I can tell, the push technology article tries (rather unsuccessfully) to describe the idea of pushing data in general, which could be Comet, but could also be push email, etc. That article should probably be completely re-written, as right now it is poorly organized, uninformative, and somewhat inaccurate. Comet, as this article explains, is push-like interaction which sends events to JavaScript callbacks in a web page. Comet requires the cooperation of client-side (in-browser) code and server-side code. --jacobolus (t) 10:19, 1 June 2008 (UTC)

I am still having problems to understand where Comet separates itself from other technologies. Sound like Comet is a subcategory of HTTP push using JavaScript (or other browser applications) to handle asynchronous notifications. Isn't this in most cases called AJAX or is the difference between Comet and AJAX that with Comet that you are using server push instead of client pull? Since the article mentions also XMLHttpRequest polling I am confused. It isn't clear for me where those technologies (Comet, AJAX, HTTP server push, HTTP client pull) overlap or which is a subset of the other. Perhaps a distinction between all those technologies could help to explain what Comet is.

Actually I was trying to find out more about Comet... I think a good description is found at http://cometdaily.com/about/. From what I read there, Comet is an AJAX related technology and a collection of non-static web technologies. For an outsider the first impression is that Comet is complex, unshaped (in terms of not focusing on a few core technologies) and has many requirements on server, client and the intermediate transport infrastructure (e.g. HTTP proxies not suited for streaming will harm or block long living server connections). Without wanting to judge, the technology seams relatively unknown (are there any notable Comet applications, where is it in productive use?) and the current Wikipedia article is very hard to understand. My suggestion would be to simplify the article and shorten it. - 83.254.208.192 (talk) 13:19, 1 June 2008 (UTC)

“Ajax” means the browser requests new information without reloading a page. “Comet” is a particular type of Ajax interaction, in which the browser requests new information, opens a long-lasting connection, and then receives updates in real time as the server has them ready. “HTTP Push” is often used to refer to a specific Netscape technology of the mid-90’s which is nearly unused. The description at this article is probably confusing if you just arrived after one of Damiens’s edits, which have chopped out all explanation and context. As for notable Comet applications, Google’s Gchat, Meebo, Renkoo, Facebook chat, etc., are all rather notable. Unfortunately, Damiens keeps removing the history section of the article —jacobolus (t) 23:57, 1 June 2008 (UTC)

Already in the late 90s there has been code that uses a web server to push notifications to clients... calling it "HTTP server push" not Comet. To give an example, a simple nph-CGI script on Apache [2] could forward notifications, no special web server needed and it works with many web browsers, no Netscape specific MIME type is used here. - 83.254.208.192 (talk) 10:08, 2 June 2008 (UTC)

The Gchat, Meebo and Facebook articles do not mention "Comet" in the article text. I think a well defined terminology is needed that describes what the Comet technology is and what sets it apart from AJAX (or alternatively both could be merged). - 83.254.208.192 (talk) 01:21, 2 June 2008 (UTC)

I've added mentions in those three articles. Google also added Comet functionality to Google Gears recently. Search down for “Comet” on this page. What Comet is is sending events to the browser as they come to the server, rather than as the browser requests them. Comet is an extension of Ajax, a way of using Ajax to provide a different interaction model, not really something completely separate. But I think the topic is complex enough that merging the articles is a bad idea, and would confuse readers. The “web application architectures” section tries to explain how Comet works. Is it not clear enough? —jacobolus (t) 01:49, 2 June 2008 (UTC)

About the “web application architectures” section, the current version is confusing in my opinion and doesn't help to understand the topic. I suggest to remove it completly and give a shorter description with images to visualise Comet compenents and the data exchange between them. - 83.254.208.192 (talk) 10:46, 2 June 2008 (UTC)

What specifically did you find confusing about it? I agree with you that images would help it substantially. —jacobolus (t) 20:27, 2 June 2008 (UTC)

It talkes a lot about everything around and little about Comet in my opinion. I don't know who is the intended audience for this section, but I would state it can not even explain what Comet is to readers familiar with main stream web development technologies. As long as people don't understand what Comet is, even when they work on related fields, this will have serious problems of being accepted as a notable technology or to find contributors. - 83.254.208.192 (talk) 02:10, 3 June 2008 (UTC)

Here we go, a short comparison between Ajax and Alex Russel's Comet (including pictures): "Comet: A new approach to Ajax applicatons" [3]. In a nutshell, it's a flavour of Ajax over a streaming (or long lived) HTTP connection. - 83.254.208.192 (talk) 23:38, 4 June 2008 (UTC)

Looks great; if you want to work info from that into the article it'd probably benefit. Restepc (talk) 00:05, 5 June 2008 (UTC)

Currently it seams that we have establised that Comet is not a new technology but a term that describes a group of technologies. I added this yesterday to the article text, correct me if I am wrong. My suggestion would be to write a short section "Comet is a combination of" and list the involved technologies including links to the corresponding articles and a few words (maximum one line per list entry). For an outsider the first impression is that Comet is complex and unshaped, that eveything can be labeled Comet that isn't traditional browser polling (request - response, pause, request - response). We need to get away from this uncertainty and contradictions (e.g. Comet requires Ajax, anything JavaScript, any browser side extension), in the worst case find out that it is just a fancy word for saying "big hack towards interactive web".[4] - 83.254.208.192 (talk) 07:32, 14 June 2008 (UTC)

Well, I tried once more to describe what "Comet" is from a technical point of view. The following is the current version of the article introduction (I personally think it is still too complicated for readers unfamiliar with the topic and would also mention "buzzword" in the introduction, let's see how much survives the lobbyist patrol):

In web development, Comet is a neologism to describe an web application model in which a long-held HTTP connection allows a web server to push data to a browser, without the browser explicitly requesting it. Comet is an umbrella term that uses multiple techniques for achieving this interaction. All methods have in common that they rely on browser-native technologies such as JavaScript, rather than on proprietary plugins. In theory, the Comet approach differs from the original model of the web, in which a browser requests a complete web page or chunks of data to update a web page. However in practice, Comet applications use traditional Ajax with long polling to detect new information on the server. The concept has been known by various names, including Ajax Push, Reverse Ajax, Two-way-web, HTTP Streaming and HTTP server push among others. - 83.254.214.192 (talk) 20:35, 4 July 2008 (UTC)

That description is at best misleading and imprecise, and it completely misses the point of the contrast with the “traditional model of the web”, which is page-by-page, and does not use polling, long-polling, or any similar mechanisms. Your distinction between “theory” and “practice” is a false dichotomy. —jacobolus (t) 22:10, 4 July 2008 (UTC)

So you say it is wrong, OK, I tried. I really wonder how long we need to be able to understand what Comet is, if this categorisation is so hard I wonder if it's me or Comet that is the source of the problem. I also work with web technologies and frankly Jacob, I fail to see what the point of the Comet terminology is. One of the problems with Comet is that it is impossible to make a clear distinction between "traditional model of the web" and "Comet model of the web", for example a hidden frame, not a iframe, for sending/streaming script-HTML-tags has already been used in the late 1990s, already before DOM, DHTML and Ajax existed (ugly hacks in a ever-changing web development landscape). So things you used to label "comet transports" existed long before the Comet term was created by Alex Russel and more web technologies will be invented (HTML5 draft does not mention "Comet" and calls its model "server sent events", AIMs WebAPI calls its push solution "Ajax long polling"). After weeks there is still no consensus, perhaps the whole 'Comet technology cloud' is a failed Web2.0 concept that historically grew... and will sooner or later make place to an alternative terminology that is more concrete. - 83.254.214.192 (talk) 22:58, 4 July 2008 (UTC)

"Implementations" section

I believe the "Implementations" section should be reduced. We don't need to link to every Comet-like piece of code ever written (see "pi.comet", for instance).

I should try to establish a neutral criteria for inclusion so that to avoid link spamming.

What could be a good criteria for inclusion on this list? --Damiens.rf 17:46, 1 June 2008 (UTC)

I agree, but have no ideas on criteria, the most basic idea would be to only include those things important enough to have their own wikipedia article. Restepc (talk) 20:34, 1 June 2008 (UTC)

I agree with you whole-heartedly. I have no idea what useful criteria would be, however. None of these, or nearly none of them, deserve their own wikipedia articles, so by that standard the section should be removed altogether. On the other hand, I think mentioning implementations is useful for readers. —jacobolus (t) 23:58, 1 June 2008 (UTC)

Use of large marginal quotations

Apparently Damiens and Restepc feel that the use of quotations at the right side of the article makes it feel like an advertisement. Let me attempt to justify them.

This article currently has no images (though it could use some), and so it is essentially a gigantic wall of text. To break up that flow, provide readers with something to look at, provide some more subjective commentary about Comet while properly sourcing it to specific individuals to avoid NPOV, and give readers some idea of what Comet developers themselves think about the technology, and also to link to a number of useful external sources without further expanding the gigantic external links section, I added a series of quotations to the right side of the article. These fall outside the main flow of the article, and so readers uninterested in them can happily skip over them, but are in my opinion relevant and informative.

I see no particular reason to remove these, and neither Damiens nor Restepc has provided an explanation for why they should be removed, or explained how any of them is particularly unencyclopedic.

I readily admit that they lend the article a format slightly unlike many other Wikipedia articles. But as far as I can tell, it does not fall outside the range of Wikipedia policy.

—jacobolus (t) 01:15, 2 June 2008 (UTC)

I think there would be much less of a problem if the quote had been added by someone who didn't know the people quoted....you actually quote yourself at one point I think Restepc (talk) 01:25, 2 June 2008 (UTC)

I don’t really “know” most (maybe any?) of the people quoted, other than myself of course. I have since writing this article exchanged emails with a couple of them. Is there any more specific problem with them than the appearance of a conflict of interest? —jacobolus (t) 01:29, 2 June 2008 (UTC)

The way you used the quotations makes it sounds like that phrases (or their importance) are being endorsed by Wikipedia. It works ok for magazine articles, but not for encyclopedic articles. Statements like "15 years of plain HTTP request/response has taught us to think in terms of static pages..." or "Comet improves application responsiveness for collaborative, multi-user applications..." are just somebody's opinions, and should not be given that prominence in the article.

And even if we are to add any (inline) quoted opinion, we should make sure we have a neutral criteria for choosing who said something relevant. The fact that you mostly quoted people from your website (commetdaily.com) and yourself shows your conflict of interests influenced your criteria. --Damiens.rf 04:03, 3 June 2008 (UTC)

History section

Damiens wants to completely remove the history section of this article. I believe such a change to be unproductive, because I feel that nearly all Wikipedia articles need history sections. Damiens, please explain what you think the ideal history section would look like for an article about a subject like this one, so we can work to form a consensus on how this article’s history section should be structured. —jacobolus (t) 01:15, 2 June 2008 (UTC)

the article almost certainly would benefit from a history section, it just needs to be a lot shorter than it is now. Restepc (talk) 01:26, 2 June 2008 (UTC)

What should be in it? What is wrong with its current length? —jacobolus (t) 01:28, 2 June 2008 (UTC)

Problems with the History section

Jacob, I agree that an article about a technology (assuming this is one (see bellow)) bennefits from an "History" section. But it doesn't implies that it have to be the one that you wrote.

A history section (of such an article) should contain some background on what led to the technology creation and then go on to explain the relevant facts since it was established. For something that is claimed to be just 2 years old, and didn't really caught-up, it's at least suspect to have such a long "history" section.

When I started trying to clean up the article, I noticed that the this section was the most broken-beyond-repair of them all, and I do believe that at this point, the article would be better without that section.

What's follow is, as you asked, a detailed account of the problems I see in this section.

"Server Push"

The (huge) History section starts by describing some technology used by Netscape, "Server push", and eventually copied by IE... and then concludes that "Netscape’s server push cannot really be defined as Comet". Just in this first paragraph I can see a dozen of (overlapping!) problems.

First, it's not Wikipedia's job to declare what can't and what cannot be described as Comet (and I explained that on my edit summary). Second, if that is not Comet, why is it in the History section? (If you really want to mention it as an relevant ancestor, for god's sake be BRIEF). Third, the whole section seem to be taken from just one self-published source (guess what? cometdaily.com). And fourth, and now for a problem that affects the article as a whole, the website cometdaily.com is terribly biased (and here I believe I'm disagreeing with Restepc as well). For instance, the article used as a "reference" for this sub-section starts with "...The increasing success of the Comet paradigm..."[5].

"Java Applets"

The second History-sub-section on Jacob's version of the article, entitled "Java applets", just as the previous "Server push", is also about something that is not Comet. At the end of this sub-section we're at a 15 lines-long history section that haven't talked about Comet yet. Jacob Rus, at this point, I respectfully believe that you should consider leaving the writing of the history section to some less-prolix contributor.

As a starting point for reducing this pre-History section, I would suggest removing the povs like "the ability to embed Java applets into browsers opened richer real-time options", the original research like "..User interfaces in Java applets were often buggy and frustrating for users, suffering graphical glitches..." and waseal words like "web application developers preferred to use browser-native technologies...". Since this all is not even talking about Comet, removing it shouldn't be that hard. Of course, Jacobs, remember to completely avoid to use links from the website you work for (cometdaily.com) as if they we're good sources. They are self-published and have a strong CometIsGreatSoAreWe-directed pov.

As for the last sentence of this sub-section, it's a good example of Synthesis of published material which advances a position: "To mitigate these disadvantages, some developers—for instance, Caplin Systems in 1998—used Java applets merely as a transport mechanism (...)[23]" (or worse, since the reference [23] is not really "published", but self-published).

"Early Comet applications"

The third section (that is still not about Comet, but just part of the "background that led to the creation of Comet") starts with plain unverifiable original research: "...Circa 2000, developers began creating the first pure-JavaScript Comet applications, using the Frame/IFrame transport". Sure. Pure and simple original research. And the original writing continues, as the article author declares that Pushlets "was the first open source Comet implementation" (make not mistake, the "reference" provided[6] doesn't support this claim at all).

(More on "Caplin" software later)

The remaining of this sub-section simply decides what applications were the next relevant "early Comet applications", and goes on to state some unverified facts about them. No reference for "The 2000 startup KnowNow built several Comet demonstrations..". No references for Netscape's Marc Andreessen's Bang-Networks bankrupting while trying to standardize web-push. No reference for "Comet found a niche in financial trading enterprise markets...".

In this discussion at Jacob Rus's blog about the writing of this wikipedia article, it becomes clear what kind of "research" that was done for this section. When his cometdaily.com fellow Alessandro Alinone disagreed about (the software called) Caplin being a pioneer in "...using Comet instead of Java applets...", Jacob Rus explains that he took this information from the blog of the author of the Caplin software! Not exactly a reliable, independent source by Wikipedia's standards. (and as if it wasn't enough, this Caplin software author is also cometdaily.com fellow).

Jacob, please take this criticism prudently. Everybody commits mistakes.

"Popularization"

The title itself is povvy. As far as I know, "Comet" is not popular. But in any case, it's just an opinion. By the way, a fix for this title was one of the my editions that were bulk-reverted.

The pov goes on with "...several prominent applications...". Some claims about googletalk, Meboo, IE go again referenced by the self-published folks at cometdaily.com. Something is said about a non-notable "'event-planning site" called Renkoo, but the reference doesn't mentions comet at all (Am I missing something?)

And at this point we have the pearl "...Meanwhile, Comet continued to make inroads into enterprise markets...". Jacob, I hope you understand how this is an unappropriated tone. This phrase has a footnote (something that also makes this article strange) saying something about "enterprise Comet companies and solutions" being created, and puts the well-established ICEfaces framework (that's not "Comet", but Ajax) side by side with Caplin (that software created by cometdaily's Martin Tyle). Not to be surprised at this point, the "reference" used is Martin Tyle's own blog at cometdaily.

The text them praises one of cometdaily's contributors for coining the term "Comet" and for "bringing attention to the approach" (POV!). Next, no source is provided for the claim that "The term quickly gained currency, and Comet became a prominent lecture topic at web-related technology conferences".

And the sub-section ends with this incredible piece: "In late 2007, the website Comet Daily assembled a group of Comet server implementors to write articles about Comet techniques and usage". That is it. The history of Comet, according to Jacob Rus, culminates with the creation of the cometdaily's website. A link to a cometdaily's press-release is used as a (verifiable, independent) source.

"Standardization attempts"

This sections seems more appropriate as a technical discussion than as a part of "history".

Jacob Rus cites himself on his self-published blog as a reliable reference and some point. In other point the article wrongly claims that "WHATWG's HTML 5 specification attempts to standardize the Comet transport", when it makes no use of the "Comet" buzzword at all.

Some good crystal balling is used in "If HTML 5 is implemented in many browsers, Comet developers will be able to avoid the current need to implement several transports."

"Future"

It's pov and non-sourced to claim that "Comet implementations and applications continued to proliferate". Passages with the tone of "...look no further than the Implementations section of External links, below..." need to be ripped from this article. And of course, using a collection of links to comet implementations as a reference of its "proliferation" is one of the best example of using primary sources I could find on Wikipedia.

The last sentence is complete pov original research: "As broadband-speed connections become ubiquitous, as the web becomes more social, and as users come to expect instant feedback, Comet stands poised for increasing adoption". Attributing it to some cometdaily contributor, or to anyone else as a matter of fact, doesn't make it less povvy. It's just an opinion, and as such, should not be stated as a fact.

...

I think this covers the whole History section. I hope you all see how broken it is, and how desperate I became when first tried to fix it (that was why I removed it completely).

Jacob, I hope you take this criticism respectfully. I understand how you may feel attached to the article you wrote, but please don't let your feelings dominate your judgment this way.

Yours truly,

--Damiens.rf 06:41, 3 June 2008 (UTC)

Self-contained article

The section of this article about web application architectures is, in my opinion, absolutely essential to explaining to lay readers what Comet is. Without a clear understanding of what the “traditional” and “Ajax” models of the web are, it is impossible for such readers to understand what Comet is, because it is defined in relation to those. I have tried to explain Comet to dozens of random people, and without a clear explanation of how the web otherwise works, they are invariably left confused. Damiens seems to feel that this section holds Comet up as a newer and better model than these others (and he therefore removed the section altogether). I have tried to keep the article free of such value judgments, but I can see how he would draw such an inference. So the question is, how can we provide necessary context to novice readers, without implying that Comet is superior? —jacobolus (t) 01:15, 2 June 2008 (UTC)

Answer: Using wikilinks. The article on Friedrich Nietzsche doesn't re-explains what Existentialism is, and the article on Existentialism doesn't re-explains what Ontology is... the Ontology article doens't re-explains what's Metaphysics neither this last attempts to re-explain what's Philosophy.

And this Comet thing is waaaay simpler than Nietzsche. --Damiens.rf 07:33, 3 June 2008 (UTC)

Yes, exactly. It is impossible to have a wiki article which explains Nietzsche’s thought to someone who has no background in philosophy (or perhaps impossible to explain altogether), and reading Wikipedia to learn philosophy is utterly hopeless, because it has no coherent structure or basic explanations. It seems to me that the attempt at an explanation of Comet which gives the newcomer a complete picture, if it can be accomplished in a few paragraphs, is quite reasonable. —jacobolus (t) 10:12, 3 June 2008 (UTC)

By "a few paragraphs" you mean something like your version? I would strongly disagree with this categorization if this is the case. --Damiens.rf 15:13, 3 June 2008 (UTC)

Yes, the section on web application architectures has about 6–7 paragraphs of background, which provide a clear context in which to explain how Comet works, so that readers are not left scratching their heads. —jacobolus (t) 00:00, 4 June 2008 (UTC)

Scalability

The chief problem facing any popular Comet application is how to scale it, and much of the reason that Comet has not been adopted by as many developers as would like to use it, is that scaling such applications is difficult. I feel that addressing this topic is worthwhile in an article about Comet, because it is one of the top questions facing Comet developers, and much has been written about the topic. Damiens removed the section altogether, with no justification. Damiens, can you explain why such a section is not relevant to this article? —jacobolus (t) 01:15, 2 June 2008 (UTC)

I think the problem with a lot of these is a combination of sources and length, hopefully Damiens can tag 'citation needed' where appropriate rather than deleting the whole section, but be aware that other people may feel it necessary to trim it down. Restepc (talk) 01:30, 2 June 2008 (UTC)

First of all, Jacob, I hope you understand that what you state above is just your opinion about "comet", and it should only consider "addressing this topic" if it can be somewhat attributed to independent reliable sources.

What I mean is, saying that scalability is the "chief problem facing any popular comet application" is an opinion. Just as I could disagree that "Comet has not been adopted by as many developers as would like to use it", preferring to say that it simply didn't catched. But neither my opinion nor yours should find they way into the article just because we can click the "edit" link.

As with the "History" section discussed above, this Scalability section is a soup of original research, self-published sources, and pov.

The first paragraph (including the 2 bullets following it) are, in the lack of sources, pure original research. The sentences saying that comet applications "...typically use more resources than other types of web applications..." or are "often interactive, allowing arbitrary groups of users to communicate with each-other" are generalizations based on the author's own experience (or lack thereof). Saying that traditional web-servers "...cannot cope with such large numbers of open connections..." also needs references.

Also like the History section, the Scalability section suffers from overly-prolix writing. Do we need to explain how threads inside apache synchronously handle http connections? Can't we simply say that the comet application requires a dedicated communication channel to the server while others don't?

Not to mention the complete lack of sources in this apache detailed discussion...

The sub-section on horizontal-scalability starts with some good original research about how web-applications are difficult to scale, and about the behavior of "powerful servers" in heavy conditions and what Jacob believes we should do on these cases: "...distribute across many servers". I would ask for references, but I believe that the greater problem is the text's verbosity. If concisely written, the text would be void of these unverifiable or hard-to-verify statements.

To simplify the arguments, the whole "Scalability" section is a combination of non-sourced material with statements attributed to a link to a slides-presentation. PowerPoint slides are not the best reliable published sources I know about. And it's also important to point here that Jacob Rus was also involved in the making of this presentation (he's credited at least with the "diagrams" (and thus, there are some pics of him among the sliders)).

I sincerely don't see what can be saved from this section in the article. --Damiens.rf 08:23, 3 June 2008 (UTC)

Most of these are factual, easily verifiable statements, but you are correct that it is difficult to find sources for them. I don't have the tens of hours it would take to track down sources to your standards, and won’t for the next month at least, but I will try at some point. —jacobolus (t) 10:28, 3 June 2008 (UTC)

As a very early developer of services based on what much later came to be known as Comet I tend to agree with many of Jacob's points in this talk page and find most of Damiens arguments to be very much biased against Cometdaily (which I'm not associated with and only came to notice after reading this talk page) and the person of Jacob himself. I believe it to be not unusual for the wikipedia to publish articles that are just written by people who happen to know about a certain subject even if every thing they say can not be backed by hard fact through reliable sources. You criticize this Jacob for speaking from his "lack of experience", but I'd rather say that to me it's your comments here that denote a lack of experience in this matter. Scalability and open connections is a known issue with this sort of technology for anyone who has had some experience with it. I believe that if you want this article to get better you have to let the information in and not rule it out based on your own opinion, that's how all you very verbose explanations in this talk section sound to me.

To me, naming the problems with Apache to serve as an event-driven asynchronous web server is important and it would also be important to name the alternatives that overcome some of Apache limitations in this matter, like the Medusa based Zope server or Twisted or the later developed Apache event MPM.

I believe that the word Comet originated in order to try to name something that many developers had already been doing for many years -oftentimes in a very hacky fashion- and thus it's very important to "tell the story". I'd prefer the scientist pose to be dropped altogether, in the end are we not talking of household cleaners here? --nicoechaniz —Preceding undated comment was added at 14:06, 30 July 2008 (UTC)

(Good) Wikipedia articles don't expect the reader to trust the author's expertise. They are backed up by reliable sources. This is my only bias. --Damiens.rf 14:37, 30 July 2008 (UTC)

Browser hacks

The other main problem in Comet development is how to make it work across browsers. I added a large section to this article about how Comet can be made to work in different browsers, because it is a topic of interest to anyone trying to understand the technology and its limitations, or trying to implement their own Comet applications. Understanding how existing Comet transports work is necessary background to understanding the standardization attempts such as that undertaken by the WHATWG. Damiens removed this section as well, without providing much justification. Damiens, please clearly explain why you feel this section is not relevant or not encyclopedic, and what you think it should be replaced by, if anything. —jacobolus (t) 01:19, 2 June 2008 (UTC)

The section is basically your original research account of what web-developers liked to use to implement server-pushing. It's ok that you post your findings in your blog (as you did), but the criteria for publishing such information here on Wikipedia is a bit stronger. And of course, using your blog post as a reference (as you did) doesn't changes a thing. --Damiens.rf 08:31, 3 June 2008 (UTC)

Use of Notes

Damiens, you removed all the footnotes from this article. Can you please explain why? —jacobolus (t) 01:15, 2 June 2008 (UTC)

Footnotes in my experience aren't overly common in wikipedia articles, but working them into the article would be better than simple deletion Restepc (talk) 01:31, 2 June 2008 (UTC)

Lack of commonality is no reason to remove them: most articles would probably benefit from more footnotes, which provide clarifications without cluttering article text. They should be dealt with one by one, if they are truly problematic. —jacobolus (t) 01:53, 2 June 2008 (UTC)

Don't forget they were all from a single source (Comet daily blog entries) and you even quoted yourself. As a whole it looked like marketing and personal opinions from a small circle of people. I am sure you can understand the problems that arise from this situation (neutrality, conflict of interest, original research), so the consensus at the time was to remove it. Please don't take it personal. - 83.254.208.192 (talk) 07:49, 14 June 2008 (UTC)

We're talking about the lettered footnotes, not the numbered citations. The two are unrelated. —jacobolus (t) 08:04, 14 June 2008 (UTC)

Sorry, my mistake. I can't remember seeing footnotes in Wikipedia articles. In that case I agree with Restepc, instead of footnotes text should be integrated into the normal article text flow. - 83.254.208.192 (talk) 09:38, 14 June 2008 (UTC)

big problems with the article IMO

severe lack of sources, which leads me to believe that much of it is original research.
length...way too long.
the quotes, they're not from notable sources, appear as if there for vanity, I would suggest (indeed I will insist on) scrapping all of them unless any can be shown to be important.
many minor things which add up to sounding adverty/NPOV problems, for example replacing "Comet is one of the ways to address this limitation" with "Comet removes this limitation"

Restepc (talk) 01:19, 2 June 2008 (UTC)

Which part is unsourced? Be specific. Some paragraphs have a single source, and therefore use only one reference link. Cluttering them with a note after every sentence is stupid (and unfortunately rather widespread at Wikipedia). But I am happy to attempt to track down sources for anything which you think is inadequately sourced.
Way too long by what standard? Many Wikipedia articles are much longer (indeed nearly every featured article is). If it is too long, some sections could presumably be made into their own articles, using summary style.
What is a “notable” source? I believe these provide useful context, improve the readability of the article, are mostly factual statements, and provide useful external links to further relevant explanations.
Please enumerate these, and suggest remedies.

Please separate your concerns into separate discussion sections so that the discussion doesn't get too cluttered. I started some of those sections above. Thanks. —jacobolus (t) 01:26, 2 June 2008 (UTC)

My biggest problem with the comet technology lies in the problem of understanding what it is -and- if it exists as an independent, notable technology. As I have suggested above, a good definition could help a lot here. - 83.254.208.192 (talk) 01:32, 2 June 2008 (UTC)

The concept is clearly notable, and in the last couple of years, the term “Comet” has come to dominate (over the alternatives, which are mentioned in the introduction and redirect here; for evidence go look at the talk titles for any web technology conference), and is used by Google, IBM, and others. As for the “independence” of the technology, it requires completely re-architecting both servers and in-browser applications, so I think it can be considered relatively independent. —jacobolus (t) 01:52, 2 June 2008 (UTC)

Do you have sources for this dominance? As soon as someone describes "Comet" the word AJAX falls, where are the hard facts of what Comet is other than a neolism for dynamic interaction via HTTP that exists since the mid 90s (e.g. nph-CGI, multipart MIME, PSYC, streaming webchats). As long as we don't have a definition of what Comet is and how it separates itself from related technologies I think it is more of a Web 2.0 buzzword. I agree with Restepc on the points he has listed, the article is way too long and leaves the impression that this is original research. - 83.254.208.192 (talk) 13:14, 2 June 2008 (UTC)

What would a source for that look like? There are no NY Times stories exclaiming “Comet is the most popular term for this!” if that’s what you’re looking for. Sure, “Comet” could be called a buzzword if you like, just like “Web 2.0” itself, or “RIA” or “Ajax” or many other terms, or for a non-web-related example, “intellectual property”. That doesn’t in and of itself make it non-notable. Also, it is “way too long” by whose standard? Who made you the arbiter of article length? I was under the impression that Wiki is not paper. —jacobolus (t) 20:23, 2 June 2008 (UTC)

If a source can't be found for something, it shouldn't be in the article. Article length is a tricky subject, and is generally down to consensus on individual articles, here the consensus is clearly that the article is currently considerably longer than it should be. Restepc (talk) 20:52, 2 June 2008 (UTC)

Grumble… the point was not that it is not verifiable. Indeed if you look around it is completely obvious common knowledge (you can take Google’s adoption of the term as “credible” evidence of a larger trend). In any case, this “consensus” view is of 2 people who happen to have been looking at this article in the past 2 days, and I am not even convinced they (you) have read the whole article. They certainly haven’t gone out to look at the dozens of linked external sources. —jacobolus (t) 21:08, 2 June 2008 (UTC)

consensus is consensus, even if it's only 3-1. I am trying to prevent this from escalating up the dispute resolution page. Restepc (talk) 21:19, 2 June 2008 (UTC)

If you like, I can easily find a half dozen people to come in here and argue for preserving or extending this article. But that would be a pointless waste of time. You need to give some reason why each part that you think it unnecessary should be cut or shortened, and then I will try to explain why I think they are useful, and then hopefully the dialog will persuade you. Just saying “I have more votes, let’s chop out 85% of the article, because I say so” is absurd. —jacobolus (t) 21:25, 2 June 2008 (UTC)

your involvement with this article has already raised concerns, I can't recommend you start canvassing for your point of view. I would consider myself to have been a neutral fourth opinion in this case, if you'd like to bring this to the attention of a wider audience you could try a request for comment. Also I would remind you that things need to be shown relevant and sourced to be included, not the other way around. For starters I would question the point of the entire 'browser compatibility' and 'scalability' sections.Restepc (talk) 21:45, 2 June 2008 (UTC)

Listen, the article has multiple intended audiences. You Restepc, Damiens, and one of the anon IP users, who had never heard of Comet before, are clearly members of an audience that doesn’t particularly care about how the technology works, but would instead like to hear a basic definition and then move on. That is completely fine. Other more technically inclined readers might instead like to learn more detailed information. This sort of diversity of audience presents a problem for any wikipedia article. Indeed, I often find articles go into much greater detail than I personally am interested in, and I skim or skip over entire sections. But that is not evidence that those articles are “too long”, but rather merely that they are not catering specifically to my interests and needs. All of the parts of this article are relevant to the subject, and are in my opinion important to some subset of readers. Why do you personally get to decide that they should be cut? —jacobolus (t) 21:59, 2 June 2008 (UTC)

Jacob, why did you personally decided what should be in? You wrote an 11 pages long article without providing 1 edit summary, why is it hard to accept that others my have a say on it? It's not that we need to convince you before making any changes, since you didn't bothered to ask anyone before writing the whole article your way. At this point, the better we can do see to work together and, without any attachments, decide what needs to be done to this text to make it look like an Wikipedia article. The more defensive you are, the painful the process will be for us all. --Damiens.rf 08:40, 3 June 2008 (UTC)

I can perfectly well accept that others might have a say. It is the deletion of the vast majority of the text with only a little, extremely vague justification which I found annoying. I generally don’t write edit summaries when adding large amounts of text, simply through habit. It certainly is not through any attempt to mislead or deceive later editors. —jacobolus (t) 10:30, 3 June 2008 (UTC)

Jacobolus, I come to believe that you are lobbying for your favourite technology. It should be easy to give a few hundred sources, articles, news coverage if this new web technology would be as popular and the market penetration as dominant as you claim. There are multiple problems with this article in my opinion. You created it without providing hard facts (original research, advertising, conflict of interest, mystification), secondly you added redirects for related terms and can not provide convincing arguments when asked for an explanation, third you effectively deleted the HTTP streaming article despite objections of others. From this point on it would be much better if you would only assist and consult with your technical expertise, the current process of regaining neutrality on this topic has been unnecessarily slow so far. - 83.254.208.192 (talk) 02:11, 4 June 2008 (UTC)

I would have to agree with 83.254.208.192 here. As I made clear in #Problems with the History section above, this article, as was written, seems broken beyond repair and needs rewriting from stub, or even become a sub-section in a broader article.

"Comet" is a buzzword created by a blogger called Alex to denote something that is not specially different or independently notable from other existing technologies.

The term became used by a small group (10 or 12) of evangelists (including the buzzword creator) that gathered on a blog called cometdaily.com to promote the use of this buzzword. Jacob Rus is one of these evangelists and is using Wikipedia to spread this meme, and to increase the perceived importance of the cometdaily website (a place where 10 or 12 people blog about comet and post comments praising each other posts).

All comet-related references in the article are from people directly linked to the buzzword creator Alex (usually by being a co-worker in cometdaily.com) or from Alex himself. No notable third part coverage of the topic seems to exist.

Jacob, at this point it seems that you're the only one believing your version of the article is good as it is. Will you help us to re-write the whole article (following Wipedia's standards this time) or we should go back to the COI page (or other conflict resolution help) to settle this dispute? --Damiens.rf 13:33, 4 June 2008 (UTC)

I don't believe further COI action or escalation up the dispute resolution ladder is needed at this point, Jacob is well within his rights to disagree with us. I think the way forward is to merge this article and the reverse ajax article into push technology, as part of a rework of the push technology article. Reverse ajax article should be deleted, and this article cut down to a stub and left for a couple of months to allow adequate time to demonstrate the notability or not of the term 'comet' (what does WHATWG refer to it as for example). Then either deleted or kept as a concise stub/article as appropriate. I am way outside my area of expertise here, but I'll do what I can. Restepc (talk) 18:08, 4 June 2008 (UTC)

I'm worried about the way Jacob ask for justifications for any change we propose, and doesn't receives the justifications very well. I would like to see a compromise from him here. He seems to sincerely believe that his writing style is clear and comprehensible, he doesn't seem to carry about consistence with Wikipedia's style, and he somehow fail to see how this article is, more than anything else, a big advertisement for commetdaily.com and it's collaborators (himself included). --Damiens.rf 18:45, 4 June 2008 (UTC)

But I'm ready to support stubfying and moving to a subsection of an article about a broader (and more solidly established) technological term. --Damiens.rf 18:45, 4 June 2008 (UTC)

In situations like this I think it's better to discuss things for a while with the other editors of an article, but yes, I think that process has pretty much finished now and I'll start cutting down this article....probably tonight. Do you think it would be better for the broader article this is moved to to be the push technology article, or a totally new one as I suggested below? If we do work on push technology as being that article, would it be a good idea to rename it to 'server push'? Restepc (talk) 19:05, 4 June 2008 (UTC)

Good questions. Regarding the last question, the terms 'server push' and 'client pull' (as the opposite) are very common terms when speaking about protocol level, developers do instantly understand what is ment. Perhaps "Push technology" is a more neutral umbrella term for concepts, technologies and protocols - but I really don't have a strong opinion which is better. - 83.254.208.192 (talk) 22:53, 4 June 2008 (UTC)

re: Restepc: As far as I can tell there has been no “process”. re:anon user: “the terms 'server push' and 'client pull' (as the opposite) are very common terms when speaking about protocol level, developers do instantly understand what is ment.” Actually these terms have several possible meanings, and are quite general, and it is therefore not usually clear what is meant without further context. —jacobolus (t) 02:02, 8 June 2008 (UTC)

I have trouble believing that you actually are looking for a compromise, since you have consistently rejected anything but complete destruction of this article. But yes, I believe my writing is clear and comprehensible when read, and I still have not heard which parts of what I wrote are hard to understand, but I would be happy to attempt to clarify them. What exactly do you understand to be the meaning of the word “advertisement”? Comet Daily is mentioned exactly once, and then several of its articles are used as sources, as they do a good job explaining various aspects of Comet. I wouldn't mind if you remove the mention from the article; it is mostly incidental, which is why it sits in the “history” section. But using a source in a wikipedia article is very different from “advertising,” in my understanding. —jacobolus (t) 02:02, 8 June 2008 (UTC)

In danger of repeating what was said multiple times by various people: it was broken beyond repair. The consensus at the time was to rebuild it from scratch. - 83.254.208.192 (talk) 17:28, 12 June 2008 (UTC)

Destroy and rebuild

The simplest way to sort out this article may simply be to cut it right down to a stub and add in more info if necessary, I'm not familiar with the notability guidelines on programming terms, but I'm beginning to wonder if this should even have its own article in the first place.

My suggestion for the form of the article if it remains is:

Comet is a programming term used to describe a method of event-driven, server push data streaming. This method uses long-lived HTTP connections as an open line of communication which the server can use to push data to the client, reducing latency.^[1] This method is used in many web applications.^{[citation needed]}

The term Comet was coined by Alex Russell in a March 2006 blog post, as a play on Ajax (Ajax and Comet are both common household cleansers).^[2] Though the concept is much older, and has been known by various names^{[citation needed]}, the term Comet has come to dominate recent discussion.^{[citation needed]}

Obviously I suggest we actually find sources where I've put citation requests, and I'm far from sure that the article shouldn't simply be merged with push technology.

Restepc (talk) 18:16, 2 June 2008 (UTC)

What the heck? Why is that the simplest way to sort out this article? Comet is not a “programming term”, and the current first paragraph is a perfectly fine description. The current article does an excellent job of introducing Comet in my opinion—and I have had several people read it, both technically inclined and not, and the response has been consistently positive. Where have you put citation requests? As far as I can tell the article is sourced just fine, but I am happy to add more citations. In any case, this is unquestionably a notable concept: seriously, go look at the list of talks at any web technology conference, and you will see large numbers of them explicitly about “Comet”. There is now at least one book about the subject, and the term is used by hundreds of companies and developers. —jacobolus (t) 19:56, 2 June 2008 (UTC)

Oh, I see. Citation requests in the little bit you've pasted there. No, what you have written is insufficient description, and badly focused. “this method is used in many web applications” is vague and nearly meaningless (no wonder it would need some citation), and the second paragraph belongs where it currently is, in a footnote. The introduction of the article should describe the concept, not dwell overly on the origin and usage of the particular name, since the concept is what the article is about. The specific name used is just a convenience, not some core aspect of the concept. (A rose by any other name and so forth.) Leaving that in a footnote is the best way to avoid cluttering the article with distractions. —jacobolus (t) 20:02, 2 June 2008 (UTC)

The reasons this should not be merged with push technology are that we have enough to say about Comet that putting push email and other types of push applications in the same article would expand it to an unreadable length, and also that many more readers will search for “Comet” than “push technology”, seeing as the latter term is nearly never used, while the former has become quite popular. If you like, you can put a summary of Comet into the push technology page, and provide a link from there to this page. Indeed, that probably should be done, but I don't personally care enough about the push technology article to completely rewrite it, which is what it needs at this point. —jacobolus (t) 20:06, 2 June 2008 (UTC)

The fact that you could (and havven) write a great deal about comet, doesn't necessarily mean you should do so (on wikipedia). How would you describe comet (in 1-4 words) if it's not a programming term? 'a word used for a World Wide Web application architecture' is a bit of a mouthful.

The relationship between push technology and comet probably needs mentioning, on either the future merged article or both if they remain separate. It seems to be that comet is essentially a modification/upgrade of push technology, in that it eliminates the need for the client to consent to being sent each piece of data, whereas beforehand although the server would be the one initiating the request, it would 'check' with the client it was okay to send it (and before push at all, the server would only send stuff when the client specifically asked for it) have I got that right?Restepc (talk) 21:06, 2 June 2008 (UTC)

How about an “interaction model for web applications” or a “web application architecture”? “programming term” is completely vague to the point of meaninglessness. No, it is not a modification or upgrade of push technology. And your sentence there describes the difference between a run-of-the-mill Ajax application and a Comet application, and has little relation to push technology, which is I think an even broader umbrella term than Comet is, requires no browser whatsoever, and can be used to describe applications like push email. But it’s unclear—the term was mostly used in the late 90’s, (e.g. here)—and I'm not sure in how broadly a sense anyone uses it today. I certainly haven’t heard it mentioned recently, other than on this talk page. —jacobolus (t) 21:20, 2 June 2008 (UTC)

Server push is clearly not only a term of the late 90s, e.g. "Scalable Server Sent Events" from the 2008 HTML draft speaks about pushing, polling, streaming and AJAX (however not about Comet as far as I can see). [7] - 83.254.208.192 (talk) 02:55, 3 June 2008 (UTC)

Restepc, I completely agree and also think that this article is a strong merge candidate -or- that it could gain by being rebuilt from scratch. I think we stumbled across one problem, that is ownership. If we can't make progress with the current article perhaps Jacob could be so kind and take a step back (you know that no information is really lost, since it is always possible to revert back to an older article version if necessary, also I am feeling unhappy that Jacob has recently added a reference to himself, this should simply not happen). Since the terminology is still unclear, my personal working title for 'Comet' is 'Streaming Ajax'. - 83.254.208.192 (talk) 02:37, 3 June 2008 (UTC)

At this point, I would also favor a complete rewrite. We could start with a definition-only stub and start adding only (GOOD-)sourced information from that point. But of course, we would have to first decide if "Comet" is really an independent technological concept, or just a buzzword defended by some clique. --Damiens.rf 08:45, 3 June 2008 (UTC)

It seems I've been confused by the state of the push technology article, many of the phrases which according to this article comet is also known as, currently redirect there. It seems both articles (and possibly reverse ajax too) need work, and the redirects need sorting out.

I'll leave this discussion going for a bit longer before cutting the article to a stub in case Jacob still has objections I haven't considered. Restepc (talk) 16:37, 3 June 2008 (UTC)

Done; the first paragraph could do with some expansion (and sources) from an expert. The second paragraph could do with a list of the various names (the list in the previous carnation came under dispute so I'll wait for sources), and would be better being more specific about how much older it is...Restepc (talk) 23:59, 4 June 2008 (UTC)

I strongly object to your most recent version. If you don't add a real description of Comet very soon, I will certainly put in a message at WP:RFC. Cheers! —jacobolus (t) 10:04, 5 June 2008 (UTC)

I took the description of comet from the Reverse Ajax article as I am not an expert in this area. If you think it's inaccurate, feel free to change it with a neutral source; our Swedish colleague had a good one the other day. As I already suggested, RFC may be the best option from your point of view if you still think that the wider community would be more forgiving of your recent input to this article and links to it. I now plan to try to help modify the push technology article, and possibly split this sorta thing into a new one if it becomes necessary, you are more than welcome to bring your expertise to help.Restepc (talk) 14:43, 5 June 2008 (UTC)

Looks like this is still not sorted out. Many parallel discussions, but at the end of the day there is still no definition of what Comet is on a technical level. On a sidenote, the so called technical experts from cometdaily didn't help that much so far in cleaning up this mess. - 213.115.160.102 (talk) 16:47, 11 June 2008 (UTC)

Long list of redirects

I have seen that there are redirects to the Comet article, which should point to other articles in my opinion. The following is my suggestion. -83.254.208.192 (talk) 04:21, 3 June 2008 (UTC)

HTTP streaming -> Push technology
HTTP Streaming -> Push technology
HTTP push -> Push technology
HTTP stream -> Push technology
Pushlet, Pushlets -> Push technology
Full duplex Ajax -> DELETE (gives <20 Google hits when ignoring "lighttpd" as keyword)
Full-duplex Ajax -> DELETE (gives <20 google hits when ignoring "lighttpd" as keyword)
Reverse AJAX -> Reverse Ajax

It's interesting that there used to be an article on HTTP streaming, but Jacob Rus/Jacobolus decided to wipe it out and redirect to his preferred term "Comet", despite the discussion at the time achieved no consensus for doing so. --Damiens.rf 08:59, 3 June 2008 (UTC)

I wiped nothing out, but rather merged all relevant content. These articles were all stubs, and the Comet article mentions the alternate names prominently in its introduction. The point is to have one article for a single concept. I honestly don't care all that much what it is called. At the moment, sending “HTTP streaming” and “HTTP push” to “push technology” is a bad idea, because they describe different concepts. Likewise, “Reverse Ajax” should not be an independent article at all, as it is an alternate name for the (“Comet”) architecture described here, a name used by almost no one (all its uses originate from a single source, and the term has gained no currency as far as I can tell). --jacobolus (t) 10:19, 3 June 2008 (UTC)

Ouch, this shouldn't have happened. I wondered why there was no HTTP streaming article. From my technical point of view 'HTTP streaming' and 'HTTP server push' are close siblings if not even synonyms for each other. Comet can be a subcategory of a streaming technology but not every HTTP streaming technology is Comet (also see comments from Lynx and Sprocketonline in other sections of this talk page). I strongly suggest to merge the old HTTP streaming article into Push technology for now, maybe it can be reinstantiated as an own article. Having commented on streaming, there also have been objections to say that 'Reverse Ajax' and 'Comet' are the same thing, in any case the term 'Reverse Ajax' has been around longer (already mentioned in 2005 [8]). Jacobus I start wondering on what grounds you decided that 'Comet' is the most prominent term for all these terms? - 83.254.208.192 (talk) 12:00, 3 June 2008 (UTC)

I decided based on looking around the web, at the discussion of the topic, at the titles and blurbs from talks at technical conferences, at the books coming out, at the demos and applications using the technology. Comet is by far the most popular term of the last year and a half or so. Which name came first seems like a poor method (by that standard we would merge Istanbul and Constantinople into Byzantium). I honestly don't care what the article is called, or what the technology gets called. My reason for merging these articles is that they all described nearly exactly the same thing, not that I want to stomp out alternative names. It is silly to have 5 wikipedia articles under different names, for the same concept, especially since it is easy to list off alternative names in the introduction. —jacobolus (t) 23:58, 3 June 2008 (UTC)

Ah, it seems we may be in agreement; look below :) Restepc (talk) 01:03, 4 June 2008 (UTC)

Merger of all these things into new article

Article title something like 'web server client interaction models'

In that title comet/whatever you want to call it, polling, piggybacking, etc would each be described, rather than have separate articles for each where the less notable ones get merged into the more notable ones when they're not the same thing. Like a more detailed version of the current reverse Ajax article but under a title which isn't a competing term/method.

It seems by far the simplest way to explain each of these concepts is by contrasting with the alternatives.

Sound good to anyone? Restepc (talk) 01:01, 4 June 2008 (UTC)

The one problem I can imagine with that is that it is a broader topic than this one, and so it might lack focus, and become confusing, if it tries to completely cover the subject. Such an article should certainly exist, but I think that the particular case of real-time web applications (as is described here) merits its own article as well. —jacobolus (t) 01:10, 4 June 2008 (UTC)

I suggest you try researching and writing an article like that, incorporating text from this article if it is useful, and then come back to this discussion once that is reasonably written, to consider what to do afterwards. —jacobolus (t) 01:11, 4 June 2008 (UTC)

I don't think that is needed right now, there is not the amount of content that justifies an own article. We could start small in the 'Push technoloy' article with individual sections and create extra article(s) if one section becomes too long. This would give an opportunity to focus on the definitions and similarities first, instead of going straight into the details. 83.254.208.192 (talk) 01:40, 4 June 2008 (UTC)

Agreed that this is the best option right now. Ripped out of original research and peripheral discussions, very few would be left from the "Commet" article. We should redirect this page to Push technology#HTTP server push and work to improve that article. If that section ever gets too big (what would be reasonable to expect), we could create the independent article HTTP server push (or simply "HTTP push").

Of course, while working on Push technology#HTTP server push, we should avoid the same mistakes committed in Comet (programming). --Damiens.rf 20:52, 4 June 2008 (UTC)

Disagree strongly, a summary article would be very useful and have a significant amount of content. The distinctions between various asynchronous HTTP paradigms are important, and discussions of the relative strengths and weaknesses of each are important for providing context. 63.64.108.5 (talk) 18:07, 16 June 2008 (UTC)

common household cleansers

We need a source for the claim that "Comet" was chosen as a play on Ajax as both are common household cleansers. I know this is probably truth, but... the threshold for inclusion in Wikipedia is verifiability, not truth. --Damiens.rf 21:25, 5 June 2008 (UTC)

There are several pages of well-verifiable material that were ripped out of this article, in your path to destroy it. Explaining the silly joke behind its name is pointless, by comparison. Go ahead and rip it out. Why leave the vandalism half finished? —jacobolus (t) 06:04, 6 June 2008 (UTC)

You're just being rude here. If you're not prepare to contribute to an article that will not blindly promote your agenda, consider stepping away of the process. --Damiens.rf 11:35, 6 June 2008 (UTC)

There is no article left to contribute to. —jacobolus (t) 14:00, 6 June 2008 (UTC)

I hope you get back when you heal your feelings. --Damiens.rf 15:51, 6 June 2008 (UTC)

Why would he waste his time when you've taken an excellent resource on a very popular technology and butchered it?. Comments above such as "by a blogger named Alex" show your complete lack of understanding of the space. Alex Russell is the co-creator of the Dojo Toolkit and Cometd, among other things, one of the most highly respected people in the world of Comet and Ajax. --Dylan.rf 06:06, 7 June 2008 (UTC) —Preceding unsigned comment added by 70.167.219.162 (talk)

Of course he's "one of the most highly respected people in the world of Comet", he made the whole thing up! I, for one, am one of the most highly respected people in the world of Cryptoflugny (that happens to be something I invented myself). Wait until some friend of mine start writing Cryptoflugny (dancing) --Damiens.rf 20:20, 7 June 2008 (UTC)

Dude, seriously? Come on now, see Greg's post below. How do you think the term Ajax came to be? Someone (Jesse James Garrett) in the field came up with a term to describe a group of techniques. Comet accomplishes the same thing, and is widely used across the industry (again, see Greg's post below). Or, take a look at the list of projects on Google Code and Sourceforge that use the term Comet, or the number of implementations on the original page before you removed the list. If Wikipedia doesn't want this information, that's a shame. I still don't really understand your objection to the work done by Jacob over many months with community feedback? It doesn't really matter... the content is open source anyways, I just don't understand why you feel the information presented is not useful? Your objections listed above seem pedantic and it feels like the goal of creating a great resource of useful information is being ignored by your desire for a process that I fail to understand. If any of the open source software projects I contribute to behaved in the manner you have behaved in, I would be embarrassed and ashamed. --Dylan.rf 00:08, 8 June 2008 (UTC) —Preceding unsigned comment added by 70.167.219.162 (talk)

Full disclosure here, as Restepc has accused me of WP:CANVASSing “against” him. Dylan, knowing that I wrote most of the Wikipedia article here, sent me an email yesterday asking me what was happening with this Wiki page, to which I have just responded, a day after his comments here, suggesting to him that discussion on the talk page is the method by which Wikipedia disputes are resolved. I have no interest in compromising wikipedia process, and Dylan speaks for himself. —jacobolus (t) 02:26, 8 June 2008 (UTC)

Damiens, trivializing this discussion with absurd straw man arguments does not further your position. —jacobolus (t) 02:31, 8 June 2008 (UTC)

Comet deserves it's own page

This page (as of 8/June/2008), is totally insufficient for the emerging technology called comet.

I'm the main developer of the Jetty webserver and have been working with internet since 1995. Certainly comet relies on pre-existing push techniques and I for one was doing similar things with netscape in the late 90s.

However comet encompasses is more than just "push" protocols - it is a about implementing "push" in a browser environment with the connection limitations and security sandbox issues. Prior to the coining of the term "comet", it was very difficult to simply describe the techniques and the issues they raise with existing network infrastructure.

Through 2007, the term "comet" got increasing traction and is now widely understood in the industry. In 2008, it is mostly possible to use the term Comet without the need to include the terms "Ajax" and/or "Push". For example there were many talks at the javaone 2008 conference that included comet in their title, their abstract, or content.

I am on the servlet expert group for JSR-315 (for the servlet-3.0 specification). The comet use-case has been a high priority for this iteration of the specification and the servlet API is being modified to cope. The popular tomcat server has provided support for this feature with a servlet called CometProcessor. I am also on the Open Ajax alliance, and the specific security and communication issues of comet are given significant consideration in their efforts for interoperability of Ajax and lobbying infrastructure providers for change.

More over, while comet is based on push technology, the terminology used is a little different. In order to discuss comet, you need to define such terms as "long polling", "JSONP callback", "forever frame", "cross domain", "2 connection limit". These phrases are the currency of comet discussions and you will not learn them from the push technology page.

I do not think this page should be merged with a push page, however it can refer to some general techniques that may be defined there. (Note that "push" is not really a good name either... there is no actual push in many techniques).

The current content of this page is totally insufficient for a significant movement in web technology.

I should also say that I'm a contributor to the cometd project at Dojo foundation, and an author on the bayeux protocol specification.

Gregwilkins (talk) 23:42, 7 June 2008 (UTC)

Please help to improve this article and contribute with your technical knowledge. I am not convinced about the notability, what makes me wonder is that the Ajax article doesn't even mention Comet. The latest HTML draft doesn't mention it, no other W3C document mentions it. Web developers here in Europe that I spoke with haven't heard of it, so it doesn't look like a well known Ajax design pattern and by far not like a recognisable web technology. Perhaps Comet is a new term used within a small circle? One big problem with this article is that we have seen lobbying (e.g. from/for Cometdaily) and very little facts, where are the sources? Let me comment on some of the "Comet currency" you mentioned: Ajax long polling is also used by AOL's Web AIM and the 2 connection limit as defined in RFC 2616 concerns all HTTP server push technologies. - 83.254.208.192 (talk) 09:03, 8 June 2008 (UTC)

"I am not convinced about the notability, what makes me wonder is that the Ajax article doesn't even mention Comet. The latest HTML draft doesn't mention it, no other W3C document mentions it." For what it's worth, the HTML5 specification does not mention Ajax either, yet the specification incorporates APIs, technologies and concepts for creating Comet-based applications: Server-Sent Events, cross-document messaging (which can be used in conjunction with SSE or legacy Comet techniques to do safe cross-domain Comet). Does the specs lack of mention on either warrant deletion of the articles? (And before someone asks, I won't edit the article, given that I work for a browser vendor that has implemented technologies for modern Comet Arve (talk) 17:21, 16 June 2008 (UTC)

Ajaxian, a leading site about Ajax, has 45 entries tagged with Comet [9]. Simon Willison, Ajax expert, has given several talks about Comet at conferences: [10] . OpenAjax Alliance talks about Comet: [11]. Every major Ajax and web conference I've attended in the past two years has had one or more Comet talks (JavaOne had approximately a dozen talks on Comet... Sun's Grizzly and GlassFish server has used the Comet term for two years [12], and IBM's WebSphere includes Comet support: [13] Of course it's a newer term than Ajax, but what facts am I missing? Comet is defined as a combination of techniques (long-polling, forever-frame, server-sent events (part of HTML 5 drafts), multi-part MIME, that encompass ways to do things more efficiently than Ajax, and the server-side implementations necessary to make this happen. I'm sorry if you feel that Comet Daily is lobbying... I created Comet Daily as a collaboration between people interested in Comet, to talk about it, and raise awareness for it. That said, I don't care if Wikipedia uses this information enough to go on proving whether a term is important or not... if you don't want it, we'll reuse it at Comet Daily under the terms of the GFDL and get on with more important things. dylanks

I have come to the conclusion that comet does warrant a wikipedia article...and it has one; I realise you preferred the old version, but it went against a number of wikipedia policies/guidelines to the extent that scrapping it and starting again was the best option in my view. This article will continue to exist; if someone was to nominate it for deletion I would argue to keep it, but it needs to be edited within wikipedias guidelines. Restepc (talk) 03:38, 9 June 2008 (UTC)

OK, I've started contributing little improvements to the text here and on the push technology page. They contained a number of minor falsehoods and a few barbs from an obvious comet skeptic. Hopefully I've corrected those and not gone the other direction nor broken any guidelines... but will sit back now for a bit and see if what I've done is acceptable. Note that it is difficult to find definitive citations other than say something like: "google for long-poll". Comet has developed in the era where we look to wikipedia to provide the definitive citations :-) Gregwilkins (talk) 08:20, 9 June 2008 (UTC)

Fine. Then the question is: what needs to be said about Comet in a comprehensive self-contained Wikipedia article. I posit that the most important aspects of such an article are:

A description of what Comet is, which requires in my view a short description of how other web application architectures work, without which the description of Comet makes no sense to those not already intimately familiar with web technology
A description of the various techniques used, such as invisible iFrames, XHR objects, dynamic script tags, and the proposed HTML5 Server-sent events, including at least how they work, but ideally also the pros and cons of each one.
A history section
A section on scalability, including some description of what publish/subscribe is would be also very useful to anyone trying to understand Comet as it is actually used, as this is one of the most difficult problems facing Comet developers, and is essential background for anyone wishing to discuss Comet.

Cheers!

Restepc: I expect you and anyone else who objects to the article as it was before to put actual effort in to trying to make a real article, instead of merely pushing for the article’s destruction as you all have so far. —jacobolus (t) 18:01, 9 June 2008 (UTC)

Also, I would appreciate a restoration of the introduction as it was before this article was butchered, and barring some convincing argument why it shouldn't be restored, I will restore it in the spirit of WP:BOLD. I believe it does a significantly better job introducing the subject, wiki-linking to key outside terms, etc., than the current introduction, and also is much more readable prose. —jacobolus (t) 18:04, 9 June 2008 (UTC)

At this point, I'm not sure we can write something like an "history section". It seems nothing has ever been published about "The History of Comet", and Wikipedia should not be the first to do so. Do you know some such publication?

About the leading section, I like the current one much, even though it's still isn't perfect.

As a side note, you have been warned before about potentially offensive statements like "...this article was butchered..." or "...pushing for the article’s destruction...". You had enough time to understand how the text you wrote was anything but an encyclopedic article. Time to stop complaining about its removal. --Damiens.rf 21:22, 9 June 2008 (UTC)

I realize you are intentionally being a jerk in an attempt to make me angry. It has worked. Please stop it now. It is utterly unproductive. —jacobolus (t) 22:38, 9 June 2008 (UTC)

Stop the name-calling right now. --Damiens.rf 12:19, 10 June 2008 (UTC)

Comet can also be seen as an Ajax design pattern. If that is the case then it doesn't need an own article and can merge with the Ajax page. The visibility outside the Ajax world is close to zero: ask a web developer if they heard about Ajax and they will confirm, ask what Comet is and the answer is very likely 'what?' 213.115.160.71 (talk) 17:43, 10 June 2008 (UTC)

And others describe Comet as a technology collection that includes any kind of JavaScript (Ajax being on of them). I don't know which is more correct, but there is one interesting question: Is the terminology "Comet" actually used outside the Ajax world or has this close relationship simply turned into a synonym for "Ajax push" over the years? - 83.254.208.192 (talk) 10:00, 14 June 2008 (UTC)

Introduction / Purpose of "Comet"

I have reverted the piece taken from the original article that used to define "Comet" as an "Ajax-based web application architecture"", to the more realistic and less far-fetched "a flavour of Ajax". I believe giving this neologism the status of an architecture is overly pompous. --Damiens.rf 12:19, 10 June 2008 (UTC)

Maybe another valid definition would be "a set of ajax techniques"? --Damiens.rf 12:20, 10 June 2008 (UTC)

What the hell is a “flavor” of Ajax? That is unsourced meaningless babble and is not suitable for Wikipedia. —jacobolus (t) 19:36, 10 June 2008 (UTC)

Please chill out and read my (upcoming) comment on the section below. --Damiens.rf 22:05, 10 June 2008 (UTC)

Not very constructive, and you know there are also other people who call it a flavour of Ajax. -- 213.115.160.102 (talk) 16:24, 11 June 2008 (UTC)

The original is much better, it is indeed an application architecture. Flavor is less specific. In the article 213.115.160.102 quotes, in addition to saying it is a 'flavor of Ajax', also says: 'Comet uses a modified form of the canonical Ajax architecture' and then goes on to describe the ARCHITECTURAL differences. Furthermore, 'web application architecture' gives wide context, which 'Ajax-based' further refines. It is much better than 'a flavour of Ajax'. Damien, you are the one being pompous in your rigid fight against this 'neologism'. Your continued focus on it being a new word coined by several 'evangelists' points to your bias against it. Who cares if it has similarity to previous architectures, or that it is a subset of 'Ajax' technologies (Ajax is also a neologism, OOOO!!!!) We create words as we need a shorter way of mentioning complex concepts. It facilitates communication. Please don't work against simple, organic progress.63.64.108.5 (talk) 18:21, 16 June 2008 (UTC)

At the end of the day, Comet is not even a flavour/variation of Ajax, it is Ajax (long) polling? - 83.254.214.192 (talk) 22:12, 4 July 2008 (UTC)

ComeComet-like technology standards proposals

I've just noticed I have accidentally removed a paragraph that mentioned html5 and json (I though I was just reading a removed mention of json). I'm sorry for that. But re-reading the paragraph, I have doubts if it should be re-added as it is:

"Two standards, Crockford’s JSONRequest and HTML5’s Server-sent events, have been proposed to ease the adoption of Comet-like technology. Neither has seen widespread browser adoption or use"

First, I believe that characterizing things like HTML5’s Server-sent events as "Comet-like technology"reveals an implicit slight pov: Considering the influence of each one, it's more like that "Comet" is a "HTML5’s Server-sent events"-like technology.

And second, we should avoid a kind o loose writing the permeated the old version of the article, where the article tries to guess the motivations behind everyone's steps: "...have been proposed to ease the adoption of...". We should not infer the reasons for the proposal of these standards (let alone to stated that the reason is to allow Commet to be used).

The loose language is seen again with "Neither has seen widespread browser adoption...". Since we have any sources, this is solely based on the author's experience (and thus is original research). In any case, as far as I know, html5 is just a planed proposal. No "widespread browser adoption" should be expected at this point (but the paragraph makes it sound otherwise). --Damiens.rf 12:35, 10 June 2008 (UTC)

The reason I put that paragraph, is that when I removed “JSONRequest” from the list of alternate names for Comet, the anon contributor who had added it reverted my removal. So by way of explanation, when I moved it out of the list of alternate names again, I instead described what it actually is. It seems perfectly reasonable to leave JSONRequest out of this current article, stub that it is. —jacobolus (t) 19:25, 10 June 2008 (UTC)

No it isn't "perfectly reasonable", stop acting like in a highschool debate club. The reason it is in because it is a push technology with JavaScript on the client side... it can't get much closer to what Comet is, compared to e.g. Pushlets which is also listed and use a different programming language. I think you try to water down this article as much as possible, in the past you did not give bare facts but mystifications instead. One reason I could think of why you are doing this is that in an attempt to lobby for Comet you try to weaken or hide all related web interaction technologies (so far: deleting articles, removing related technology terms, ridicolus claims about how popular Comet is, etc). -- 213.115.160.102 (talk) 16:33, 11 June 2008 (UTC)

Erm… Damiens.rf removed it, and I termed his removal reasonable, in an attempt to be conciliatory. I have no interest in “weakening” “related technologies”. (aside: out of curiosity, do you understand technically what JSONRequest is, who uses it, which browsers support it, which standards bodies have discussed it, etc?) I did merge several synonymous terms into one article, in an attempt at consolidating the information about a single topic. Please explain which claims of Comet popularity are ridiculous—I certainly am not aiming at ridicule here—and which explanations are mystical. —jacobolus (t) 14:28, 16 June 2008 (UTC)

Citation needed

Since we're re-writing the article from stub, there's no need to keep non-cited information in the article. I've just removed some disputed pieces. Anyone should feel free to re-add them with proper sources. --Damiens.rf 12:43, 10 June 2008 (UTC)

Removal of secondary source

That source is all the article has for the 'definition' of comet; if you want to remove it you'll need to find a better one, if you can't even find a source for the definition we're not going to get anywhere. Restepc (talk) 20:14, 10 June 2008 (UTC)

Go ahead and put the article at AfD if you like. Or go look through the dozens of sources that were previously linked: each one was specifically chosen because it was worthwhile for interested readers. The Ajaxian source, as well as the nameless mailing list post, are both throw-away sources, apparently used to give the appearance of legitimacy to this stub, like putting lipstick on a pig. Neither one is worth linking readers to, and using the former to justify calling Comet a “flavor” is intellectually dishonest. —jacobolus (t) 20:21, 10 June 2008 (UTC)

I do not recall there being any sources at all in the introduction of your article. Which source from the old article would you suggest using for the definition? Restepc (talk) 20:30, 10 June 2008 (UTC)

I am sure you understand why Wikipedia editors prefer external references over personal opinions. - 83.254.208.192 (talk) 08:00, 14 June 2008 (UTC)

Leading paragraph (or "what's comet anyway?")

Powell, Thomas (2008). Ajax: The Complete Reference. McGraw-Hill Osborne Media. p. 654. ISBN 007149216X. {{cite book}}: Unknown parameter |month= ignored (help)

This book on ajax by McGraw-Hill refers to comet as a "networking pattern" at page 280 , and then as a "...push-style communication pattern generally dubbed Comet..." at pages 483 and 515, and as a "...push-oriented communication pattern..." at page 516. I believe it means a desing pattern when it says pattern.

I suggest we use this terminology on the leading paragraph. --Damiens.rf 22:07, 10 June 2008 (UTC)

Comet sources

Here is a list of Comet sources. Many if not most of these should be linked from the article, and some others besides. I'll try to keep tracking down more. I'm not going to bother writing prose. It will just be reverted, so there is no point. —jacobolus (t) 21:39, 10 June 2008 (UTC)

Alessandro Alinone (December 2005). “Changing the Web Paradigm”. Lightstreamer white paper.
Alessandro Alinone (October 2006). “Next Generation Push: a viable technology for web-based FX market data delivery”. e-Forex magazine, issue 24.
Alessandro Alinone (19 October 2007). “Comet and Push Technology”. Comet Daily.
Dion Almaer (29 September 2005). “Jotspot Live: Live, group note-taking” (interview with Abe Fettig). Ajaxian.
Brent Ashley (May 2006). “Ajax Transport Layer Alternatives”. Presentation at Ajax Experience.
Brent Ashley (6 June 2006). “Ajax transports still need to evolve”. Brent Ashley’s blog.
Andrew Betts (4 December 2007). “Cross Site Scripting Joy”. Comet Daily.
Andrew Betts (January 2007). “Browser Techniques”. Meteor website.
Clint Boulton (27 December 2005). “Startups Board the AJAX Bandwagon”. DevX News.
Engin Bozdag, et al. (2007). “A Comparison of Push and Pull Techniques for AJAX” (pdf). Delft University of Technology technical report.
Rob Butler, et al. (June 2006). “Bug 14392: Add support for multipart/x-mixed-replace to XMLHttpRequest”. Webkit Bugzilla bug tracker.
Michael Carter (25 October 2007). “HTTP Streaming and Internet Explorer”. Comet Daily.
Michael Carter (18 November 2007). “IE ActiveX("htmlfile") Transport, Part II”. Comet Daily.
Michael Carter (24 October 2007). “Comet for Highly Scalable Applications”. Presentation at The Ajax Experience Boston. (Link goes to PDF slides).
Jon Chambers (10 December 2006). “Flash-to-JavaScript communication with a streaming HTTP connection in Internet Explorer”. Blogature (Thinkature blog).
Grzegorz Daniluk (25 January 2007). “Full-Duplex Ajax module for Lighttpd”. Refwell.
Andi Egloff (2 October 2006). “Comet Basics”. Andi Egloff's Weblog.
Jesse James Garrett (18 February 2005). “Ajax: A New Approach to Web Applications”. Adaptive Path.
Ian Hickson, et al. (2005–2008). HTML 5 specification, § 6.2: Server-sent DOM events. WHATWG.
Bob Ippolito (5 December 2005). “Remote JSON – JSONP”. from __future__ import *.
Rohit Khare (August 2005). “Beyond AJAX: Accelerating Web Applications with Real-Time Event Notification”. KnowNow white paper (link from the Wayback Machine).
Tobie Langel (28 May 2008). “An Interview with Piotr Dachtera”. Prototype JavaScript library blog.
Michael Mahemoff (5 March 2006). “Comet: It’s Ajax for “Push” (Podcast)”. Software As She’s Developed.
Michael Mahemoff (6 June 2006). “Portable Comet? It’s the IFrame, Stupid!”. Software As She’s Developed.
Matt Marshall (15 December 2006). “Renkoo launches event service — in time to schedule holiday cocktails”. Venture Beat.
Philip McCarthy (17 July 2007). “Write scalable Comet applications with Jetty and Direct Web Remoting”. IBM Developer Works.
Gregor Roth (20 March 2008). “Asynchronous HTTP and Comet architectures”. JavaWorld.com’s Java Enterprise Edition.
Daniel Rubio (20 February 2008). “Comet: Reverse Ajax for streaming data from the server”. SearchSOA.com’s The Web Services Advisor.
Jacob Rus (11 December 2007). “The Future of Comet: Part 1, Comet Today”. Comet Daily.
Jacob Rus (10 January 2008). “The Future of Comet: Part 2, HTML 5’s Server-Sent Events”. Comet Daily.
Alex Russell (6 August 2005). “Toward server-sent data w/o iframes”. Continuing Intermittent Incoherency.
Alex Russell (12 February 2006). “What else is buried down in the depths of Google’s amazing JavaScript?”. Continuing Intermittent Incoherency.
Alex Russell (3 March 2006). “Comet: Low Latency Data for the Browser”. Continuing Intermittent Incoherency.
Alex Russell (22 July 2006). “Cross Domain Comet”. Continuing Intermittent Incoherency.
Alex Russell (21 December 2006). “Adventures In Comet and Multipart Mime”. Continuing Intermittent Incoherency.
Alex Russel et al. (June 2007) “Bayeux Protocol”. Standard proposed by the Dojo Foundation.
Dylan Schiemann (18 October 2007). “Why Comet Daily?”. Comet Daily.
Jonathan Share (3 May 2007). “Scalable Server Sent Events”. Opera Web Applications blog.
Maciej Stachowiak (26 June 2007). “Re: XMLHttpRequest and readyState==3”. Webkit-dev mailing list.
Johnny Stenback, et al. (March–April 2004). “Bug 237319 – Add support for server push using multipart/x-mixed-replace with XMLHttpRequest”. Mozilla Bugzilla bug tracker.
Martin Tyler (30 November 2007). “The Evolution of Comet at Caplin”. Comet Daily.
Just van den Broecke (3 January 2007). “Pushlets: Send events from servlets to DHTML client browsers”. JavaWorld.
Joe Walker (23 October 2007). “Why Comet is of Growing Importance”. Comet Daily.
Greg Wilkins (6 January 2006). “Ajax, Comet, and Jetty”. Webtide white paper.
Greg Wilkins (6 November 2007). “Comet is Always Better Than Polling”. Comet Daily.
Simon Willison (5 December 2007). “Comet works, and it’s easier than you think”. Simon Willison’s Weblog.

Also, at least one book about Comet is forthcoming:

Dave Crane & Phil McCarthy (July 2008). Comet and Reverse Ajax: The Next Generation Ajax 2.0. Apress.

Thanks for the step, but please, save us from the self-published ones, like the ones from cometdaily.com (your blog), webtide.com (User:Gregwilkins's blog), lightstreamer.com (commetdaily folk Alessandro Alinone's blog), meteorserver.org (commetdaily folk Andrew Bett's blog), alex.dojotoolkit.org (the Comet "creator"'s blog) and ajaxian.com (some other guy's blog).

Anyway, I'm working on finding good published sources for the article. --Damiens.rf 22:00, 10 June 2008 (UTC)

All of the best sources in the web development world are self-published. Several of the sources listed above come from high-level developers at large companies, or project leaders of prominent open-source projects. Their posts are certainly more reliable than a book from Apress, or a blog post at Ajaxian. —jacobolus (t) 22:15, 10 June 2008 (UTC)

It seems you completely fails to understand what a reliable source is, since you believe these self-published blog posts (including some by yourself) are "more reliable" than an independent, third-part, published book. --Damiens.rf 22:32, 10 June 2008 (UTC)

Yes, that is correct. A blog post by Alex Russell is more reliable than a book published by Apress. —jacobolus (t) 23:07, 10 June 2008 (UTC)

Great. So, please, keep the good work on your blog and leave the Wikipedia article for us. --Damiens.rf 23:15, 10 June 2008 (UTC)

I have stroked out the blog-entries. Anyone feel free to revert if believing more discussion is necessary, but make sure to read WP:RS and understand the concept of (and the problems with) primary sources. --Damiens.rf 22:39, 10 June 2008 (UTC)

And I have un-stroked them, as it makes the links unreadable. All but a couple of the rest of those sources are also self-published (notice, for instance, conference talks at tech conferences are just as self-published, and certainly no more reliable than, blog posts at edited blogs). You're missing the point. I am putting those links up because I suggest you (or any editor actually interested in productive contribution) go read them, at which point you will have some understanding of what Comet is, and will stop introducing factual inaccuracies into this page. —jacobolus (t) 23:06, 10 June 2008 (UTC)

Ok, as just prefer. But please chill out and try to stop attacking me on every interaction.

About "conference talks", at least they have to be accepted by the conference organizers, it's not just writing something you like and clicking "[Post to My Blog]".

DISCLAIMER: For those willing to follow jacobolus's suggestion above of "go read them", it's never too much to disclaim that the user jacobolu's (Jacob Rus) contributes for the Comet Daily blog (the single-most-linked website above), and most of the other blogs mentioned are personal blogs for other Comet Daily's contributors.

As a last point, what am I missing in the "Webkit Bugzilla bug tracker" reference? --Damiens.rf 23:58, 10 June 2008 (UTC)

EVEN MORE URGENT WARNING: jacobolus may be the root of all evil. Be warned that reading his writing could result in blindness, insensitivity towards small animals, or eternal damnation. —jacobolus (t) 00:20, 11 June 2008 (UTC)

So if the cometdaily site had been unsuccessful in attracting all the major comet developers to contribute, the these citations would be valid. But because there a web site has been created that does link to many of the sources of comet work, that some how invalidates those citations. The white papers I wrote on comet pre-date the creation of the cometdaily website - yet because I now contribute to the cometdaily website, my papers no longer count as separate citations??? Gregwilkins (talk) 02:46, 11 June 2008 (UTC)

Not exactly....citing any self published work, which definitely includes cometdaily, is strongly discouraged...although personally I make exceptions, many don't (or don't except in very obvious non-controversial cases, for instance citing rusells blog as him having come up with the term). Also wikipedia in general is highly suspicious of people citing/writing about themselves, their site, or sites they're related to. To give an extreme example, at one point in Jacobs version of the article he actually quoted himself; I'm sure you can see the problem with that. cometdaily isn't absolutely uncitable, but if it were to be used at all people may object to it and remove it purely because it's a blog (and they'd be completely in the right as far as wikipedia policies are concerned).

So unless you were planning on writing 'a beginners guide to comet' and getting it published, I doubt you'd be citable anyway. (but maybe you are...what've you got that isn't self-published?)

You are more than welcome to edit here, but reading WP:COI and WP:RS might help you understand the situation better. Restepc (talk) 03:11, 11 June 2008 (UTC)

I don't quite understand the problem with cometdaily. I can understand that if i start my own website and write something on it myself then that could be questionable - but as a contributer to cometdaily i am not self publishing, i am submitting an article which is reviewed and published by cometdaily. I dont really see how that is different to publishing a book.. or maybe a better example which has been mentioned - a conference organiser (just like cometdaily) gathers speakers/contributers that they see as experts in the subject. Martintyler —Preceding comment was added at 11:12, 11 June 2008 (UTC)

Maybe because their credibility is down close to zero? Javobolus did quite a lof of lobbying and quoted (exclusively) himself and his friends for the cometdaily website. Right now it is hard to say if he damaged cometdaily's credibility by dragging them into this conflict or if cometdaily is much more than an advertising factory. Greg (also from same website) didn't help much in regaining credibility on cometdaily's behalf. -- 213.115.160.102 (talk) 16:38, 11 June 2008 (UTC)

I am writing to ask for some clarifications, since I admit I am not an expert of Wikipedia policies and guidelines. I started working on Push Technology in 2000, when I created Lightstreamer. When the "Comet" term was coined within the Ajax community in 2006, we decided to adopt it and to define Lightstreamer as a "Comet server", because this new word describes very well what Lightstreamer is. Before 2006, during trade shows we needed to use long sentences to describe Lightstreamer (something like "a push engine for delivering real-time data to html clients", etc.). Then we started to employ the Comet term, which seems to be pretty well understood, especially at Ajax shows (while at finance shows we are still evangelizing on its meaning; actually I'm right now at SIFMA in New York, explaining people what Comet is and what advantage it can provide to financial institutions). In 2007 I was invited to contribute to the Comet Daily blog, and I was glad to accept. Now, does that mean that all my previous and current writings and talks should be invalidated because I write on Comet Daily too? By the way, Comet Daily has done a lot on effort on helping define Comet and teaching the many aspects of it. It is a very open community which gives space to anybody who has some solid background on Comet (which typically means being the creator of some Comet product, be it open-source or commercial). So, what is bad in having some of the major Comet experts collaborate and interact on the same blog? --Alessandro Alinone 207.41.180.157 (talk) 18:11, 11 June 2008 (UTC)

On Wikipedia, to be considered reliable sources should have gone through fact checking and/or editing by a reliable third party which is normally a trade magazine, a print publisher, a newspaper, etc. Blog sites don't qualify. So any articles that you wrote and got published in a magazine would fine as sources, while any articles that you wrote and just published on a blog or a private website would not be. See Wikipedia:Reliable sources for more detail. This has nothing to do with whether you post on Comet Daily or not - though there is a related point which is that writings that have only been published on Comet Daily are not going to be considered reliable sources according to wikipedia's policies. - MrOllie (talk) 18:29, 11 June 2008 (UTC)

Thanks for the clarification. So I think that my article that was published on e-Forex magazine in 2006 should be added as Reference number 8. e-Forex is no doubt a leading magazine in the financial domain. In this article, the Comet term and some of its synonyms are explained. I guess I cannot add the article myself and perhaps it would be better if nobody from Comet Daily does. So it would be good if any of the other contributors to this discussion adds the reference. Here are the details to create the link. Article title: “Next generation Push: a viable technology for web-based FX market data delivery”; published on e-Forex Magazine October 2006 Issue 25. Digital version (PDF) available upon subscription from e-Forex or freely available from Lightstreamer web site (http://www.lightstreamer.com/docs/Lightstreamer-eForex.pdf). --Alessandro Alinone 207.41.180.157 (talk) 21:25, 11 June 2008 (UTC)

E-forex? They say that they "can accept no responsibility for inaccuracies that may appear ... [on] ... editorial and other content provided by other parties". This surely applyies to the website, but I haven't read the disclaimer at their magazine.

Maybe we would have to use our discernment to judge the reliability of this source. In any case, remember that a published opinion is still an opinion, and we should restrict ourselves to take the only the facts from reliables sources.

After a superficial reading, the article looks promising. It's informative, and apparently low on evangelism and mystifications. Thanks for the colaboration. --Damiens.rf 22:01, 11 June 2008 (UTC)

Note that I'm not "from cometdaily" any more than damiens.sf (or the anonymous IP address guy) is from wikipedia. Cometdaily is just a media service to which I contribute. But to be constructive, here are a couple more citations:

JSR-315 Is the Java Community Process standards process for the upcoming 3.0 release of the servlet spec. It has proposed features for "Async and Comet support" and refers to "The comet style of Ajax web application" as sufficient explanation for these features. Note that I am on the JSR-315 EG and I have proposed a solution to that feature request, but I did not write the 315 Specification Request. It is reviewed by the JCP and the EG includes experts from Sun, IBM, BEA,SAP, ericsson, etc.
InfoQ have a page dedicated to comet news and articles.
push vs pull is an article on InfoQ reviewing a study (cited above) done by Delft University of Technology about Ajax push and pull. The study (self published by the university) describes comet as "The application of the Service Streaming scheme under AJAX is now known as Reverse AJAX or COMET", but cites Alex Russell for this.

Actually, this is pointless.... every good article I find cites Alex's blog or my blog or one of the other cometdaily contributors. Everybody self publishes now - even the university. Any of the news organizations that review and link to to the self published articles appear to be labelled "blogs" and dismissed. I'm going to comment more on the standdown message below Gregwilkins (talk) 00:20, 12 June 2008 (UTC)

Can someone explain why the currently referenced articles are more 'reliable' and less 'self publishing' than all the ones being complained about please? I mean, a usenet post asking a question about a perl cgi script? Come on. Reliable sources and self publishing aside, I think the changes to this article (and any future changes if the current trend is kept to) simply make this article far less useful to anyone wanting to know about Comet, which is what Wikipedia is for is it not? —Preceding unsigned comment added by Martintyler (talk • contribs) 12:57, 12 June 2008 (UTC)

I believe the usenet post is being used as an evidence that some term was used back in 1995, and not to endorse what the post's author says. I believe, in this case, this is not a problem of reliable sources, but more of using using primary sources. I'm not completely sure but I would say that it's ok, since the article only make descriptive claims about the information found in the primary source, and make no analytic, synthetic, interpretive, explanatory, or evaluative claims about the information found in the primary source. --Damiens.rf 13:58, 12 June 2008 (UTC)

Thanks for the explaination, but I cant say I understand really - what about the bluishcoder blog, I thought blogs were pretty much not allowed based on what's been said here? Anyway, I've been digging around to find 'reliable source' information, and found some old articles about my company (citynet, now caplin) - [14], [15] and [16] - I actually do not believe these are in any way more useful to someone reading the Comet article than the more recent self published work, but if those are the rules then these are reliable sources at least. —Preceding unsigned comment added by Martintyler (talk • contribs) 16:26, 12 June 2008 (UTC)

Do these articles really talk about comet? What excatly info from them do you think can be useful for our Comet article? (I'm not saying there isn't one (I haven't read them all). It's a sincere question) --Damiens.rf 16:48, 12 June 2008 (UTC)

They mention server push and http tunnelling. As I said, I dont think they are particularly useful compared to other articles that arent allowed, but in a similar way to the perl cgi post they demonstrate that comet was being used back in 1998 in a commercial environment (whereas the perl cgi thing is just a usenet question vaguely related to the subject) —Preceding unsigned comment added by Martintyler (talk • contribs) 16:59, 12 June 2008 (UTC)

The reason that I used the usenet posting for 'server push' was that I couldn't find the original Netscape documentation online anymore (this usenet posting refers too it). I could only find later dated documents on the Netscape site, perhaps if someone digs deeper we can find a primary source from 1995 or earlier. Maybe there is something in old Apache or CGI documentations, you could use a simple nph-CGI-script to deliver a streaming (long lasting) HTTP connection in the late 1990s. - 83.254.208.192 (talk) 17:11, 12 June 2008 (UTC)

Will the combatants please stand down!

I think it is time for those that have been butting heads on this page to step down and let others have a go.

Specifically Damiens.rf! You may think you are trying to be objective, but you are letting your POV intrude - if not in the text (nothing much of that left) then in your reasoning for your edits. Examples are many: Changing "used" to "created" because you think comet is barely used. You dismiss Alex Russell as a "non-notable tech-blogger", yet Alex is the creator of one of the most notable Ajax framework and I have seen him fill auditoriums with thousands of people interested to hear his opinion at several major conferences. You demand citations, yet dismiss those provided simply because the sources have collaborated. You've whittled away at the introduction that I provided, taking out word or a phrase at a time, until it was meaningless and then you delete saying it is "inessential". Well only because you've taken away all the meaning!

jacobolus, I think that you to should perhaps stand down as well, as you've gotten a little heated.

Following my own advice, I too have gotten a little hot under the collar at Damiens.rf's approach, so I'm going to back off as well! I've got better things to do with my time.—Preceding unsigned comment added by Gregwilkins (talk • contribs)

There's a minor implicit POV in "used". That the term "comet" was created, we have no doubt. But we should be careful when we say it's "used for this or that", since it seems very few people use this term for talking about the same concept. Of course, the fact that there's an perceived effort to push this meme makes things more delicate.

This Alex may have created notable things, but he doesn't necessarily becomes notable for that (at least not on Wikipedia's meaning of notability, that was what I was referring to).

Also, it was not me, but jacob that said that introduction piece was "inessential"[17].

About the "citations" provided, please red Restepc's reply to you on a thread above. He put it better than I could have.

But in general, thanks for the step on the right direction on this conversation. --Damiens.rf 11:58, 11 June 2008 (UTC)

Will the lobbyists please stand down! It will make this article much better if they give notable, verifiable resources or simply let neutral editors write this article. It does not help to complain LOUDER if complaining didn't help. The Comet meme is even misunderstood in the Ajax comunity, it is associated with "Ajax Push" but is there anyone who really has a definition of what Comet means? -- 213.115.160.102 (talk) 15:35, 11 June 2008 (UTC)

The “neutral editors” are getting their chance. So far no writing is happening. Yes, there is a definition of Comet. Look at the version of this page from before this brouhaha, and it is clearly defined and explained. —jacobolus (t) 20:39, 11 June 2008 (UTC)

The old version that you wrote has been described that it can not even explain what Comet is to readers familiar with main stream web development technologies, it lacked a definition and provided rather mystification that facts. - 213.115.160.72 (talk) 11:05, 12 June 2008 (UTC)

Editors - neutral or otherwise, should not be writing this article. Editors should edit and guide the contributors to improve the quality of their submissions.

The experts in this field are trying to provide content to this article. Perhaps they are violating some wikipedia fundamentals, but surely the editors should be giving them guidance and encouragement to improve the quality of the submissions.

Unfortunately the editors have revealed themselves to be non-nuetral. Their comments show that they are cometd skeptics and are making their POV rather than upholding the standards of wikipedia. These editors resort to name calling (lobbyist, un notable, zero credibility), and simply delete contributions rather than engaging in a process to improve them.

Even when it was accepted the build the page from scatch again idea, the editors have insisted in reverting every contribution to a simple "comet is just some name some guy though of for some technology that already existed" content.

I really think the rules are being over-vigorously applied by non nuetral editors who really want to push their own POV and are using rules as a defense to prevent the experts from contributing.

And yes I know you will say that the experts are self appointed,... but at least we have self published on the topic, there are several business owners whos company sell product described as comet, there are members of standards bodies, there are speakers at conferences etc etc. The contributors have some standing. The editors half the time can't even bother to login so we know who they are.

Is there some process to get a more senior editor involved? —Preceding unsigned comment added by Gregwilkins (talk • contribs) 00:41, 12 June 2008 (UTC)

Aren't you pushing your point of view? What makes you the neutral expert and gives you the authority to acuse others as not suited for reviewing this article? Your attitude stands against core policies of Wikipedia, reaching consensus and providing neutral facts. If people want to promote an idea, meme or concept they are simply at the wrong place. Here you have to give verifiable facts (read: this is mandantory by Wikipedia policies) and not complain that others will remove personal opinions and other material that is not suited for an encyclopedia. As long as consensus hasn't reached it isn't helpful to try to push trough one or the other opinion. In this situation, I liked the idea of building up this article from scratch. - 213.115.160.72 (talk) 11:44, 12 June 2008 (UTC)

Reply to Gregwilkins points: I don't quite understand the distinction you make between "editors" and "contributors". What exactly do you mean by each of these? (Aren't Wikipedia's editors its contributors?)

Here's guidance for "the experts in this field"": Read and understand the following core principles:

From my side, I've read a lot about comet.

>"...These editors resort to name calling (lobbyist, un notable, zero credibility)..."

Tell me about. I've been called a "vandal"[18] and a "jerk"[19] by one of the "experts in this field"".

> "...even when it was accepted the build the page from scatch again idea..."

It was not quite "accepted". The old author insists in calling this "destruction"[20] or "vadalism"[21] (and other strong synonyms).

> "...the editors have insisted in reverting every contribution to a simple "comet is just some name some guy though of for some technology that already existed" content..."

The whole idea of scratching the article was that we could restart from beginning and include only verifiable/attributable information. So far, that's what we have (either you like it or not).

> "...the contributors have some standing..."

Credentials have no value on Wikipedia. You should be prepared to attribute all material you submit to a reliable, published source if asked to do so.

> "...The editors half the time can't even bother to login..."

Just like some of the "experts in this field" [22][23]. But that's no problem. Anonymous editors are welcome on Wikipedia, and their edits are treated under the same standards than those by logged in editors.

> "Is there some process to get a more senior editor involved?"

Wikipedia doesn't quite have the a concept similar to "a senior editor". Every editor is treated the same (even the anonymous editors, as said above). We do have the concept of "Administrator", that are trusted members of community that can do the technical taks of enforcing policies.

If you want an administrator to take a look at this case, you may ask for help at Wikipedia:Administrators' noticeboard/Incidents.

But if you simply want more editors to take part in the discussion, you may try opening a Wikipedia:Requests for comment.

If nothing works, you may try a Wikipedia:Request for Arbitration.

Thanks, --Damiens.rf 13:49, 12 June 2008 (UTC)

I called you a “vandal” because you insisted on disruptive and destructive edit-warring repeatedly while being pointedly asked to take it to the talk page. I called you a “jerk” after you made several unnecessary insulting and inflammatory statements and edit summaries, intentionally designed to rile me up. These descriptions were not intended as name calling (indeed, I would be surprised if you didn’t yourself find the latter to be accurate); please consider them notices that you should avoid the excessive incivility you have so far brought to this article, and instead work towards constructive consensus. Yes, a requests for comment should be opened at this point. When I have more time, I will open one. For now, I will merely shake my head at the pitiful inaccurate two-sentence stub this page has devolved into. —jacobolus (t) 20:08, 13 June 2008 (UTC)

Damien.rf - Just because some have called you names, does not make it OK for you to name call back! Gregwilkins (talk) 15:31, 14 June 2008 (UTC)

And I haven't done that. Please, be careful not to formulate advices in a way that gives the impression of an implicit accusation. --Damiens.rf 19:59, 14 June 2008 (UTC)

Sorry if I have used the non-wikipeadia term of "contributor" and "editor". What I meant by them, is that contributors are actively trying to add text to explain the term comet, while those acting in as editors are trying to correct those contributions with regards to the rules of wikipedia. OK - so everybody is an editor. None the less, some here are trying to be constructive and add explanitive text (with citations) to this page. Some other appear to be comet sceptics (in their comments at least) and are using over-rigorous application of the rules of wikipedia to remove explanative text and not replace it. Gregwilkins (talk) 15:31, 14 June 2008 (UTC)

We did start building this page again from scratch. But the comet sceptics kept pushing their POV that this page was nothing more than advertising for an phrase that was hardly used and have kept cutting it back to nothing. Gregwilkins (talk) 15:31, 14 June 2008 (UTC)

OK, I get it... experts have no standing here... only citations. Well citations have been provided, but that still does not appear sufficient. The POV is being pushed that "comet" is a term being pushed for promoting a meme, advertising or google rank purposes by a band of collaborators. Well I would ask you to cite your evidence for that! Saying that the many of the editors all contribute to cometdaily does not indicate a conspiricy any more than the fact that we all contribute to wikipedia Gregwilkins (talk) 15:31, 14 June 2008 (UTC)

The contributions to this page have been accompanied with citations for articles, products, specifications, open source project and companies that self describe using the term comet. While not all of these may have been up to the obviously high standards of wikipeadia, there are now certainly enough of them to say that it is FACT that there are products/projects/companies/spefications using the term "comet" to communicate their function. The contributions to this page have not been to promote that as an idea, meme or concept. They have been to explain what it means when something self-describes as using comet techniques. As far as I can tell, none of the contributions have tried to push through "one or the other opinion". If fact the opposite is true, in that the deleted content was trying to describe all the different flavours of comet that exist. If you read the cometdaily site, you will see that there is heated debate about the nature and best practises for comet. This page should not try to pick a winner, nor should it wait for 100% agreement between all parties that use the term comet - that will never happen. Sure this page should explain with citations how this phrase is being used and if there are any contradiction or ambiguities, then it should point those out. Gregwilkins (talk) 15:31, 14 June 2008 (UTC)

Comet implementations

The link to comet implementations was really just a link to cometd implementations (the d is significant). Here is an initial list of companies/products/projects that self describe as comet. I think this should be on the main page, but not sure where?

http://cometd.org/
http://lightstreamer.com/
http://webtide.com/
http://www.freeliberator.com/
http://code.google.com/p/erlycomet/
http://java.sun.com/javaee/community/glassfish/
http://code.google.com/p/pi-js/
http://tomcat.apache.org/tomcat-6.0-doc/aio.html
http://www.kaazing.com/ —Preceding unsigned comment added by Gregwilkins (talk • contribs) 01:49, 12 June 2008 (UTC)

It shouldn't be there at all. Wikipedia is not a link directory, please don't do original research this includes talk pages. For more information see WP:MOS. - 213.115.160.72 (talk) 11:20, 12 June 2008 (UTC)

Wikipedia isn't a link directory, but a few examples of notable things that use comet (and more importantly say themselves that they use comet (or streaming ajax or whatever they call it that we're all agreed is the same thing)) in the article would be useful. Restepc (talk) 09:36, 13 June 2008 (UTC)

I suggest that entries need to have an existing Wikipedia article. Some projects describe themself as "Comet-like" (does that qualify as Comet or not, do they simply mean "Ajax push"?), perhaps we should wait until a definition was agreed on to be able to verify that everyone speaks about the same. - 83.254.208.192 (talk) 12:15, 13 June 2008 (UTC)

Yes, they almost assuredly qualify as Comet (though as discussed above I think these links are somewhat problematic). Some developers have come to the conclusion that “Comet” is limited to a particular transport. “Ajax push” and “Comet” are synonymous in this instance. The definition as this Wikipedia article stated it a few weeks ago predominates, but I'm not sure what exactly you would like to verify. Can you better explain your confusion? —jacobolus (t) 01:57, 14 June 2008 (UTC)

Can you please calm down, it is getting harder to talk to you. The threshold for inclusion in Wikipedia is verifiability. When developers use a vague description like "Ajax and Comet-like" the question arises if they know what Comet means, therefore if the product uses Comet (or just some sort of server push). For Wikipedia editors this means we need an agreed and comprehensive definition of Comet. - 83.254.208.192 (talk) 06:58, 14 June 2008 (UTC)

I am perfectly calm, and I am not seeing the difficulty of discussion you reference. It is indeed unclear whether some of those developers have a precise idea what Comet means, but that doesn’t mean that their products do not “use Comet”. Indeed “just some sort of server push” sent to a javascript callback in the browser is the same thing as “Comet”. The definition is widely agreed upon, even if still murky for some minority. I’m not sure what you mean by comprehensive, and I still don’t understand your general confusion about the definition. Can you further clarify? —jacobolus (t) 08:10, 14 June 2008 (UTC)

Not sure if everyone agrees with you. Some people think of Comet as an Ajax design pattern, for you everything that pushes or long-polls JavaScript is Comet (strange that you accept pushed XMLHttpRequest but tried to remove JSONRequest Duplex), for others JavaScript is not even needed (Pushlets) and it describes EVERY non-static web development. Do you see the problem for Wikipedia editors of not having a clear comprehensive definition when categorising software? - 83.254.208.192 (talk) 09:01, 14 June 2008 (UTC)

The use of JSONRequest to send real-time updates would be “Comet”. JSONRequest itself is merely a proposed but unlikely-to-ever-be-used-in-the-real-world Comet transport. Most uses of Pushlets, as far as I know, are “Comet”, and I have seen the general idea of sending real-time events to browsers referred to as “Pushlets”, in sources written before there was a more widely accepted name. I'm not sure what you mean by “every non-static web development”. Those who call Comet “an Ajax design pattern” are not using the normally understood technical definition of “design pattern” for such usage. —jacobolus (t) 11:24, 14 June 2008 (UTC)

That's funny, describing JSONRequest and Ajax as "Comet transports". Perhaps all web technologies are "Comet core technologies" and HTTP the primary "Comet transfer protocol". :) However, you didn't answer my question, how can editors categorise implementations (Comet software) without having a definition of the concept? - 83.254.208.192 (talk) 15:02, 14 June 2008 (UTC)

Will the anonymous editors please login and sign their posts!!! (pot calling kettle black) Gregwilkins (talk) 14:58, 14 June 2008 (UTC)

Nobody is asking wikipedia editors to go around "categorising software". The products and projects cited all self classify as comet. So when a reader finds JSR315 saying that it addresses the comet use-case or the lightstream product saying it is a comet server or jquery providing a comet client, some of them might like to turn to wikipeadia to find out what the term means. The self describers obviously believe that term contains enough meaning that further explanation is not required! OK, so the self describers are not 100% consistent in their use of the term comet and there is some degree of ambiguity. But that is exactly why it would be valuable for wikipeadia to have a longer article that describes (with citation) the different uses of the term comet and the different technologies to which the phrase has been applied. I can't believe that you expect 100% external agreement on the meaning of a term before you will allow any detail to be included on wikipedia. Surely if there are conflicting citations, then both should be listed with text saying that they are in conflict? Gregwilkins (talk) 14:58, 14 June 2008 (UTC)

You can't win! First you demand citations, and when they are provided you say "wikipedia is not a link directory". I give up - Bye Gregwilkins (talk) 15:47, 14 June 2008 (UTC)

Maybe you should try not seeing discussions about how to improve the Encyclopedia as something one can win or loose. --Damiens.rf 20:25, 14 June 2008 (UTC)

JSONRequest is a “Comet transport”. “Ajax” is an application architecture (just as “Comet”, but “Ajax” has a looser definition than “Comet” does). HTTP is indeed a “protocol”, but I’m not sure what that has to do with anything. The definition of “Comet” is rather simple, and I don’t understand your difficulty with it. Maybe try emailing Alex Russell and he can try to explain it to you himself? In any case, what does any of this have to do with categorizing implementations? (And note, I have not suggested that wikipedia editors should categorize implementations). —jacobolus (t) 23:41, 14 June 2008 (UTC)

(undent)"The self describers obviously believe that term contains enough meaning that further explanation is not required! OK, so the self describers are not 100% consistent in their use of the term comet and there is some degree of ambiguity. But that is exactly why it would be valuable for wikipeadia to have a longer article that describes (with citation) the different uses of the term comet and the different technologies to which the phrase has been applied. I can't believe that you expect 100% external agreement on the meaning of a term before you will allow any detail to be included on wikipedia. Surely if there are conflicting citations, then both should be listed with text saying that they are in conflict?" Yep; I agree with that and that's what the article will be (and is) working towards. I left you a message on your talk page btw Greg. Restepc (talk) 16:01, 14 June 2008 (UTC)

Who are these “self-describers”? What is the definitional problem? What are these different technologies? Comet can use a number of possible “transports”, which were incredibly well described by the previous article. —jacobolus (t) 23:41, 14 June 2008 (UTC)

Image

I added a picture of a server to the article, as an image always makes things look better IMO. It wasn't perfect for the topic though and has been removed, does anyone know of a more appropriate image we could use? Restepc (talk) 09:19, 13 June 2008 (UTC)

I would be happy to add images, but it is not worth it given the current state of this article —jacobolus (t) 01:57, 14 June 2008 (UTC)

Sounds like a good idea, a picture says more than a thousand words. - 83.254.208.192 (talk) 09:28, 14 June 2008 (UTC)

What happened to all the content?

A few weeks ago, this article was full of content, and now it's all gone. I understand that a lot of it was unsourced, original research, but there was a lot of good, valid information in there, such as the HTML 5 proposal, that got removed. I haven't been watching this article for long and I'm not about to read every discussion on this talk page, but I see no reason for removing half the stuff that was there before. Who wants to help me build it back up? — ^Fatal Error 01:21, 15 June 2008 (UTC)

The attitude of some editors and the scale/number of problems with the old version made me think that essentially rewriting the article from scratch would be better than trying to 'fix' the old version of the article. There was some useful information in the old version, and I've been meaning to go back and recover it...but simply haven't had much time to spend on wikipedia, so the process of building the page has been incredibly slow...although a large part of that could be the problems with defining exactly what comet is. So Yep; the article needs work...I'll be doing what I can when I get the time. Any help is greatly appreciated. Restepc (talk) 19:15, 15 June 2008 (UTC)

Okay, well I was going to restore some information but I didn't know if there was a good reason behind removing it other than rewriting it. I'll see what I can do. — ^Fatal Error 19:33, 15 June 2008 (UTC)

The much people working here the better. Be welcome. The last version before the restart was this one. Feel free to work any usable information from there. But of course, be sure not to re-add the unsourced, badly sourced, and original research pieces.

And since we're at it, why did you removed "Pushlets" and "JSON Request" from the list of other names to the same concept in the article's current version? --Damiens.rf 22:28, 15 June 2008 (UTC)

The Definition of Comet?

The current opening paragraph of this page (on 15 June) is close to a pretty good definition of what comet is. But I have two niggles with the definition. Firstly, is the "long-lived HTTP connections" phrase. I know this phrase is used in the cited articles for the definition, but this really relates to specific transports (used at the time) and is also confusing as it is really long held HTTP requests that these transport rely on. Also there are non HTTP implmentations that use a pure TCP/IP socket available from the flash plugin as a transport. The question I have about these, is that the use of flash plugin may render these transports non-Ajax?

This brings me to my second niggle: the definition does not make it clear that comet is a subset of Ajax. But while there are many article that do describe Comet as an Ajax technique, the original coining of the term only says that it is Ajax-like. So perhaps the use of a flash-plugin is OK - in which case non-HTTP transports are probably OK.

So I think both these niggles go to the heart of the current lack of an agreed definition for the word comet. Specifically small disconnects between the original coining of the term and what I believe to be it's current usage. However, as wikipedia is not the place for original thought, I guess we cannot resolve this disconnect here.

Thus I'd like to propose to slightly defocus the current definition, so that it does not lean one way or the other with these niggles. I'll update the definition accordingly, but I'm not 100% committed these changes and will take no great offense if they are backed out. I'll start some efforts elsewhere to resolve these questions, so the definition can be tighted up here. Gregwilkins (talk) 00:45, 16 June 2008 (UTC)

What Just Happened Here?

This revsision covered the technology perfectly. Why was it deleted? --Armin Ronacher (talk) 09:42, 16 June 2008 (UTC)

Fortunately someone reverted the article again. --Armin Ronacher (talk) 09:43, 16 June 2008 (UTC)

The content was deemed inappropriate (WP:NOR, WP:RS, WP:V, WP:NPOV, WP:COI) and broken beyond repair. Restarting the article was the best choice.

By the way, if you allow me a barely-related question, how did you know about this article? In the last 6 hours there was a small surge of inactive accounts to this talk page[24] and article[25] [26] [27] [28] [29], together with a surge of anonymous vandalism in user pages of users working in this article's rewriting[30] [31]. Maybe just a big coincidence. --Damiens.rf 12:32, 16 June 2008 (UTC)

It's linked on Reddit

http://www.reddit.com/goto?rss=true&id=t3_6njht

--.../Nemo (talk) 12:35, 16 June 2008 (UTC)

Oops, you linked to the story's out-going url, Nemo. The discussion on reddit is here:

http://www.reddit.com/r/programming/info/6njht/comments/

--Ywg.dana —Preceding unsigned comment added by Ywg.dana (talk • contribs) 13:10, 16 June 2008 (UTC)

I'm no contributor but that doesn't stop me from having a very bad feeling if a good article is started over. I actually read the original version of the article (which is now restored) a week ago or so, shortly before it was deleted. --Armin Ronacher (talk) 15:13, 16 June 2008 (UTC)

The wiping of the article is inappropriate. Damiens.rf, rewrite it yourself if you want to start over, but don't just delete a large body of useful content. --Tim Jun. 16, 13:29:25 (UTC)

Agreed. The only "consensus" I see is Damiens.rf and restepc against all the subject matter experts. Now that others are paying attention, you two will have to be specific in your criticism, or be constructive in your edits, rather than destructive. And stop marking as "vandalism" the edits of those who are clearly trying to improve the article over your personal preferences. Ken Hirsch (talk) 14:24, 16 June 2008 (UTC)

"be specific in your criticism..." Have you actually read the discussions? Is this specific enough for you: #Problems_with_the_History_section, #Scalability, #.22Implementations.22_section, #Browser_hacks, #Use_of_Notes, #Conflict_of_interests.

Ignoring the discussion and the edit summaries to just revert to a deemed version defended by of-site rants sounds as vandalism to me.

Thanks, --Damiens.rf 14:50, 16 June 2008 (UTC)

Some of those criticisms are good points. Why don't you do something about them rather than delete the whole thing. Is is laziness or vindictiveness? I fail to see the relevance of the off-site "rant". Most of us have a life, therefore we cannot monitor every WP page full-time. When we are alerted to a useful page than has been rendered useless, we should take action. —Preceding unsigned comment added by Ken Hirsch (talk • contribs) 15:02, 16 June 2008 (UTC)

Stop this madness!

I'm getting pretty tired of this "if it's not sourced it's not true" kind of attitude. There was a lot of good information in that article. Rather than scrapping it and replacing it with what is essentially a stub why not go back and try to source areas without citation? Bactoid (talk) 14:23, 16 June 2008 (UTC)

The threshold is not truth, but verifiability. And lack of reliable sources was just one out of the many problems this article had. --Damiens.rf 14:51, 16 June 2008 (UTC)

Don't bite. WP:V isn't going anywhere, but many of the newcomers from Reddit will go. And Bactoid has a point: excision is the last line of defense for unverified technical content in a non-BLP article. This is only a controversy because someone amputated instead of treating the wound. --- tqbf 15:50, 16 June 2008 (UTC)

The threshold is verifiability -- that is, "is able to be verified", not "has already been verified and cited". 68.42.74.83 (talk) 22:43, 19 June 2008 (UTC)

Ad/PR-Hits and too much technical detail

There's a little too much detail at some points that would be better off just being linked to in a one-liner. For example, the web architecture comparison. It seems better to just mention the other architectures and focus on what Comet is, not what it isn't. Also, since Comet Daily was created to promote Comet, it may be difficult to include quotes from them and we should get rid of anything that smells like an ad. --OMouse (talk) 14:11, 16 June 2008 (UTC)

ATTN REDDIT PEOPLE

The old version of the article which many of you are reverting to was not removed because of overly technical details but because it was almost all Original Research, written by one editor to advertise his website, including adding quotes from himself and his colleagues. There was originally an attempt to address specific problems/sections in the article as some of you are suggesting, but this was met by bulk reverting and edit warring.

The current 'consensus' version is a work in progress and any help improving it is appreciated, but please do not revert to cometdailys version. Restepc (talk) 15:58, 16 June 2008 (UTC)

I see the article is now protected thankfully....on the advert version for now, but oh well. Restepc (talk) 16:00, 16 June 2008 (UTC)

A further comment, if content previously ear-marked as Original Research, can be supported by verifiable reliable sources- please source and include. AtaruMoroboshi (talk) 16:04, 16 June 2008 (UTC)

I assume someone is writing a RFC as I type this (or someone will be soon at any rate), I'll wait until it's up and running and work on a well thought out description of events from my view when it is. strikeout, see discussion below. I suggest the page protection is left for now, even though it's protected on what I think is the 'wrong' version, to prevent further edit warring and let things calm down....perhaps the protection should remain until after the RFC is finished. I also advise against admins being over zealous with blocks, I believe the majority of these reddit people are acting in good faith, if misguidedly. Restepc (talk) 16:20, 16 June 2008 (UTC)

RFC

I believe an RFC is the way to go now. Who will be so king as starting it? Should we wait for the REDDIT people go away before? --Damiens.rf 16:12, 16 June 2008 (UTC)

I wouldn't necessarily discredit edits coming from Reddit users. It is a technology community, and the article could see overall improvements as a result. Cheers, AtaruMoroboshi (talk) 16:20, 16 June 2008 (UTC)

you said yourself, the more the merrier....let's see what they can offer and avoid throwing the baby out with the bath water. As far as RFC, I had assumed that you'd be rushing to write one already, but if not....I am willing to...with respect...I think it would perhaps be better if either myself or IP guy started it instead of you(Damiens) or Jacob. No offense intended. I'll start drafting one now, but if anyone else has a particular desire to start it instead I don't mind yielding. Restepc (talk) 16:25, 16 June 2008 (UTC)

I agree with AtaruMoroboshi that it might actually be helpful if Reddit users choose to comment on the RFC. I'll write a notice for the top of this page describing what's going on for the benefit of any visitors. I'm happy to help write the RFC if needed, but I'm sure Restepc has it well in hand. Canderson7 ^(talk) 16:29, 16 June 2008 (UTC)

Restepc, please by all means begin the RFC. Offense not taken. AtaruMoroboshi (talk) 16:31, 16 June 2008 (UTC)

Yes. Restepc is by far the more cool-headed here, and I one of the few capable of writing an neutral description of the situation. Please, start it, and I will help with my views as appropriate. --Damiens.rf 16:33, 16 June 2008 (UTC)

For someone who labeled as pompous a definition of Comet as an 'Ajax-based web application architecture', you sure do have a high opinion of your own neutrality...63.64.108.5 (talk) 18:41, 16 June 2008 (UTC)

Can this RFC stick to the same rules that have been recently enforced on this article? ie stick to the facts, rather than conjecture on jacobs motives and conspiracy theories about cometdaily please. Martintyler (talk) 16:47, 16 June 2008 (UTC)

~~Canderson, I'm not sure how neutral my statement is, especially 'discuss allegations of canvassing', it's just the best I could think of. If you'd like to modify it feel free.~~ On second thought the two issues should be kept seperate, I don't want my annoyance at the canvassing to interfere with the discussion of which version of the article to use.

Martin, it is likely that people will give their opinons on the RFC as well as facts, this probably includes Damiens speculation that Jacob is/was using the article to advertise cometdaily (although you'd have to ask him or wait and see to be sure). I'm sure people are capable of telling the difference between opinion and facts though. Restepc (talk) 17:03, 16 June 2008 (UTC)

Thanks, I am not familiar with the process, I got the impression from the comments above that some kind of initial statement is to be written, and that is what i was refering to. Subsequent comments are obviously going to be more opinionated. Martintyler (talk) 17:20, 16 June 2008 (UTC)

I'm not overly familiar with it either, the initial statement I've chosen can be seen below. "To attempt to resolve the dispute over which version of the article to use, and discuss allegations of canvassing". Restepc (talk) 17:24, 16 June 2008 (UTC)

I added a somewhat more verbose summary below yours; hopefully it will give a newcomer to the discussion some sort of foothold to get himself started. If any aspect of it is objectionable, however, feel free to edit away. The main parties in this dispute should now go ahead and state their grievances / reccomendations. Canderson7 ^(talk) 17:44, 16 June 2008 (UTC)

I've also added a box at the top of this page that I think will provide some guidance to any visitors. Everyone should feel free to edit it. Canderson7 ^(talk) 17:48, 16 June 2008 (UTC)

Article Fully Protected

Toddst1 and I have both fully protected this article. We did so following discussion on the Administrators' Noticeboard, and the article should remain protected until community discussion can resolve the content dispute. It happens that the longer version of the article is the one that we protected, this is by pure happenstance and does not reflect our opinion one way or the other. Canderson7 ^(talk) 16:12, 16 June 2008 (UTC)

Under the circumstances I strongly support your actions. Restepc (talk) 16:21, 16 June 2008 (UTC)

Thanks. Canderson7 ^(talk) 16:26, 16 June 2008 (UTC)

Thanks for the so-much-needed intervention. Please, stick with us since we'll benefit from some admin oversight on the upcoming rfc. --Damiens.rf 16:35, 16 June 2008 (UTC)

Sure thing. I'll be here. Canderson7 ^(talk) 16:37, 16 June 2008 (UTC)

Request For Comment: Article version dispute.

To attempt to resolve the dispute over which version of the article to use.

The two versions in question are Version 1 and Version 2. Version 1 was in effect until a couple weeks ago, when concerns began to mount that it violated certain Wikipedia policies and guidelines. The article was then stubbified, and Version 2 produced over the course of the next several days. Recent publicity has renewed the dispute, and the opinion of neutral third parties would be welcome: Should Version 1 be used? If so, what editing (if any) is required to bring it within Wikipedia policies and guidelines? Is it easier to just use Version 2? If so, what improvements (if any) are needed? 17:41, 16 June 2008 (UTC)

Past discussion threads (please expand this list): [32], [33], [34], [35], [36]

Having read both, I think Version 1, with judicious pruning and editing, is vastly preferable because it actually explains what Comet is by mentioning the specific techniques and giving a bit of context as to when, how and why it developed. Much concern seems to have been expended not on the sections of the article which actually describe what Comet is and the various techniques, but rather on the sections which explain how it differs from the traditional pull-based model of the Web; the ideal would probably be to start from Version 1 and then pare down those sections to an introductory paragraph or two summarizing the difference between push and pull models. Ubernostrum (talk) 17:59, 16 June 2008 (UTC)

I agree with Ubernostrum. Though anonymous here, I'm user 'dhogarty' at reddit, and am very familiar with the field under question. The information in the first version needs to be retained, with pruning and editing 63.64.108.5 (talk) 19:13, 16 June 2008 (UTC)

Looks like a conflict between two untenable extremes. We should refactor from the larger version, instead of tossing good content and rebuilding it.

Pull quotes --- WP:MOS, and probably the reason this page got flagged as promo content. These obviously need to go.

Web app architectures --- Strike from article as not directly relevant to COMET and per WP:SIZE; this content belongs in its own article, not in a little island of "web architecture as seen by COMET enthusiasts" here.

Browser compatability --- Good core content. Should never have been struck from the article. Browser support for persistant connections is the core technology behind COMET and server push.

Scaleability --- As with "Web app architecture", too much general content. Vertical scalability and Horizontal scalability deserve their own articles. An article on COMET can't be wasting time and focus explaining Apache's threading model.

Publish-subscribe --- not at all clear how this is germane to the topic of COMET, as any web app architecture can be built on a messaging layer (plenty of second-gen WebSphere apps are built on MQ). --- a well-known app that used XMPP with Comet would mitigate. Maybe cut this down to one sentence.

History --- Way too much editorializing. This isn't an article about Java applet UI. One graf ends "This can't really be called COMET" --- that's a symptom of bad writing. Things like "Splash in the Silicon Valley Imagination" don't belong in a WP article. Just the facts, please.

Footnotes --- Too many. Work into the article text or excise.

--- tqbf 18:01, 16 June 2008 (UTC)

I concur with above two comments. Version 1 (with a lot of judicious pruning) is preferred. (The quotes definitely should go.) Version 2 doesn't actually describe what comet is. --lk (talk) 18:38, 16 June 2008 (UTC)

tqbf, the "Browser compatability"" section was struck from the article because it's original research. When references are used, they are from self-published tech blogs, including an "article" by Jacob himself. --Damiens.rf 19:24, 16 June 2008 (UTC)

Damiens, having a hard time believing that everything in this section is OR, since much of it is common knowledge to HTTP devs. Maybe you could propose specific rewords for things that are contentious? The fact that progress bars remain active on invisible IFRAMEs is true, very relevant to server-push web apps, and an example of the kind of thing that could reasonably be sourced back to the COMET team. --- tqbf 19:46, 16 June 2008 (UTC)

The section at least feels like original research for those that do not share the "common knowledge" of HTTP devs, since it lacks good sources. I see no benefit (at least not for Wikipedia) in using cometdaily.com as a source for what would be "common knowledge to HTTP devs". If it's so common, it shouldn't be hard to find reliable independent published sources (eg: books) supporting the passages.

Anyway some passages really seem like the authors musings, like "...application developers mostly use..." or "...main benefit of the IFrame method is...".

And there is a general pro-comet tone in that it gives me the impression that everything web-developers have being doing in the last years were failed attempts towards this new "Comet" future: "...until recently, browsers were not designed with Comet in mind..." - They still aren't, and probably will never be. Maybe the author meant something like "The web was not designed for push programming". More comet-praising passages is the characterization age-old IFRAME hacks as Comet, the (more grave) caracterization of HTML5 server-sent events as "...an attempt to avoid the disadvantages of other Comet transports" or the mention of "Comet developers" when it probably means "web developers".

This section, as much of the article, gives the false impression that "comet" is more than it is, by calling a lot of related concepts "comet" or "almost-comet". --Damiens.rf 20:12, 16 June 2008 (UTC)

Comment --- I'm confident I can find sourcing for the technical points made in the "Browser Compatibility" section. However: (1) I am not a COMET advocate (I've never contributed to the project or used the technology in a product), and (2) I am concerned that based on User:Damiens.rf's strong position on COMET, that work would be wasted. I can see why this article hit RFC. There are some strong personalities here. --- tqbf 20:35, 16 June 2008 (UTC)

FWIW, Opera already [supports] [Server-sent DOM Events], and I believe support from other vendors is in the pipeline (Huge disclaimer, I wrote the article I'm pointing to, and I work for Opera, so I'm not going to be editing). Further, browsers already mostly support some of the patterns used in Comet, as a side-effect of supporting HTTP. Arve (talk) 22:46, 16 June 2008 (UTC)

Do Operra/Mozilla/IE/Safari call it "Comet", or just "Server-sent DOM Events"? My concern is exactly with this appropriation of a general concept by Comet evangelism in the article. --Damiens.rf 23:20, 16 June 2008 (UTC)

"Server-sent DOM Events" constitute only a very particular implementation of how to do Comet, with other techniques such as injecting script tags being another. I guess what you are really asking is "Is Comet an accepted industry term"? Yes, to the point where the OpenAjax alliance, with prominent members such as IBM, Google, Yahoo, Oracle and others have developed the OpenAjax Hub, claiming [comet support] Arve (talk) 23:44, 16 June 2008 (UTC)

To elaborate further, Comet is a set of design patterns where you can give the appearance of pushing data or events to the client, by means of a long-lived HTTP request in some form. In this vein, it should be viewed in a similar manner as "Ajax", which can be implemented with other methods than XMLHttpRequest, such as form submission and document retrieval through an invisible frame/iframe. —Preceding unsigned comment added by ArveBersvendsen (talk • contribs) 23:49, 16 June 2008 (UTC)

Good. Independently published mentions of "comet" is exactly what this article is in desperate need for. --Damiens.rf 23:55, 16 June 2008 (UTC)

Restepc's view (and comments)

The below is intended to give a description of events from my point of view, it contains some speculation on my part, and my opinions on what has happened as well as simple facts, feel free to mentally insert the phrases 'I think' and 'In my view/opinion' as necessary.

I have the COI noticeboard on my watchlist, and responded to this post, which now includes some further discussion.

I observed on the article a minor edit war between Damiens, Jacob and an anonymous user I've taken to calling IP Guy.

It appeared that Damiens had started to edit 'Jacobs version' after being of the opinion that it needed changing considerably for a wide variety of reasons. Jacob bulk reverted his edits and the edit war began, IP guy joined in on Jacobs side on the basis that Damiens wasn't engaging in discussion on the talk page and Jacob was repeatedly inviting him to.

When I came to the article I attempted to stop the edit war and get Damiens onto the talk page. Myself and IP guy quickly came to the conclusion that Damiens was more than justified in his edits, and Jacobs version was severely unsuitable for wikipedia. It also became clear that Jacob did indeed have huge COI/Vanity problems with the article.

After some discussion I once again changed the article to 'Damiens version', which although not perfect I feel was a massive improvement.

Jacob started up the edit war again, so I left it and suggested that it may be best to simply stub the article and rewrite it from scratch. After a few days of discussion (plagued by incivility, some from Damiens but mostly from Jacob) a consensus of 3v1 in favour of my suggestion was reached, and I personally was (and still am tbh) of the opinion that Jacobs opinion should be ignored due to his severe COI problems. Although 3 people is a small consensus, the only opposing view was that of Jacob, who was arguing that his version should be left as it was, which I'm almost certain every experienced neutral editor will agree is not an option.

So on the back of those discussions I stubbed the article with a basic definition (which I got dead wrong knowing nothing about the subject myself), with the intention of going back to Jacobs version and salvaging what useful information/sources there was. Four things happened that stopped the article making much progress at all: Jacobs attitude became even less civil/more disruptive, it emerged that there is no universally agreed definition of comet in the first place, I didn't actually get the time to salvage anything, and a handful of other people from cometdaily turned up on the article and enlongated/complicated all discussions/edits. Jacob has been edit warring throughout.

I would like to note at this point that all the 'comet daily' people with the exception of Jacob have acted in good faith and Greg in particular has been helpful. The article appeared to be slowly beginning to get somewhere when someone canvassed reddit and an edit war erupted between the new consensus version and Jacobs version. Mine and Damiens user pages were also vandalised.

I think there are three options to be discussed, in all cases the version of the article that should be used is simply a starting point, they all need a lot of work.

1 Start working from 'Jacobs version' 2 Start working from 'Damiens version' 3 Start working from 'consensus/stub version'

I am also very concerned by the canvassing which all common sense suggests has taken place, but would suggest keeping that issue and the issue of how to proceed on the article seperate. Restepc (talk) 18:23, 16 June 2008 (UTC)

Those are, indeed, the options. Now, (1) which of them do you favor, and (2) what steps do you propose we take to get from that option to "good article candidate"? --- tqbf 18:41, 16 June 2008 (UTC)

An important issue to resolve is how to treat emerging topics such as this. Increasingly in the software field, and especially in smaller, more bleeding edge communities, the authoritative sources are online, and published works that meet all of wp's guidelines regarding self publishing lag two or three years behind. I think the major issue here was not one of self-published sources, but instead the conflict of interest issue. CometDaily is a valid tech source, particularly for this kind of resource. If a non-CometDaily editor had posted most of the information, and then not bulk reverted there would be far less of an issue. Surely there must be a case for allowing self-published sources when they ARE the most authoritative? --- 63.64.108.5 (talk) 18:52, 16 June 2008 (UTC)

Comet Daily is a Self-Published Source. WP doesn't "lag" on this issue; WP:SPS issues come up constantly. Limited citation from a popular, highly relevant SPS seems reasonable. But accepting that SPS as a blanket valid source and using it to wholesale lift promo copy into a WP article is obviously not reasonable. It is unlikely that this RFC is going to end with a consensus that "if it appears on Comet Daily, it's fair game". --- tqbf 18:59, 16 June 2008 (UTC)

I agree, no reason to accept CometDaily as a 'blanket valid source', what I mean is, to find the right way forward, we need to define what kind of limited citation is appropriate. I read the SPS guidelines, and it seems that the relevant guidlines are 1) Self-published sources may only be used as sources about themselves 2) The material used is relevant to their notability 3) it is not contentious. The information regarding 'Comet' is largely found in such sources, and so perhaps a reasonable approach to this article would be detailing the specific definitions of various writers that self identify as 'Comet' developers (e.g. Alex, Jacob, Greg, etc.). 63.64.108.5 (talk) 19:10, 16 June 2008 (UTC)

I add the obvious clarification that the COMET developers themselves can never write directly in this article and cite themselves and their status as a source --- that's the definition of WP:OR. I may have misread you, though, so no response necessary. --- tqbf 19:17, 16 June 2008 (UTC)

I propose option 2. Revision 1 is damaged beyond repair, so I don't think we need to even look at it. Damien's revision has a lot of good information in it, and I'd say about 2/3 of the article could stay as-is (if we can find more sources to prove them). Most of the stuff that doesn't belong can simply be deleted; there isn't much that needs to be rewritten in that version. My two cents. — ^Fatal Error 18:55, 16 June 2008 (UTC)

It would be helpful if you could enumerate the damage in Jacobs version, so we don't have to guess as to what your issues are from a diff. --- tqbf 19:00, 16 June 2008 (UTC)

Comment can you expand on what you qualify as "damaged beyond repair"? The version that "has a lot of good information in it" is a stub comprised of four sentences. Is that really what you intended to state? I want to make sure I'm understanding you statement correctly. Thanks, AtaruMoroboshi (talk) 19:04, 16 June 2008 (UTC)

I once wrote a detailed line-by-line analysis of the History section, explaining the many problems that permeate the article. That was a reply to a request by Jacob, but it was ignored. You may still read it above at #Problems_with_the_History_section. --Damiens.rf 19:16, 16 June 2008 (UTC)

Strong agree with everything in Damiens.rf's critique of the "History" section of this article. --- tqbf 19:18, 16 June 2008 (UTC)

Agree much of the History section - is synthesis or not of real value to the article. Also agree with Damien's statement "A history section (of such an article) should contain some background on what led to the technology creation and then go on to explain the relevant facts since it was established - However, this does not constitute "damaged beyond repair." The baby should not be thrown out with the bathwater, and I'd support building from version 1 beginning with a remove of much of the editorializing, and synthesis, and CN tagging (but not removing) unsourced statements. AtaruMoroboshi (talk) 19:29, 16 June 2008 (UTC)

I also agree with much of the criticism of the History section, though I also disagree with much of it. Damiens’s line-by-line analysis mixes alternating excellent points with frivolous hyperbole. I will try to respond in more depth sometime tomorrow or in the next few days. Most of his conclusions boil down to concerns about inadequate sourcing. It would be wonderful if, among other things, many more sources could be tracked down and cited. —jacobolus (t) 19:44, 16 June 2008 (UTC)

I believe the main point of confusion to be differing interpretations of the purpose of this section. I wrote it to describe the history of the model/pattern/architecture that is Comet—the thing itself—while Damiens wants it to be mainly a history of the term Comet and its usage. The same misunderstanding is evident in other conflicts, such as whether alternate names for the same technology, including Pushlets and Reverse Ajax, and the largest section of the article about Push technology, should be merged into this article, or whether those should instead duplicate all this content several times. If this article discussed only the term, as a word, then having those all be separate articles would be much more reasonable. —jacobolus (t) 19:55, 16 June 2008 (UTC)

3 points:

Both Restepc and Damiens have in their statements above mischaracterized my statements on this talk page, my edit summaries, and my edits themselves. Damiens has also mischaracterized his own edit history. Restepc seems to have been editing all along in good faith, and I think the cause of said mischaracterizations has simply been a miscommunication between us, and to some extent a lack of careful examination of Damiens’ actions, without which context some of my statements may lose their intended meaning.
My “incivility” was based on a series of clearly condescending and abusive edits and statements by Damiens, which I construed to be intentionally designed to incite anger. I trust going forward that he has improved his attitude, and will be able to work towards constructive consensus; I will from this moment renew my assumption of good faith, unless it is again proven misplaced.
I don't have much time to spend on this for a few days (but I'll try to look at least once a day). I certainly do not object to seeing massive changes to the article, assuming that they are reasonable, are adequately discussed, and are based on a real consensus. What I don’t want to see is blanket removal of content, simply for lack of sources (most of the justifications for removal basically boiled down to this).

—jacobolus (t) 19:36, 16 June 2008 (UTC)

What do you think should be done about the content here? --- tqbf 20:01, 16 June 2008 (UTC)

I've been doing a little research to try to figure out where to head from here. The article evolved from 27 March 2006 to 27 May 2008, largely driven by Jacobulus since 18 October 2007. On 27 May 2008, Damiens.rf began a sequence of edits, changing the article size from 48K to 12K in approximately 46 commented edits and summarized with this diff: http://en.wikipedia.org/w/index.php?title=Comet_%28programming%29&diff=216002173&oldid=215165715. then the conflict started with two back-to-back reverts:

Jacobulus: rv 40ish edits which amount to utter destruction of the article. please don't vandalize wikipedia.
Damiens.rf: rv last - this is not vandalism. this article is a shameless advertisemt of a group of people

I think the place to start from should be that diff, because everything since has been an ignoring of consensus. It's debateable who started it, whether it was Damiens.rf's (rather biased IMO) edits, or Jacobulus' full revert.63.64.108.5 (talk) 20:18, 16 June 2008 (UTC)

I have not reduced the article size in just one edit. I have done that in almost 40 edits, so that they could be contested independently (i.e., fixing the article instead of blanking it). Jacob refused to address the my concerns independently and insisted on bulk reverting. --Damiens.rf 20:24, 16 June 2008 (UTC)

Perhaps you didn't see the 'sequence of edits'? Or the plural 'edits' at the end? I never said you did just one edit.63.64.108.5 (talk) 20:28, 16 June 2008 (UTC)

(edit conflict) Just a comment. Where does he say you did it in one edit? They state "began a sequence of edits" - and gave a diff that was illustrative of the type of edit, and a place to start from. Generally speaking, it's good to shy away from editing other peoples comments. You should make your response below his. AtaruMoroboshi (talk) 20:30, 16 June 2008 (UTC)

I have a problem with "and summarized with this diff:...". I don't agree that diff is illustrative of the type of edit. --Damiens.rf 20:42, 16 June 2008 (UTC)

I added 'in approximately 46 commented edits'... The purpose of the diff is not to point a finger, but to provide a basis from which to rebuild the content 63.64.108.5 (talk) 20:51, 16 June 2008 (UTC)

I also should add that I don't mean to imply that all of your edits are biased or unmerited. Many were good. But there is an overall defensiveness to the tone of your edit comments. It is that defensiveness that I find to be biased 63.64.108.5 (talk) 20:55, 16 June 2008 (UTC)

I notice one of the other issues brought up here is the preponderance of references to a single web site; it should be noted that useful reference material on Comet is fairly abundant. See, for example: 1 (from an independently-notable industry figure), 2 (from a public university), 3 (from the coiner of the term "Comet"), etc., and the JavaScript-themed commentary blog Ajaxian has a topical archive of Comet material and references as well. —Preceding unsigned comment added by Ubernostrum (talk • contribs) 00:41, 17 June 2008 (UTC)

Greg's view

Adding my 2c to the post mortem of the dispute, I think that the root cause was that damiens.rf massive edits were for mixed motives. Yes there were many wikipedia rules and guidelines that were broken and needed to be corrected, but Damiens.rf was also an obvious comet sceptic and questioned the need for the topic in the first place. Thus the edits were made with a strong bias, which I think provoked the opposing biased view to reassert itself. To that effect, I think a good effort was made and some valuable citations and links were located and added. But provision of those sources did not appear to alter Damiens.rfs sceptism towards comet and his edits continued to be biased (eg replacing the word "used" with "created" as he says that comet is hardly used), at the same time making efforts to discard the contributions of comet developers as biased, COI, POV, etc. etc. The parallel edits on the "Push Technology" page were also biased against comet - in that they grudgingly acknowleged that comet existed as term, but that somehow it was not deserving enough of a mention in the general section and had to be burried deeper in that page. This was a continuation of the attitude that the efforts to document "comet" were simply an attempt to advertise it or to hype it. There have been very few nuetral editors involved here and the wikipedia processes appear to descriminate against those with a bias due to their close involvement with the subject in favour of those with a bias due to their lack of knowledge. There has been very little effort to coach the contributors in the ways of wikipedia. eg. There was no grace period given to obtain citations, by the time the citations were obtained, the text they supported was long gone. Gregwilkins (talk) 01:32, 17 June 2008 (UTC)

(replying to Ubernostrum, not sure if i am meant to insert or add to the end). You havent demonstrated your point very well, in fact the opposite. Of the links you have given, the first is good (and i believe was already used in the article), the second link is broken, the third is a cometdaily contributer and the ajaxian link contains mostly links to cometdaily articles. You have demonstrated very well that most of the useful content on comet comes from cometdaily or its contributers. I think, as was mentioned much further up, a good part of this discussion should focus on what should be done with a topic like Comet where most, if not all, of the available information falls foul of the WP rules if they are strictly applied. Martintyler (talk) 07:22, 17 June 2008 (UTC)

The second link works fine for me, pulls up a nice paper from Delft University discussing Comet and related subjects with a lot of useful material that corroborates the reliability of the sources Damiens has been questioning.

Also, I could easily turn your argument on its head: if industry publications frequently cite a given source when talking about Comet, isn't that an indicator of that source's reliability on the subject? Why do you instead interpret it as those sources apparently being infected with unreliability? Ubernostrum (talk) 01:14, 21 June 2008 (UTC)

I prefer the shorter version, the longer version doesn't explain the topic better and mixes web development with Comet in a way that you don't understand in the end what this is about, something crisp and short would be better right now. - 62.127.211.2 (talk) 07:28, 19 June 2008 (UTC)

Greg, I mostly agree with your assessment of what's happened here, and with the source of the issues. I've been trying to keep an eye on this because I'd like to think I can help out a bit with the root problems -- I'm unaffiliated with any of the disputed sources and have no investment in Comet as a technology, but I do work in the field and so have the requisite knowledge regarding technical issues and reliable sources -- but I don't know that there'll be much progress without broader input and, of course, any attempt to obtain such input will be dismissed as "canvassing" Sigh. Ubernostrum (talk) 01:11, 21 June 2008 (UTC)

Canderson7's sandbox

OK, discussion seems to have abated somewhat, so it's time to move forward. It seems to me, reading the discussion above, that most editors feel that we should work on editing the longer version of the article. I have created a Sandbox for this purpose (I've intentionally put it in the Talk namespace). Please feel free to edit this page in ways that conform to the consensus being achieved on this talk page. For example, it seems to be commonly felt that the history section should be re-factored, perhaps using Damiens' suggestions here as guidelines. Several other ideas that most seemed to find agreeable were raised by Tqbf here. As for the reliable sources issue, I've spent a while reviewing policies / guidelines / essays and it is my opinion that CometDaily is an acceptable source for technical information on Comet. A diversity of sources is always preferable to over-reliance on one, but I do not believe that content should be removed from this article only because CometDaily is its source. If content from CometDaily is POV, non-encyclopedic, or has some other editorial problem, it should be edited first and removed only as a last resort (and after discussion). For the time being, content that is unsourced should be marked citation needed rather than removed. I, for one, plan to start editing in the sandbox sometime today, and I'd appreciate assistance from more knowledgeable editors. Canderson7 ^(talk) 21:04, 18 June 2008 (UTC)

Conderson7, totally agree with all you have to say. I'd like also to contribute to the sandbox page, but am away for a while. But don't let my silence be interpreted as not supporting that effort and I hope that I'll be able to help resolve with any remaining citation needed on my return Gregwilkins (talk) 09:00, 19 June 2008 (UTC)

As long as there's enough volunteers to fix the long version, it's o.k. to use it. But I have to disagree with your position when you (just once) departed from the consensus above by accepting cometdaily as a source. Cometdaily is a collection of blogs, where comet enthusiast express their point of view. We can't use what they write about comet because they somewhat hype it, and we shouldn't use what they write about web-developent in general because there are so much better published sources for that. --Damiens.rf 11:19, 19 June 2008 (UTC)

I disagree with the statement development blogs in general are "hype." There is value in using comet daily as a source. To put it into context, a development community culls information from developers, often provides source code and pertinent information. A note about comet daily's contributors, they are vetted, so anyone can't just post. A quick distinction as well, the site isn't "a collection of blogs" It's a single blog, with multiple posters. I support using this source, but also strongly suggest other sources are used as well. AtaruMoroboshi (talk) 12:53, 19 June 2008 (UTC)

I didn't say "development blogs in general are "hype."". I said cometdaily hypes comet. The point of view exposed there is overly enthusiastic about comet adoption, power and influence. --Damiens.rf 15:18, 19 June 2008 (UTC)

As with most dev blogs, the developers are excited about a product/platform they are developing for. I don't think enthusiasm in it of itself makes the source invalid. However, I think it would be fair to advise editors to be cognizant of this when using comet daily as a source. i.e. a quote from cometdaily that says "X is the best thing ever!" is no good, where as "X is a tool for Y, here's an analysis, Z has used it for the following application..." are. Fair enough? AtaruMoroboshi (talk) 15:51, 19 June 2008 (UTC)

Why wouldn't we want to use published sources for information like "X is a tool for Y...", "Z has used it for the following application..." and mainly "...here's an analysis..."? --Damiens.rf 16:29, 19 June 2008 (UTC)

It is a published online source and given the nature of the topic, you'll be more likely to find online sources as opposed to print. The Ajaxian comes to mind. I am not advocating the exclusive use of comet daily as a source. I refute your statement that essentially the information coming from the site is no value because it's features contributions from enthusiastic comet developers. If there are additional sources that better support a statement in article, those could be sourced in place of a cometdaily source. But we'd just be arguing hypothetical right now. I suppose we should focus on editing the actual article, right? :-) Cheers, AtaruMoroboshi (talk) 18:06, 19 June 2008 (UTC)

AtaruMoroboshi, if you really believe that cometdaily is a "published online source", let me respectfully point you to the Wikipedia policy called Wikipedia:Reliable Sources. --Damiens.rf 18:26, 19 June 2008 (UTC)

Damiens.rf, Thank you. I've read the policy before. Perhaps then, the issue should be addressed at Wikipedia:Reliable sources/Noticeboard as Canderson7, Gregwilkins, and others have stated they believe it to be an acceptable source. AtaruMoroboshi (talk) 18:35, 19 June 2008 (UTC)

There's already a thread there. Gregwilkins, and few of the "others", are not exactly third-part nor independent, but they are all invited to expose their arguments. --Damiens.rf 18:57, 19 June 2008 (UTC)

Thanks, I see no one else has contributed to the dialog aside from you and Restepc. I'll try add my comments over there, though I'd be re-iterating my points here. I was unaware Greg Wilkins was affialiated with cometdaily. Not withstanding, a serious consideration needs to be made, and consensus should not be a result of two editors only. (It is my opinion that this is largely how we got into this mess) AtaruMoroboshi (talk) 19:27, 19 June 2008 (UTC)

Some sources that support Comet Daily can be an acceptable source. The O'Reilly Media Open source-con [37] features speaker Michael Carter [38] who "shares his expertise in bi-weekly articles for the technical publication Comet Daily. (emphasis added) AtaruMoroboshi (talk) 19:27, 19 June 2008 (UTC)

I'm not sure how this can help. Anyone can "shares his expertise" in his/her own blog. That's exactly the point of WP:V#SELF. --Damiens.rf 20:28, 19 June 2008 (UTC)

I think you're seriously misunderstanding several issues here regarding reliability of sources, mostly because you seem to be operating without context.

O'Reilly is a well-known, respected and reliable publisher in this domain, and the OSCON conference vets and selects speakers and talk subjects; thus, an endorsement of reliability from O'Reilly should carry some weight in this discussion. For example, if I were to say on my blog that I'm "sharing my expertise" on a technical topic, that wouldn't necessarily mean anything, but if a respected technical publisher or a major industry conference were to say that I have expertise or am an expert on that topic, that would mean something as far as Wikipedia is concerned.

Also, the fact that a particular publication or collection of publications is run by people who are enthusiasts about a particular technology does not mean that their statements must be discarded; Wikipedia should be providing verifiably-correct information about, for example, the technical nature of Comet and the specific techniques involved, and one source of such information is Comet Daily (there are others, and I've pointed out a couple in other comments; the fact that multiple sources corroborate the information weakens your arguments quite a bit). Learning to differentiate the verifiable technical information (which Wikipedia should include) from marketing material (which Wikipedia should not include) is an important skill, and when applied to this article would, I believe, resolve your concerns quite neatly. Ubernostrum (talk) 00:14, 20 June 2008 (UTC)

I think you hit the nail right on the head there regarding O'Rielly. Just observation this conversation has forked into two locations - I suppose there should be a centralized discussion, or even excizing this from the current section. As there is alot of dialog going on right now. I've added your resonse to the relevant discussion at the reliable sources board. AtaruMoroboshi (talk) 16:11, 20 June 2008 (UTC)

Replied there. (Ubernostrum, would you agree in migrating this discussion for the Reliable sources Noticeboard, or do you prefer to keep it here?). --Damiens.rf 16:42, 20 June 2008 (UTC)

In my experience, the biographic blurbs about an author or a conference speaker are written by the author/speaker. (My experience is from consulting firms in the 90s and 00s. We sent many speakers to programming conferences and had employees write articles for magazines and outside websites. Our firm always wrote the bios for our authors and speakers, which were then used by the conference, website, or magazine.) Brilliant Pebble (talk) 20:00, 23 June 2008 (UTC)

Summarizing changes needed to start fixing the long version

Here's my suggestion of a TODO list for beginning the fixing of the long version:

Remove the quotations.
Replace every blog-references by a {{fact}} tag - Minor exception is the factual reference for the blog post where the term "comet" was invented.
Refactor the History section - my suggestions at #Problems_with_the_History_section
Refactor or throw away the Browser_compatibility section - since it's mostly a copy of Jacob's blog. A section on "compatibility" may stay, as long as it's written based on published sources.
Fix the definition of comet in the first paragraph - The short version made a good job on that.
Remove the Notes - and incorporate any useful information in the article's text.
Cleanup the Implementations section - Some minor discussion here (it's also funny how the only one implementation described with a non-neutral term is the "cross-language scalable Comet server" Orbited, that happens to be Jacob's own project).
Put lots of {{or}} and {{who}} tags, and fix them.

I suggest volunteers work from the Canderson7 created sandbox and, as soon as the article get unprotected, we move to it (since it's so trivial to make progress at this point). The {{advert}}, {{cleanup}}, {{onesource}} and {{COI}} should of course be kept until the problems are fixed.

We could wait even for one month (more needed?) for the {{fact}} tags to be fixed, and have a compromise not to remove the problematic content for the time being. But if that approach doesn't work (if volunteers interest fade, for instance), we would have to take the stubing path again. Please note that restarting an article from scratch is not throwing content away, as suggeested by the reddit rant, but instead, it's a commom last resort Wikipedia practice for dealing with problematic articles (usually, but not necessarily, bios) [39] [40] [41] [42] [43]. --Damiens.rf 11:58, 19 June 2008 (UTC)

If so, a few of these sources should be factored into the article content, as they are worth linking readers to.
Damiens, I don’t think you’re going to win this one—on a topic like Comet, there just aren’t newspaper explanations of every relevant detail, but that doesn’t mean that the sources used aren’t reliable. Not all blogs are created equal, and several of these authors have outstanding reputations in the web development community.
Most of this criticism boils down to criticism of the sources, plus some criticism of the style of a few of the sentences. But I would like to hear your suggestions as to the overall organization of a history section. What do you think is significant in the history of the technology?
This, if you read this comments section, is not going to happen (throwing it away), as it is essential knowledge for anyone planning to discuss Comet. Sources can easily be tracked down for it, though few if any of them are as clear or concise as my article about the subject, and some are misinformed to one degree or another, so pointing readers at them is not ideal either.
The definition placed in your stubbed article was inaccurate and POV.
Why? Several featured Wikipedia articles use this device.
This section needs major work, and I accept your criticism of my description of Orbited, which should be changed ASAP.
Go for it.

I don’t think you have the same definition of “last resort” that I do.

Cheers. —jacobolus (t) 20:57, 19 June 2008 (UTC)

Incidentally, here is the “reddit rant” you refer to, for the benefit of other editors:

There is something inherently broken about the way Wikipedia works. Recently, _why's article was nominated for deletion, by someone not at all familiar with the subject at hand, claiming WP:NN. Now there is the Comet article, which went from being decidedly useful (to the point that I have pointed customers and users towards it as an introduction to the topic), into a complete fluff piece devoid of any real information beyond the first sentence.

The problem, as always here, is that for any subject outside of the complete mainstream, domain experts are not allowed to contribute, since they, per the broken Wikipedia policies, can't possibly maintain a NPOV or avoid conflict of interest.

IMO, Wikipedia needs to get a policy against dumb revisionist assholes, or someone needs to start an alternative with less broken policies. In particular, I'd like to see a wiki concentrated around computer science and related topics where the wikitards can't destroy perfectly good articles

—Arve [Bersvendsen, Opera]

I don’t necessarily agree with his conclusions about the brokenness of Wikipedia—apparently when enough eyes come, it is possible to un-break the process (at least so far)—and he was misinformed at that point about NPOV. But I can certainly see where he’s coming from. —jacobolus (t) 21:11, 19 June 2008 (UTC)

Completely agree with the eyeballs-number theory. You wouldn't have being able to write this article the way you wrote if more editors were aware of it at the time. --Damiens.rf 22:03, 19 June 2008 (UTC)

And thanks for posting Arve's rant. You just forget the title: "How Wikipedia deletionists can ruin an article" (but I've learned I shouldn't touch other editors comments). --Damiens.rf 22:05, 19 June 2008 (UTC)

Consensus seems to be that we should start with the long version and make roughly, but more selectively, the same edits Damiens made originally, so I'm going to do some broad brush strokes on the sandbox now, I haven't been around much cos my computers on the blink, but I'll make a start. Restepc (talk) 01:46, 20 June 2008 (UTC)

I would strongly disagree with your characterization here; making "the same edits Damiens made originally" lands us back at the highly-disputed version of the article, and the fact that it is disputed means it does not represent the consensus. Ubernostrum (talk) 00:54, 21 June 2008 (UTC)

Restepc was not talking about the "restart-from-stub" version when he said "the same edits Damiens made originally". A lot of things happened to this article way before that rant was posted to readit, although the the rant's author somewhat let the impression that it was just somehing like "(1) This article is not perfect, (2) Let's make it a stub instead of trying to fix it, (3) Profit." --Damiens.rf 01:51, 21 June 2008 (UTC)

Done, just quickly cut out major chunks that needed to go; if anyone thinks something I did has gone against the consensus feel free to revert that particular edit, but please don't start bulk reverting again. I'll come back and work on the remainder more closely (and sort out the broken refs) when I get my fancy (alright; cheap) new laptop. Restepc (talk) 02:06, 20 June 2008 (UTC)

I mostly reverted your edits, which are not remotely aligned with “consensus”. Among the parts that I left taken out, more discussion is needed:

Alex Russel’s quotation about the purpose of Comet the term is useful, as it helps justify the use of a “buzzword”, exactly the problem that several of these unfamiliar editors had, as they brought up questions of the term’s notability and usefulness.
The implementations section is a problem, but it also serves a useful function for readers. Is there some way to slim it down without being POV?

—jacobolus (t) 03:03, 20 June 2008 (UTC)

Seems like the consensus to me, but I'm sure it won't hurt to have the same discussions with the same conclusions yet again. I will re-revert one revert you made; we definitely do not need your website in the external links twice. I'll leave the other changes to other people for the time being. Restepc (talk) 03:16, 20 June 2008 (UTC)

Please start a discussion about a specific large change you’d like to make, and we can work through these one-by-one, without confusion. This is a sandbox page; there’s no hurry to do everything at once. —jacobolus (t) 03:27, 20 June 2008 (UTC)

A few (1-4) examples would be nice in a perfect article, the most notable/sourcable ones probably; I felt that this function is covered by the 'popularisation' section, and that'd probably be the best place to include more examples if you want to.....the most notable one should go in the lead to, but that's probably trying to run before we can walk.....Restepc (talk) 03:20, 20 June 2008 (UTC)

Yes, but I don’t see how those can be picked neutrally. How do you draw a line? Whoever does it will be arbitrarily inserting their own judgment. Anyway, no, implementations definitely do not belong in the lead. —jacobolus (t) 03:31, 20 June 2008 (UTC)

I believe changes and opinions made by those who have zero understanding and no experience with the topic should be weighed lighter than those who are experts on the topic. For example Damiens.rf referred to ajaxian.com as 'some other guys blog'. Ajaxian.com is the most authoritative and well known news source on all things AJAX and AJAX related. In addition, cometdaily is the most authoritative and well known news source on COMET. Clearly this user has no knowledge or understanding of the topic and edits and opinions this user has to the 'worthiness' of the content should be valued lower than a user who has knowledge on the topic.Sembiance (talk) 13:16, 20 June 2008 (UTC)

So you should seek for a change in the core Wikipedia policies. Here, we trust in reliable, third-party, published sources with a reputation for fact-checking, not on editors who are "experts on the topic". --Damiens.rf 15:36, 20 June 2008 (UTC)

Comment The sources, not the editors. A note for all editors, let's all try to separate the issues. Whether is source is reliable on it's own merit, and whether there are editors with vested interests editing this article are two separate issues. Both which need to be addressed. Damien, Ajaxian.com is a reliable source [44] with it's founder(s) speaking at the Google I/O conference. Before anyone says the webware blog is "some other guys blog" It is a CNET.com blog, subject to editorial oversight, and the link leading to the article from CNET's main site is here [45] Quick Edit: Here is a recent eWeek news article which interviews both founders of Ajaxian, about a ARAX technology. [46]AtaruMoroboshi (talk) 16:40, 20 June 2008 (UTC)

You can re-phrase what I said as "Here, we trust in reliable, third-party, published sources with a reputation for fact-checking, not on ~~editors who are~~ experts on the topic". Try to think in terms of "Reliable Sources" instead of "Reliable People", as you seem to be doing. Google I/O reviewed and accepted a speech by Ajaxian's founders, but that doesn't make what the Ajaxian's founders write on Ajaxian.com any more reliable, since Google I/O doesn't review and accept everything those guys post there. That's the whole difference between self-publishing (i.e: what they write on Ajaxian) and being published by a third-part (i.e: being granted a speech on Google IO). --Damiens.rf 16:55, 20 June 2008 (UTC)

That's not what I'm doing Damiens. Take any field (software development or not)- if a person (or persons) are notably recognized in their field by their peers or the industry as a result of what they've created or improved, and this is confirmed by third parties, then a website where they speak to their area(s) of expertise is likely credible and can be used as a source. Example: Game developer John Romero, who rose to prominence as the co-founder of id software. Say he makes commentary on the state of the video game industry or a video game product, maybe it's in a blog post and later it's picked up by GamePro. It is credible as he's a well recognized and notable individual in his field. The same would apply here. AtaruMoroboshi (talk) 20:09, 20 June 2008 (UTC)

I strongly disagree with points 2 and 8 here; the blanket approach you advocate does not represent the consensus on this matter and does not feel like an NPOV treatment of the available sources. Ubernostrum (talk) 01:04, 21 June 2008 (UTC)

Quotes

OK, almost everyone has suggested removing the quotes (although jacobolus has reminded us that any valuable content should be merged into the article). There are currently seven quotes in the article. I list them below with my recommendation.

Joe Walker from CometDaily. Here's where I agree with Damiens.rf. This is a highly POV quote that has no meaningful content and simply serves to aggrandize Comet. Saying that Comet lets us treat the web as a network of people interacting more naturally than they can behind a request/response barrier does nothing to enhance the reader's understanding. Check out Wikipedia:TIGER. POV quotes can be stuffed nicely and put on display, but simply plopping this quote next to the article and not even addressing it in the body makes it seem like we agree with what Joe is saying. This makes the quote a tiger loose in the museum, and it should be removed.
Lars Rasmussen as quoted in Computerworld. This quote is better. It's not from a Comet evangelist and it conveys fact not POV. However, I note that Comet is not mentioned by name once in the Computerworld article, so using that quote here has a vague hint of OR about it. I'm neutral on whether we remove it.
Alex Russell from his personal blog. Same problems as #1. Should be removed.
Greg Wilkins from CometDaily. This quote, unlike 1 and 3, actually has a small amount of fact in it. It remains POV but is closer to permissible than those two. If we comment on the quote in the article, pointing out that Comet proponents claim certain advantages, then I think it could stay. Overall, I'm neutral.
Alessandro Alinone from lightstreamer.com. This quote is hyperbole and does not seem to me to contribute useful information. I think it should go.
Michael Carter from his presentation at "The Ajax Experience Boston". Doesn't egregiously violate policy or guideline. Keeping it in vs. taking it out is an editorial decision, and I'm neutral.
Just van den Broecke from javaworld.com. Same as #6.

Summary: I feel that 1, 3, and 5 should be urgently removed. The rest should be allowed to sort themselves out during the editing process. I'll act on my own advice if no one objects. Canderson7 ^(talk) 19:28, 20 June 2008 (UTC)

Okay, I somewhat agree about 1 & 3. #5 is mostly a simple statement of fact. #2 is from an article which doesn’t mention Comet because it is a quotation about Google Maps, which is not a Comet application. [That section is here for context, so that readers can quickly learn (in 1 ¶ each) what Ajax is, hear a brief example of an application it enables, and why Ajax (without “Comet”) is still insufficient for real-time applications.] #4 explains why “Comet” is preferable to polling, with a more specific quotation than the summary in the article body. #6 could probably be removed. #7 is I think worth leaving. #8 (“common terminology…”) provides some answer to the objections that “another buzzword” is unneeded. Those objections should perhaps also be covered in the article, but I’m not sure where they would be most appropriate. —jacobolus (t) 19:52, 20 June 2008 (UTC)

In general, I'm against the use of the quotations because, even when they are not endorsing and opinion, they give the impression that the person being quoted is somewhat "Important". I would be o.k. with quoting Albert Einstein in an article about black-holes, because Einstein's opinion/sayings about black-holes are highly regarded as important. But otherwise, the quotations should only be used to showcase citations that are going themselves to be discussed in the article;

I agree with the immediate removal of the opinion-endorsement #1 and #3.

I point to the fact that #4 is misleading: It starts with "...This analysis shows that..." but actually points to a blog post. The "analysis" mentioned is just the calculations the author made and the conclusions he draw, and that have never being subject of any serious peer review. This is a good example of the big quotes augmenting the perceived importance of some saying.

The other quotations, to me, seem to add nothing that couldn't be (or already is) addressed in the article's text (like #5, that Jacob characterized as a "simple statement of fact"). Their presence seem to serve solely to exalt their authors. --Damiens.rf 20:24, 20 June 2008 (UTC)

OK, I'll remove 1, 3 and 6. I'll keep 5 in there against my inclination and pending comment from other users. I'm missing something obvious, I'm sure, but where's quote 8? For some reason, I can't seem to see it. Canderson7 ^(talk) 20:12, 20 June 2008 (UTC)

Apparently #8 was already removed from the sandbox. —jacobolus (t) 20:48, 20 June 2008 (UTC)

Editors with COI

What worries me is that jacobolus is very vocal on this talk page and the direction of this article, a big amount of comments are from him. I think that everyone who has a conflict of interest should not be actively editing this article or define its outcome. I don't understand why he is allowed to be that active here. Is there a way in Wikipedia to make sure that a) not only the loud voices are heared and b) people with COI do not actively participate? I am concerned about the neutrality of this article. - 213.115.160.70 (talk) 09:37, 21 June 2008 (UTC)

You have an inaccurate understanding of WP:COI. —jacobolus (t) 10:56, 21 June 2008 (UTC)

You have been already reported to the COI noticeboard once, short reminder: you promoted yourself by quoting/linking to your own blog articles. - 83.254.214.72 (talk) 16:56, 28 June 2008 (UTC)

Despite COI concerns I don't think anyone should be excluded explicitly from making this article better. Jacobolus is an editor who is closely involved in with COMET, his involvement is established and spoken to multiple times on this talk page. We should not simply ignore his contributions, especially if they can be adequately sourced. As per consensus regarding the sandbox article, discuss, refactor and other wise tag the sandbox so that we can address the neutrality and content of the article. One of my concerns, the tone of the article needs to change, it reads like an academic paper/white paper more than an encyclopedia article in some places. AtaruMoroboshi (talk) 12:33, 21 June 2008 (UTC)

Academic paper? For me, it has the tone of a paid magazine article. --Damiens.rf 16:08, 21 June 2008 (UTC)

Great. So we're in agreement the tone of the article needs to be addressed. Your edit summary however, was not necessary. Please try to be civil. I know this can be a frustrating and long process, but there is no need for that. Thanks AtaruMoroboshi (talk) 03:20, 22 June 2008 (UTC)

Apart from COI concerns there is a problem of ownership WP:OWN, I am concerned that Jocobolous is too possessive and will (again) revert changes so that the article matches his ideas. -- 83.254.214.72 (talk) 16:48, 28 June 2008 (UTC)

Could you please WP:AGF and stop the baseless aspersions? Unless you have something more specific to say here, as far as I can tell you are merely being disruptive for the sake of disruption. —jacobolus (t) 05:42, 30 June 2008 (UTC)

Such comments foster incivility. Please don't verbaly attack concerns from others, more than one person raised the problem of ownership in connection with your name. - 62.127.211.2

Comments such as those by 83.254.214.72 do seem to be particularly designed to foster incivility. I agree wholeheartedly. —jacobolus (t) 00:49, 1 July 2008 (UTC)

The problem of ownership WP:OWN has been first raised by a wikipedia admin, already before the edit war escalated. Please stop attacking concerns from others, it is perfectly fine to raise concerns on a talk page. By denieing these problems (COI and ownership in particular) you just make it harder for your fellow co-editors to work with you. - 62.127.211.2 (talk) 10:04, 1 July 2008 (UTC)

Intro Section

Several users have raised concerns that we're trying to write an article about too ill-defined a term. For example, see the edit summary on this edit or some of the above discussion threads such as #Comet_means_pushing.2C_polling.2C_streaming.3F or #Merger_of_all_these_things_into_new_article. With these concerns in mind, I've tried to edit the intro so that it's clear what we're talking about. My edit is here. The two chief goals of this edit were to use the words neologism and umbrella term prominently in the intro, because it seems to me that those best sum up the significance of the term Comet. As someone who had never heard of Comet before reading this article, it was not immediately apparent to me that this is what Comet was. My first thought was "OK, so where can I download this?" I believe that it is now more clear what Comet is and what it is not.

I've also taken the opportunity to organize the intro into three paragraphs each of which conveys only one idea. The first paragraph explains what Comet is, the second explains how it differs from past things, and the third briefly informs the reader where the neologism came from. As part of this edit, I've split the content of Note A between the intro and the popularization subsection of the history section, a decision which I believe to be aligned with the below consensus. Anyway, please let me know what you think. Keep in mind that I'm not very knowledgeable about the subject and could use any pointers you have. Anyone who want to polish up my writing is also more than welcome to do so. Thanks. Canderson7 ^(talk) 02:07, 22 June 2008 (UTC)

I haven't looked over the changes in detail. But I do agree with the change that integrates information about the coining of the term comet into the body of the article (as discussed below). AtaruMoroboshi (talk) 03:25, 22 June 2008 (UTC)

I agree it's important to emphasize the neologism and umbrella term nature of the term. I do however have some minor suggestions about the current leading paragraph:

Instead of "In World Wide Web development, Comet is...", why not just say "In web development, Comet is..."?
Instead of "Comet is a neologism that describes...", I believe it would be more accurate to state "Comet is a neologism created/coined/devised to describe...", as one of those verbs would be more inline with the just mentioned "neologism" nature of "Comet".
I believe it's technically wrong to call comet a "web application architecture". It's more of a "set of techniques", or even better, a "communication design pattern", as Thomas Powell described in his book (published by McGraw-Hill) that I mentioned at #Leading paragraph (or "what's comet anyway?").
Fix the tone in "...without any need for the client..." to "...without the need for the client..." (the usual enthusiastic tone of the article).
Simplify "...use long-lived HTTP connections between the client and server" by removing the unnecessary "between the client and server" (since HTTP is a request-response protocol, its connections are always "between the client and server").
The intro (and the article?) says that Comet differs from ajax, but I thought it was a subset of it (since it's a term created to describe something that already existed and that, at the time, "that something" was simply considered one of the many ways of doing Ajax (the book mentioned above supports this)). We should somehow made it clear that Comet is a flavor/style of ajax. --Damiens.rf 00:16, 23 June 2008 (UTC)

--Damiens.rf 00:16, 23 June 2008 (UTC)

Sounds good to me.
That would be wordy and redundant.
Not according to the common understanding of “Comet”. Like “Ajax”, “Comet” is best described as an “architecture” or an “approach towards architecture” or similar. “Set of techniques” implies that it is the techniques which make the concept, which is wrong, and “design pattern” has a different technical meaning. If there are those who call “Comet” a “design pattern”, then that is only through their own sloppy and incorrect use of the term. (edit to add: in the earlier quotations you pasted above, he did not ever call it a “design pattern”; is “communication design pattern” actually a direct quotation or are you making that up?). Further edit: in any case, to call Comet a “pattern” would be puffing it up, giving it more credit and making it more abstract than it deserves or is. You could perhaps call it an “architectural pattern applied to the browser” or something, but that would be confusing and verbose.
Go for it. I don’t see how that is enthusiastic, but whatever. How about “with no need for …” to be even more concise?
Go for it.
It says that Comet differs from the normal architecture of Ajax applications, which is true. Saying it is “a way of doing Ajax” is confused and inaccurate (if a book says this, it is imprecise at best). Comet instead is built on top of Ajax, an extension of it, which uses many of the same browser objects in a quite different manner. To say that Ajax encompasses Comet would be accurate. “Flavor/style of Ajax” is vague enough to be nearly meaningless.

—jacobolus (t) 00:43, 23 June 2008 (UTC)

Good.
Replacing "that describes" by "coined to describe" doesn't make the text "wordy and redundant". It's just one word more and it isn't redundant to anything.
The book says "communication pattern" and, as I explained above (yes, direct quotations from the book pages I've mentioned), my understanding is that it means a design pattern. I hope you understand that even if you believe a book to be "sloppy and incorrect", we will prefer accepting the reliable, reviewed, published book to your self-appointed expertise.
Yes, “with no need for" is even better.
Good.
Yes, saying that "encompasses Comet" is what I intended to mean. I believe the intro should made it clear. Can you do that?

--Damiens.rf 02:12, 23 June 2008 (UTC)

There’s no reason to add “coined to describe”, as the relevant point is the description, not the coining, which is implied. Any web technology term is understood to have been “coined” sometime within the past 20 years. Also, “pattern” ≠ “design pattern”. —jacobolus (t) 03:40, 23 June 2008 (UTC)

I disagree. Using saying comet was "coined to describe something" instead of saying it is "used to describe something" server exactly to tell it apart from well established terms like HTTP, Web, TCP, E-mail, etc, that are around for decades and were not made up in a blog post. I believe this characteristic of comet is seminal and the article should not give the impression (even by omission) that the term is that established (for instance, in the book mentioned above, the author found it important to note that "Comet isn't on the developer's lips just yet..")

I kind of understand that pattern isn't the same as a design patter, but do you think the author could be meaning a communication design pattern when he said "communication pattern" in the book? I think it's plausible. But if not, don't you (all) agree that using "communication pattern" will be better than the unverified "web application architecture"? --Damiens.rf 04:17, 23 June 2008 (UTC)

As a further follow-up to #3, Jesse James Garrett said of Ajax (and I think this would hold for Comet as well):

“A. Ajax isn’t something you can download. It’s an approach — a way of thinking about the architecture of web applications using certain technologies. [...]”

“Q. Is Ajax a technology platform or is it an architectural style?

“A. It’s both. Ajax is a set of technologies being used together in a particular way.”

—jacobolus (t) 01:06, 23 June 2008 (UTC)

Sorry, who's JGG? --Damiens.rf 02:12, 23 June 2008 (UTC)

Jesse James Garrett. The guy seems influential. I believe we can use this. Where was this interview published? --Damiens.rf 04:17, 23 June 2008 (UTC)

It is from the (q/a follow-up to) essay where he coined the term Ajax. Sorry, I thought that was clear. Go read that, and then Alex Russell’s similar post introducing the term Comet. I think it will clear up some of your points of confusion. —jacobolus (t) 05:55, 23 June 2008 (UTC)

Makes sense now. I see how Ajax can be described as an architectural style, and also does Comet. And I don't think this is far away from saying it's a "web application architecture". But I somehow have the feeling that "web application architecture" implies an architecture for the whole web application (e.g.: The GMail web application, or the e-bay web application), but Comet/Ajax is always used to build just some of the components of a complete web application.

To make myself clear, I'm afraid that when we say "Comet/Ajax is a web application architecture", some people will think it would make sense to say "The architecture used for the FaceBook web application is Comet/Ajax".

Is there any concise way of saying that Comet is an architecture for web-app components, instead of an architecture for web-apps? (Or am I just elaborating from a false assumption?) --Damiens.rf 16:10, 23 June 2008 (UTC)

Ah, I see where you are coming from. I hope readers don’t come to that conclusion, but talking about it as an “architecture for components” could also be confusing. It is unfortunate that all of these terms have such imprecise definitions (i.e. “pattern”, “architecture”, “framework”, “application”, etc.). Maybe “umbrella term for a number of technologies supporting an application architecture” or whatever is the best we can do. Either way, I think with the addition of a picture, and a fuller description later on in the article, it should be fairly clear to readers what Comet is all about. —jacobolus (t) 03:59, 24 June 2008 (UTC)

Sure it would help, but if we're to let the article confusing as it is and wait for a diagram to pop-up, we would be back to the stage when I first found this article.

What about a "In web development, Comet is a neologism coined in 2006 to describe a component architecture in which..." ?--Damiens.rf 12:25, 24 June 2008 (UTC)

Honestly, I think that makes for a poor first sentence. The first half of it is meta-description, and “component architecture” is confusing, with multiple interpretations. You'll notice that the page on blog begins “A blog is a website…”, the page on podcast begins “A podcast is a series of media files…”, the page on microformat begins “A microformat is a web-based data formatting approach…”, the page on web feed begins “A web feed is a data format…”, etc. None of these say: “In web technology, «foobar» is a neologism coined in 200x to refer to a «blah blah blah»…” where “«blah blah blah»” is carefully and confusingly worded to minimize the suggested import of the term. I just don’t think it’s a useful pattern to follow. But that’s just me. Describing it as a neologism coined in a 2006 blog post, etc., seems like a fine thing to leave to the last paragraph in the lead section, at most. —jacobolus (t) 13:23, 24 June 2008 (UTC)

I believe this is because you have the wishful thinking that "comet" will become a popular term as 'blog', podcast or web feed. It's important for Wikipedia to establish the buzzword nature of Comet in the definition, instead of joining the campaign to push the meme.

I don't see how "component architecture" can be notably confusing in this article as it stays today. Will someone reject the changing? --Damiens.rf 14:29, 24 June 2008 (UTC)

Regarding item 2, how about something like:

In web development, Comet is an umbrella term used to refer to any of several techniques by which a long-lived HTTP connection allows a web server to push data to a client at the server's initiation.

It's concise, it's accurate and it's built on the (easy to cite) article in which the term was coined. Ubernostrum (talk) 03:32, 23 June 2008 (UTC)

The one trick is that I think we also want 'neologism' in there. How about changing the first paragraph to the following and eliminating the third paragraph (which is covered in the body of the article).

In web development, Comet is a neologism coined in 2006 to describe application architecture in which a long-lived HTTP connection allows a web server to push data to a client asynchronously and without any need for the client to explicitly request it. Comet is an umbrella term that encompasses numerous specific techniques for achieving this user-interaction model. These methods rely on browser-native technologies, such as JavaScript, rather than on proprietary plugins, but have few other unifying features. Each method has its own trade-offs in terms of browser support and side effects, latency, and throughput.

Canderson7 ^(talk) 15:23, 23 June 2008 (UTC)

I like it better. I would only change "without any need" to "with no need" (as discussed above). --Damiens.rf 16:10, 23 June 2008 (UTC)

Also, I just noted the protection expired. Could you copy the sandbox to the article so we could start working directly there? --Damiens.rf 16:13, 23 June 2008 (UTC)

Thanks. I've made the suggested change (I meant to do so earlier) and updated the sandbox. I'll go ahead and update the main article in just a minute. Canderson7 ^(talk) 20:23, 23 June 2008 (UTC)

I think Ubernostrum’s sentence is reasonably fine, if a bit wordy and awkward sounding. The next 2–3 sentences can clarify the definition. —jacobolus (t) 13:29, 24 June 2008 (UTC)

I'll leave the fine-tuning of the intro to people who know more about the topic than I do. jacobolus and Damiens.rf, if you can agree on suitable wording, then don't hesitate for a second to totally re-write the current version. What I'll focus on in the next day or so is tackling the body of the article. My major goals are as follows:

- Condensing the section on pre-existing architectures. I have no interest in cutting out good content, but I also want to keep the article focused on Comet throughout.
- Adjusting the organization of the material on Comet. Right now, just looking at the table of contents, it seems a little unclear how the reader should use the article to get information on Comet that is more detailed than the introduction but less specific than, say, vertical scalability.
- Removing peacock terms and the like as I encounter them.
- Tagging uncited information and original research.

By the way, this is just a heads up, but starting Saturday June 26 and ending Saturday July 2, I will not have internet access and will not be able to help out with this article. Hopefully, I can compensate for this by being extra useful in the next couple days. Canderson7 ^(talk) 03:50, 25 June 2008 (UTC)

Notes Section. Necessary?

I'm working the sandbox version of the article. I removed what I would categorize as 'commentary'. I broke the notes in the process (setting the out of order by one) which Jacobolus pointed out when the edit was reverted [47]. Regarding the other notes though, if there is something pertinent to an explanation, why not integrate into the body of the article and do away with the notes section entirely. For example, the origin of the term "Comet" can easily be spoken to and sourced in the body of the article. AtaruMoroboshi (talk) 18:49, 20 June 2008 (UTC)

Those were things I found irrelevant to the main narrative of the article, and therefore not worth interrupting its flow for (but still relevant enough to the topic, and some readers, to keep in a note). —jacobolus (t) 18:55, 20 June 2008 (UTC)

I think the "This is not a criticism of Wikipedia" needs to be done away with - it adds little value. Note #1 about the name 'Comet' can be slightly expanded upon and sourced in the lead, eliminating the need for that note. Generally speaking, I don't think these notes add much on their own, because as you said, they aren't relevant to the main 'narrative'. If they are important place them in the body of the article. Comments? AtaruMoroboshi (talk) 19:02, 20 June 2008 (UTC)

There was already a discussion along these lines where (IMV) the consensus (only Jacob objected that I recall) was that the notes should be removed with any relevant information worked into the article. Restepc (talk) 19:12, 20 June 2008 (UTC)

I think that's fair. Jacobolus, do you agree? AtaruMoroboshi (talk) 19:25, 20 June 2008 (UTC)

No. Discuss them case by case. —jacobolus (t) 19:36, 20 June 2008 (UTC)

Ok. We are addressing "A" below. I'll address "C". What is the justification for noting that a technical explanation of portion of wikipedia is "not a criticism"? As I've stated, I don't think it adds much. AtaruMoroboshi (talk) 19:46, 20 June 2008 (UTC)

We are going to make any progress if we expect Jacob's blessing for ever edit we want to make. He's reverting most of what's being fixed in the sandbox version, and he repeatedly asks for more and more discussions for every topic, no matter how clear is the consensus we've reached. Jacob refuses to accept the consensus for fixing the History section, he refuses to accept the (broad) consensus of removing the quotations, refuses to accept the (broad) consensus of removing the notes... and mainly, he refuses to accept the consensus that Comet Daily editors should contribute to the discussion but avoid directly editing the article. --Damiens.rf 19:47, 20 June 2008 (UTC)

I agree with you, Damiens, that jacobolus could greatly expedite this process if he stopped reverting so many edits and stone-walling so many changes. On the other hand, though, other editors could expedite the process by justifying themselves briefly on the talk page before making edits to the sandbox that push the limits of consensus. Additionally, I do not believe that it was ever decided that Comet Daily editors should not edit the sandbox. I, for one, would disagree with such a measure. Canderson7 ^(talk) 19:58, 20 June 2008 (UTC)

I would very much like to see the history section “fixed”, I think that some quotations can be removed, others worked into the text, and some left, but I do not think that any of these changes have nearly the “broad consensus” you think. I have no interest in “stone-walling” anything. I am however interested in having a comprehensive, accessible, high-quality product. —jacobolus (t) 20:10, 20 June 2008 (UTC)

I agree. Take it out. Just renumber the other notes. —jacobolus (t) 20:07, 20 June 2008 (UTC)

Putting more information about the name “Comet” in the lead section adds a pro-Comet (as a term) POV to the article (i.e. unless every other name’s origin is equally described), and is distracting for readers simply trying to figure out what it is. It could probably be worked into the history section though. —jacobolus (t) 19:36, 20 June 2008 (UTC)

The article itself is named Comet (programming). It is not unreasonable to explain to the reader why it has such name, and where it came from, while also acknowledging it is not the exclusive term for, HTTP Push/Server push, etc. AtaruMoroboshi (talk) 19:46, 20 June 2008 (UTC)

I don’t think that is particularly beneficial to the lead section, being mostly a distraction from a technical summary of what Comet is, and I think it reads fine to leave it in a note, or in the history section. But it could potentially also be made into a short section immediately after the introduction. I suppose it depends on the audience is of the article. My expectation is that they’d care more about the definition than its etymology. How do you imagine the lead section would go, with such an explanation worked in? —jacobolus (t) 20:15, 20 June 2008 (UTC)

It doesn't explicitly need to be in the lead section, it could go in those other places as well. For example the AJAX article has similar info in the "history" section. The main thing is it doesn't need a note. AtaruMoroboshi (talk) 20:29, 20 June 2008 (UTC)

Canderson7 has addressed this in their most recent edit, discussed above. AtaruMoroboshi (talk) 03:27, 22 June 2008 (UTC)

Not necessary, text should go into article text or deleted if not relevant (no need for a "rants and trivia" section). I have tried to fix that in my recent edit -- 83.254.214.72 (talk) 21:10, 28 June 2008 (UTC)

Article no longer protected

The protection on the article has expired, and I have merged into it the contents of the sandbox. All future edits should be directed there (i.e. to the actual article). It is my hope that the cooperation we have enjoyed over the last few days will nonetheless continue. Canderson7 ^(talk) 20:45, 23 June 2008 (UTC)

I hope we could could still count with your help on this article!

Also, since we haven't yet fixed the whole article, shouldn't the cleanup tags ({{advert}}, {{cleanup}}, {{onesource}}, {{COI}}) be kept? They were improperly removed in the middle of the redit edit war. --Damiens.rf 22:20, 23 June 2008 (UTC)

I've implemented all of those tags with the exception of COI. I feel that the advert tag sufficiently addresses this article's POV problems, and I really don't feel that WP:COI, which describes an editor who forgo[es] advancing the aims of Wikipedia in order to advance outside interests, applies to this situation. Canderson7 ^(talk) 03:30, 25 June 2008 (UTC)

Connection vs Request

The article as it stands is very inconsistent with its usage of the term HTTP Connection vs HTTP Request. The reason for this is that comet articles and documents themselves are often lax in making the distinction between these two things.

HTTP/1.0 and HTTP/1.1 both support persistent TCP/IP connections, over which many request/response pairs may be sent without closing the connection. This is totally orthoganol to comet. Thus the definition of comet should not be about long held HTTP connections - as most browser that implement HTTP/1.1 will hold the TCP/IP connection open for long periods of time. Comet uses long held HTTP Requests! The definition of long-polling says: "After sending such an event, the server closes the connection, and the browser immediately opens a new one.", when it really should say: "After sending such an event, the server completes the HTTP response and the browser issues a new HTTP request to obtain further events" - not that there is no requirement of immediate as both DWR and cometd support configurable pauses between long polls.

I know this as a fact as a comet expert, but it is difficult to cite this as it is a very common mistake to for comet documentation to use the word connection, when request should be used. So the summary sections of many cited articles will say connection, but the detail will talk about request. This confusion is something that really should be pointed out... but I guess that would be OR. I don't want to update this page myself... as I'm probably too close to this. But I will seek to have an article published that discusses this frequent mistake and also search for other citations, so that this page can at least refer to the issue. If a neutral editor wanted to read the details of the cited articles and see that indeed it is requests held (not connections), then it would be great if they updated this page to reflect that. —Preceding unsigned comment added by Gregwilkins (talk • contribs) 06:55, 27 June 2008 (UTC)

Please just change the places where 'connection' should say 'request'. Those changes shouldn't change the verifiability of the content. —jacobolus (t) 18:17, 27 June 2008 (UTC)

Reorganization

I've just implemented a pretty major edit that I've been working on for the past day or so. Here's a quick breakdown of what I did and why:

I first tackled the "Web application architectures" section. This is the only section from which I actually removed content. My overall goal was to keep the article focused on Comet and to avoid vague and thus potentially POV wordings. I also reorganized this section so that it now focuses only on Comet's predecessors. Considering that Comet is the whole point of this article, ===Comet=== seemed somewhat incongruous. I instead formed a new section in the article called "Implementation of Comet". As I said earlier on this talk page, I wanted a reader to be able to look at the table of contents and know where to click to find technical information on Comet that is less niche-oriented than "horizontal scalability." I think this new section does that. Much of its content comes from the old Comet subheading. I also merged into it the material from the "browser compatibility" section, a section that was not really about Comet support in various browsers but instead about different methods for implementing Comet. I tried to give the new section structure by tapping the streaming vs. long polling dichotomy. However, my efforts in that regard could use some expert review. In particular, I wasn't sure how server-sent events should fit in (as is obvious from the fact that I gave it its own subheading). Anyway, you can see what I was trying to do and hopefully you can improve on my treatment of the topic.

Once I got down to the scalability section, I found that less and less revision became necessary. The technical material is well-written and comprehensive. The one thing that I did do was merge most of the notes into the article. Often a really cool piece of information was hidden away at the bottom of the page when it made just as much sense in the middle of the body. I did leave two of the notes, because I though incorporating them would disrupt the flow. I also created a note of my own. Throughout my editing, I removed peacock terms and POV words wherever I found them (such as "unfortunately" from the "Multipart XHR" section).

Overall, it's my opinion that the article is easier to use and less POV than before. I would appreciate it, though, if the experts would look over what I did and make sure that the accuracy of the article's content has been preserved. A couple other things that need attention are the following, in no particular order:

The pub/sub section seems a little out of place under horizontal scalability.
Several people are addressed by name in the article, but it is never made clear who they are or why we should care what they have to say. For instance: Michael Carter in the horizontal scalability section.
Should there be a section that clearly outlines disadvantages / criticisms of Comet?
When should Comet be italicized. Right now it sometimes is and sometimes isn't?
There remain instances of original research or uncited material that I have not yet had time to tag.

Those critiques notwithstanding, I think the article is coming along really well. I've removed the advert tag from the list of problems, because I think that most of the concerns in that direction were dealt with by this last edit. Please comment. Canderson7 ^(talk) 22:25, 27 June 2008 (UTC)

There is no particular reason Michael should be mentioned; I attributed that statement to him because despite being good advice, it is certainly not a universal point of view (at least judging from some systems people have actually built); Yes, there should be some integrated criticism, though I don't know whether it will be easy/possible to find reliable sources making well-argued criticisms; Comet should be italicized when the term itself is being discussed as a term, while it should remain roman when referring to the concept. I’ll actually take a look at your edit in detail in a few days, and comment on it. —jacobolus (t) 01:19, 28 June 2008 (UTC).

Best of luck. I'll have internet access again in a week. Hopefully things will have come along well. Canderson7 ^(talk) 17:17, 28 June 2008 (UTC)

I also gave it a try and mainly reviewed the technical parts of the article, plus tagging/removing some original research. When looking at the article today there will be work for weeks to come, in order to get it into a decent, neutral and informative state. The term 'Comet' in the article text should not be italicized in my opinion, this just confuses readers (if there are different opinion we should check WP:MOS). -- 83.254.214.72 (talk) 13:23, 29 June 2008 (UTC)

Damien.rf please don't just delete!

Damien.rf you are continuing to edit from your own point of view. Yes we all know that you think nobody uses the term comet and that there are no notable applications blah blah. But these are just your opinions and not sufficient for deleting text. You deleted the reference to JSONRequest because it is your opinion that JSONRequest people do not use the term comet. But if you typed JSONRequest and comet into google, you would see that many many people do associate the two. Note least Douglas Crawford when he introduced the proposal to the whatwg, when he said: "I am proposing a new mechanism for doing data transport in Ajax/Comet applications. It is called JSONRequest.". If you think something is unjustified, please just add ^{[citation needed]} rather than just deleting it and give the contributors time to justify it. Gregwilkins (talk) 06:28, 30 June 2008 (UTC)

I haven't deleted anything, just moved to other place in the article. Please, chill out. We're making progress here. --Damiens.rf 06:55, 30 June 2008 (UTC)

(A) There are no “JSONRequest people”—it is a proposed object not yet implemented or used by anyone, and (B) These people, should they exist, would not be able to replace the word “Comet” in any sentence where it made sense with the word “JSONRequest”. They may or may not use the word “Comet” (for all I know they prefer “HTTP push” or some similar term), but that is completely beside the point. —jacobolus (t) 10:11, 30 June 2008 (UTC)

Damiens.rf - you did not move it to another part of the document. You deleted the whole JSONRequest point from the list of attempts to "standardize non-polling HTTP communication". It is not so much your deletion that I object to, but the reasoning you gave. You justified your deletion with your POV that "JSON Request people don't call it 'Comet'", yet a simple search showed the very creator of JSON Request using the 'comet' word to describe it's intent.

I am chill and we are making progress here. Unfortunately you do not appear to be making as much progress as others and are continuing in the behaviour that resulted in the original dispute. When called out - you deny you did it! Can't you just say "oops, I should have researched that one more" and then we can move on? Gregwilkins (talk) 13:50, 30 June 2008 (UTC)

Reverse Ajax

Please note that Reverse Ajax is not the same as Comet. Reverse Ajax is any way to get asynchronous data from the server to the browser, and includes polling, comet and piggyback. For more information, see http://directwebremoting.org/dwr/reverse-ajax. JoeWalker (talk) 09:10, 1 July 2008 (UTC)

Can you please add a comment also to Talk:Reverse Ajax which Jacobolus wants to merge/delete (Quote from Jacobolus: everything said on this article [Reverse Ajax] is explained in great detail at the Comet page). - 62.127.211.2 (talk) 10:11, 1 July 2008 (UTC)

Since it "includes comet", and is an older term, then I guess it's correct to say that the "concept comet" was (among other concepts) once called reverse-ajax. --Damiens.rf 20:25, 1 July 2008 (UTC)

The content of a hypothetical perfect Reverse Ajax encyclopedia article has at least 90% overlap with a hypothetical perfect Comet article. I don’t think anyone benefits from completely writing that material twice. Any differences between the meanings of the terms can easily be explained in the “Comet” page. Therefore, the “Reverse Ajax” page should be merged and redirected. —jacobolus (t) 03:10, 3 July 2008 (UTC)

Can we please merge it with buzzwords next?. - 83.254.214.192 (talk) 22:38, 3 July 2008 (UTC)

Honestly, if you think that would provide a better explanation for readers trying to understand how real-time data can be sent to web browsers, then be WP:BOLD. I suspect other editors would disagree with you. I really have no attachment to the name “Comet”—if it stopped being used tomorrow, I wouldn’t bat an eye, though I think having some understood term for it is potentially useful—but I want Wikipedia to have a decent explanation of this technology, both accessible and comprehensive, and I want it under a single title, because many are curious about it, and go to Wikipedia as the first step when trying to learn about something new, and putting several low-quality versions of the same article with slightly different titles serves them poorly. —jacobolus (t) 00:31, 9 July 2008 (UTC)

If "Comet" really stopped being used tomorrow, would you change the name of your software from "Cometd" to something else? --Damiens.rf 03:44, 9 July 2008 (UTC)

Given that I have nothing to do with Cometd, I doubt I could change its name. But I think Orbited would be a fine name for a streaming web server, with or without the pun. —jacobolus (t) 04:58, 9 July 2008 (UTC)

Oops... my mistake. I could have asked "Would you prefer "Comet" to be called "Orbit"?"... but nevermind... I've lost this joke. --Damiens.rf 12:44, 9 July 2008 (UTC)

Hidden IFrame

I found about a book by O'Reilly that, among other things, describes the "hidden-frame" technique for building dynamic applications. There are other reliable books that talk about this. Fortunately, we no longer need to rely on 'cometdaily' blog-posts to source these passages in the article.

For now, I've replaced a piece about the 'downsides' of hidden iframes (that used to be based on Jacob Rus self-published original research) by content based on what is said on the published book.

Maybe more changes are needed in the paragraph to conform it with what is said on reliable sources. --Damiens.rf 16:09, 3 July 2008 (UTC)

"First Comet applications"

I'm worried about the section in the article about the First Comet applications. It says that some applications "exposed Comet to a wider audience", and mentions Gmail, Jotspot, Meebo and a lesser known service called Renkoo.

Unfortunatelly, we currently have any reliable sources to support the statements that popular applications like GMail, Jotspot or Meeboo use "Comet". We do, however, have a link to a press-release-like article mentioning the usage of "Comet" by the non-notable service Renkoo. But I'm not sure Reenkon is capable of exposing anything to a "wider audience".

Indeed, we used to have a link to Comet creator Alex Russell's blog where he writes about his impressions on GMail, but I don't think that qualifies as a reliable independent source for saying "GMail uses Comet".

However, I remember seeing some reference to an official Google blog where Comet was explicitly mentioned, but I can't remember where I've seen this post. Can anyone find that (or something like that)?

Also, the article used as a reference for Jotspot doesn't mentions the use of Comet at all. And the Meboo claim goes completely unsourced.

How can we improve this section? Where can we find independent reliable sources confirming that popular applications ("consciously") use Comet? --Damiens.rf 15:04, 21 July 2008 (UTC)

The biggest problem I see here is that some people describe their product with something "Comet-like", often in the same sentence with Ajax. Now that leaves the room open for speculations what was ment. Since it is often unclear what Comet means (and Comet tries to be so many things) some people use it as "Ajax with asynchronous notifications" (e.g. long polling since it is the most reliable implementation in terms of browsers, proxies and firewalls).... on the other side who really knows without looking into the code. - 62.127.211.2 (talk) 10:17, 23 July 2008 (UTC)

I see this problem too. But since "comet" is just a new name for something that is as old as web-programming, I believe we should only call something "comet" when the creator himself calls it "comet". Otherwise we will end up calling a lot of code "comet", even when the code predates the coining of the term, or when the developer prefers the use of other terms (see the IceFaces mention in the article). --Damiens.rf 12:12, 23 July 2008 (UTC)

I don’t care what you call it. But since this is the page which describes the technology, those web applications which *use* said technology should be discussed. What the application developer prefers as a name has little bearing on whether he is using the same technical building blocks (note, these developers are aiming their sites at end users who don’t care whatsoever how the site works, and therefore don’t put out detailed technical descriptions; that they don’t publicly name the technologies they use is irrelevant to this discussion, and serves as evidence of nothing). —jacobolus (t) 12:19, 23 July 2008 (UTC)

What the application developer prefers as a name does have a weight when we're talking about a buzzword/neologism, especially an ill-defined and all-embracing one as "comet". --Damiens.rf 13:14, 23 July 2008 (UTC)

We’re writing an article about a technology/set of techniques/architecture/pattern/whatever you like to call it. If developers use that pattern, it is absolutely relevant to the article, even if one developer chooses not to publicize it using the same buzzword as another developer (or any name or technical description at all). There are certainly extremely credible sources which say “Google chat uses Comet”. And as far as I can tell, there is no one (besides you) who disputes that. It is a verifiable “fact”. —jacobolus (t) 09:15, 24 July 2008 (UTC)

By all means, please list these sources in this talk page and I would be glad to add them to the article. Of course, I'm assuming when you say "extremely credible sources" you really mean reliable independent sources (as opposed to self-published blogs by self-proclaimed "experts on the field") --Damiens.rf 17:20, 24 July 2008 (UTC)

“Comet-like” == “Comet”. “Ajax with asynchronous notifications” == “Comet”. Reading these descriptions, it is usually completely obvious what is meant. If you have an example which leaves room for speculation, please be explicit (paste the offending description here, or name the product specifically). Otherwise I would classify your comment as FUD. It is quite easy to see exactly what is being used (no need to examine code) by opening up a network monitor (e.g. Firebug running in Firefox) and noticing the form of the communication. You are correct that most Comet-using sites/products use long polling. Some, such as Google’s Gmail Chat, use streaming instead. —jacobolus (t) 12:24, 23 July 2008 (UTC)

I'm affraid opening up a network monitor and mind-parsing the network communication dump can't be considered a verification of trivial information. That is to say that we need verifiable sources for the claim that gtalk, meboo and the like use "comet". --Damiens.rf 13:14, 23 July 2008 (UTC)

I don’t think you understand. A somewhat analogous situation: A country has an election, and a wikipedia article mentions that they voted in the election, and you say “STOP THE PRESSES! No one said they voted! They just said it was an election! Who knows whether they actually voted, it might have been any kind of election! Just because you were on the ground and watched people there vote, doesn’t mean it happened. You need a reliable source that says it was a vote!” It’s a simply ridiculous argument to make, and certainly has no grounding in Wikipedia policy. It makes it incredibly frustrating to attempt to collaborate with you if you replace every substantive reference with a link to a completely unrelated (and for readers’ purposes useless) book chapter, and if you remove all substantive discussion and replace it with your own POV commentary on how bad a term you think Comet is. The article is at this point vastly less informative, less neutral, and less substantiated than it was a few months ago. I'm not sure where you’re going with your edits, but my impression is that your end goal is to strip it back down to nothing, paragraph by paragraph, sentence by sentence. —jacobolus (t) 09:27, 24 July 2008 (UTC)

I believe the problem relies on what you consider a "substantive reference" and what Wikipedia accepts as a references. --Damiens.rf 17:20, 24 July 2008 (UTC)

"The article is at this point vastly less informative, less neutral, and less substantiated than it was a few months ago." - Jacob, at some point you'll have to acknowledge that the article you single-handled wrote was completely unacceptable for any encyclopedia. There's no way in the world that it could become less neutral (you even conceded that it was written as an advertisement for "comet"). I'm, affraid I no loger have the willingness to explain you why it's wrong to cite your own blog and those of your co-worker as if they were reliable sources. How many critics does it take for you to admit that the article you wrote was unsuitable by any standards? Stop being nostalgic about that beast and lets work on the far better work-in-progress we have now.

The book chapters I cite are not "completely unrelated" nor "useless". They are straight-to-the-point and reliable. Of course, they don't always support what you originally wrote on the article, since you didn't base your witting on published sources. And since they are from respectable publishers, you should also don't expect them to share your enthusiasm or mystification about comet.

My "end goal" has never been to strip the article from its content. But instead, to have an neutral article, containing only information attributable to realiable sources. Of course, this may mean an article completely apart from the one you wrote months ago. --Damiens.rf 17:20, 24 July 2008 (UTC)

I don't answer Jacobulus comment, but if you want to engage in a civcil and productive discussion you should change your attitude. - 62.127.211.2 (talk) 16:43, 23 July 2008 (UTC)

How can you not answer the question? It was a very simple question. Do you have any concrete examples? As it is, your criticisms are vague and unsubstantiated, so you leave me no way to make a substantive response. —jacobolus (t) 09:15, 24 July 2008 (UTC)

Try being less confrontational and people will be more willing to answer your questions --Damiens.rf 17:20, 24 July 2008 (UTC)

This a reliable enough source for you? “How Comet Brings Instant Messaging to meebo” by Howard Wen at news.oreilly.com. Hey look, it even links to this article. Also, hey look, it provides a source for several paragraphs of text which were ripped out of this article without justification sometime in the past two months. —jacobolus (t) 23:39, 27 July 2008 (UTC)

I'm not sure that is peer reviewed, since o'reilly itself classifies that as a "blog"-post, instead of as an "article". --Damiens.rf 01:02, 28 July 2008 (UTC)

You are not paying very careful attention. We are trying to establish that “meebo” is “Comet”, as you have challenged that claim. If a “blog” interview by a reputable tech journalist at an O’Reilly website with employees of Meebo claims that it is Comet, that is more than reliable enough. Newspapers/magazines/&c. are in general not “peer reviewed”, but “peer reviewed” is not the same as “reliable”—this isn’t rocket science we’re dealing with here. You are trying to apply strict “rules” without understanding the purpose of those “rules” (verifiability), or taking into account the context. The effect is procedural stalling which prevents progress. —jacobolus (t) 01:42, 28 July 2008 (UTC)

It might not be the best reference, but I'd count it as one (the purpose of references is verifiability not truth). - 62.127.211.2 (talk) 13:48, 28 July 2008 (UTC)

Yeah, it's probably ok for sourcing that meebo uses "comet", but considering the conflict of interest problems we had with this topic, I'm affraid adding this blog reference may (re-)open a can of worms...

And in a completely unrelated topic, isn't it very interesting that this blog post was posted on July 23, just as we were having this very discussion? --Damiens.rf 15:48, 28 July 2008 (UTC)

Stop the presses! I've found some references for Meebo+comet on the official blogs for meebo developers. I'll work them to the article later, since I can't access meebo sites from work. In short, yes they use "comet" and acknowledge that, although they make it clear that they have been using it long before the term was invented and that it's a crucial buzzword for describing meebo workings. --Damiens.rf 16:38, 28 July 2008 (UTC)

Okay, here’s an example of the kind of know-nothing edit Damiens has recently made. He removed the sentence explaining that when using the XHR object with multipart streams, it is impossible to determine when the request has been closed. The reason this is a significant removal is that this downside makes the transport all-but-useless for production use by Comet developers. The justification given by Damiens was: “(That the Alex guy can't determine when it has been closed, it doesn't follows that "it's currently impossible to determine" (Argumentum ad ignorantium. Anyway, this is almost trivia, so removing.)” Now the problem with this is that the question is not what “This Alex guy” can or can’t determine: the browser simply does not give developers any access to this information. There is no interface for retrieving the information, no legitimate way for it to be obtained. This is not an an argument from ignorance. (Why the patronizing wiki-linked Latin phrases? I find the condescension quite insulting.) It is simply a factual statement, made by a recognized expert in the field, and then verified by other recognized experts in the field (and also easily verified by anyone who is capable of downloading Firefox, writing a trivial HTTP server, and doing some testing). It is not a statement disputed by anyone (except Damiens). If you like, Damiens, you may as well remove all discussion of using multipart with the XHR object altogether, because without the caveat, the information is misleading to aspiring Comet developers. —jacobolus (t) 09:43, 24 July 2008 (UTC)

For the 1000006th time, Wikipedia's articles can't rely on "experts in the field". Every non-trivial information must be attributable to a reliable source. A piece of knowledge is not trivial if one needs to be "capable of writing a trivial HTTP server" in order to verify it.

And in any case, even if Alex can’t determine something, or you can’t determine something, and every http-server-writer can’t determine something, it doesn't follows that this "something" can't be determined. This is the very essence of an argument from ignorance (sorry for the Latin, my mother-language is Latin based).

For instance, so far, no one (even Alex!) knows how the Flagellum evolved. But that doesn't make Creationism a fact. --Damiens.rf 17:20, 24 July 2008 (UTC)

Yes, but Firefox is a piece of open-source software, whose moving parts we can rather precisely understand, and its XHR object provides no access to the information that a request has been closed. If you were a web programmer, this would be immediately clear to you after reading his explanation. Essentially, Firefox has a mechanism for knowing if any XHR object has finished. The problem here is that after every part of the multi-part response, that mechanism is triggered, with no distinction between intermediary and final parts, and no other mechanism if the connection is severed. This is not something that “Alex can’t determine”. It is simply the way that Firefox works. —jacobolus (t) 21:48, 24 July 2008 (UTC)

Damien - You are not going to find and references to say "Gmail uses comet". Comet is not something to be used, but a label that describes a set of techniques. Open Source was a phrase coined to describe a flavor of Free software. Most Free software is indeed open source, but they do not self describe as such. Emacs is open source, but self describes as free software. Another example that comes to mind is that a vandal is unlikely to self describe as such, but we recognize vandalism when we see actions that match the definition of vandalism.

So just because Gmail does not self-describe as comet, does not mean it is not a comet application. Indeed it was the needs to say things like gchat-like that inspired the coining of the term comet.

With regards to the issue of your removal of multi-part streams. For the 100000006th time - if you are unhappy with the citation given, can you just insert a ^{[citation needed]} and give others time to provide one. You continue to simply delete text without reasonable notice or discussion.

Gregwilkins (talk) 05:06, 25 July 2008 (UTC)

Please make yourself more familiar with Wikipedia policies before telling other editors what to do. Let me quote "just because Gmail does not self-describe as comet, does not mean it is not a comet application". You understand that when a self proclaimed Comet advocate puts the "Comet-inside" label on a product that other editors have the duty to verify this statement, verifiability is mandatory for encyclopaedic entries and the burden burden of evidence lies with the editor who adds or restores information. Under Wikipedia policies this could mean to aggressively remove unverifiable statements, to add a [citation needed] marker (e.g. when it makes collaboration easier) or to discuss the issue together on a talk page. Now, if we can not find a reference that Gmail uses "Comet", than this is original research and should be removed, period. For the 100000006th time.... :) - 83.254.214.192 (talk) 11:37, 25 July 2008 (UTC)

Exactly. If we can't find a reliable source, stating the information is original research. Gregor, the source doesn't have to be a self-description (btw, where did you take this idea from?). Just like we can find reliable sources claiming that e-macs fits the Open Source Definition (e.g.: ISBN 0596006489) we should be able to find some reliable source saying "Gmail uses comet". --Damiens.rf 13:30, 25 July 2008 (UTC)

So the wikipedia policy is to work toward a consensus versionWikipedia policies. [citation needed] marker suggested by the Wikipedia policies policy as a way to make collaboration easier. So one would think that collaboration is a good way to get to consensus? No? Specially for an article that has been protected and much under dispute. Aggressive deletion is for unverifiable statements. Did either of you two (IP dude or damiens.rf) attempt to verify the statement? Did you try searching or researching? It's pretty easy to find lots of usage of the word comet applied to Gmail. Not least in the very article that coined the term "comet", which cited Gmail/gtalk as a comet example. So Gtalk is comet by definition! here is some reportage from JavaOne 2008 talking about comet BOFs and describing GTalk as a site containing a comet application. There are plenty of other examples of term comet being applied to gmail/gchat in other blogs (for example), but I know you will discount all blogs. So here is a peer reviewed article that uses Ajax Chat as the example for comet applications - and gtalk is an Ajax Chat accorind to Ajax Magazine. Gregwilkins (talk) 09:54, 1 August 2008 (UTC)

So what you say is that you haven't found a single good reference that names Comet and Gtalk, only blog postings. - 83.254.214.192 (talk) 10:54, 1 August 2008 (UTC)

Exactly, Greg. I also missed the point of your post. Your links actually seem to show that the Comet enthusiasts like to associate gtalk to comet, but that association can't survive peer review.

By the way, I believe I can find reliable sources for such association (if they exist), and I'm (slowly) working on this. Stay tunned. --Damiens.rf 12:29, 1 August 2008 (UTC)

What is the criteria for a product/site being mentioned here?

As previously mentioned, I am from Caplin Systems. Until recently there was a sentence or two in the history section mentioning our product. I think the edit comment was something like 'its was nothing special' (one persons opinion). This was in the history section, and (although i am obviously biased) i think we were significant in the history of comet. The java applet section is pretty biased anyway, with various descriptions on how bad applets were (which is all opinion) - which was why the mention of us using applets merely for the communication was important and links in nicely with the following sections on pure javascript implementations - it shows the progression of the technology. 88.96.214.6 (talk) 08:50, 5 August 2008 (UTC)

Sorry, didnt realise i wasnt logged in, above comment is by me. Martintyler (talk) 08:52, 5 August 2008 (UTC)

If you are related to a product/company please see Wikipedia:Business' FAQ. Maybe you could suggest an updated section here on the talk page? - 83.254.214.192 (talk) 13:16, 5 August 2008 (UTC)

I am aware of that FAQ, I was asking about this particular page, since some stuff is deleted without discussion and others stay. I don't want to write a piece about how Caplin should be included, I thought letting someone else do that was more in keeping with Wikipedia, and that is what happened, but then it got deleted based on one persons opinion on it being 'nothing special' - I was wondering what was special about the companies/products that remain. —Preceding unsigned comment added by Martintyler (talk • contribs) 13:30, 5 August 2008 (UTC)

Martin, I can't answer your question about the history section and which products should be mentioned there. Why not make a suggestion on how the history section should look like (from your perspective )... if your product is about Java pushlet maybe you can suggest it there as well. - 83.254.214.192 (talk) 13:47, 5 August 2008 (UTC)

The 'early push applications' section used to be something like 'early comet' which is why Caplin was mentioned in the Java section, since some people prefered to keep 'comet' to javascript only. This change of title now makes it inaccurate, Caplin Liberator was an 'early push application' (using a java transport) before 2000, so deserves a mention in this section if the heading is to stay as it is. It seems someone has decided that the term 'comet' can only apply to things since the term was coined, which is why the headings have changed. The sentence about 'some people switched to pure javascript' which used to also contain a reference to Caplin is now probably in the wrong place too (since it happened before 2006) - and arguably a bit redundant without the reference to Caplin that was there before anyway.

The current java applet section seems to imply that communication was blocked and nothing really worked.. removing the reference to Caplin from there is missing a chunk of history, which was (and i dont mean just down to us) the HTTP tunneling side of Comet/Push.. you've gone from applets trying to do direct socket comms and failing, to pure javascript implementations magically working. It implies that using native browser technologies somehow made communications with the server work where it previously failed. Martintyler (talk) 14:20, 5 August 2008 (UTC)

[1] ttp://alex.dojotoolkit.org/?p=545

[2] ttp://alex.dojotoolkit.org/?p=545

[1]

[2]