User talk:Oberstet

Conflict of Interest on WebSockets comparison

An editor has questioned my contributions to the page Comparison_of_WebSocket_implementations because of COI. See also the conversation below.

The information here is to allow other Wikipedia users / editors judge for themselfes

whether I have a COI and
whether my contributions are non-objective/non-neutral.

The comparison table in Comparison_of_WebSocket_implementations contains a row for "protocol compliance test reports". The purpose of the row is to provide detailed information about the conformance of a WebSockets implementation to the protocol specification. Conformance to the specification increases interoperability between implementations, which I think is of particular importance to WS to find broad adoption and fulfil the promise to end users (more reactive apps ..). Hence, the conformance of WS implementations is valuable information to users (developers).

I am the author of a WebSockets protocol test suite that can do automated tests of client and server WS implementations, and above row contains links to test reports produced with that test suite. The links point to reports produced/hosted by others or by me. The test suite is part of an open-source (Apache-licensed) project (Autobahn WebSockets), which is freely available to anyone to use. Or not use. I have no commercial interest in that test suite whatsoever.

Further, that open-source (Apache-licensed) project also includes a WS client/server framework on it's own. And for that latter part, the company I am working for (Tavendo GmbH) may offer commercial support/services. Or not. Since it is quite a side-project for us (as company).

The test suite is designed to cover the WS spec. It has >200 test cases. It is not designed to favor any specific implementation. It is designed to produce objective, useful information regarding protocol conformance.

The test suite found several bugs in browsers, and I got positive feedback from several WS implementation authors saying how the test suite helped them improve their implementations. The test suite is continued to be used by several independent projects, which I think is a clear sign that the results are useful/objective (otherwise they wouldn't use it).

I help people using the test suite for free. I help browser developers by testing and filing bugs. The interest I have in this is to help making WS implementations interoperable, and thus help in WS adoption.

So I admit:

I am a subject-matter expert
I help the WebSockets community
I want to see WS being broadly adopted

Regards, Tobias

Emails about WebSocket article

Email 1

To: AlistairMcMillan Subject: WebSocket From: Oberstet Date: Sat, 17 Sep 2011 17:18:15 +0000

Well, taking away information useful to people is never helpful. I disagree that a list of WebSocket implementations falls under the WP:LINKFARM policy. Maybe this kind of blind following of bureaucratic "rules" has fueled the bad reputation that Wikipedia seems to take up in recent time.

I spare you some time: here is another page for you to rip apart: http://en.wikipedia.org/wiki/Comparison_of_WebSocket_implementations

Disclosure: I started that page and invested considerable time in collecting information. It is obviously not complete, and also "a mere collection of some useful" bits. So if that doesnt comply to some "rule", just tell me, I'm happy to contact the authors, host it myself, and stop working on Wikipedia.

Regards, Tobias

Reply 1

To: Oberstet Subject: WebSocket From: AlistairMcMillan

Please read http://en.wikipedia.org/wiki/Wikipedia:USEFUL

Wikipedia is an encyclopaedia. It is supposed to explain things. It is not just supposed to be a list of links to external sites. It is definitely not supposed to be a venue for people to advertise their newborn half-implemented projects.

If I had a penny for every time someone told me one of my edits was indicative of the end of Wikipedia, the moral decline of our civilisation and/or the heat death of the universe, I'd be a rich man.

Email 2

Date: Sun, 18 Sep 2011 17:24:30 +0200 Subject: Re: Wikipedia e-mail From: Tobias Oberstein To: AlistairMcMillan

> Please read http://en.wikipedia.org/wiki/Wikipedia:USEFUL

So what? How does that apply to your deletion of actually useful information?

> Wikipedia is an encyclopaedia. It is supposed to explain things. It is not just supposed to be a list of links to external sites. It is definitely not supposed to be a venue for people to advertise their newborn half-implemented projects.

I fail to see how you are able to judge the maturity of dozens WebSockets implementations written in different languages and environments. Are you an expert in the field?

> If I had a penny for every time someone told me one of my edits was indicative of the end of Wikipedia, the moral decline of our civilisation and/or the heat death of the universe, I'd be a rich man.

Don't take yourself so important.

As I see it, you are likely not an expert in the topic involved, and deleting information you can't even judge is not helpful. It's not the end of the world.

Reply 2

1) Please stop sending me emails. If you want to discuss this either use the article talk page Talk:WebSocket or if you want to contact me directly please use my talk page User talk:AlistairMcMillan.

2) Stop making judgements about my knowledge on a particular topic. You know nothing about me. If for no other reason, because it makes you look bad.

3) You shouldn't even be editing these articles as you clearly have a conflict of interest.

AlistairMcMillan (talk) 16:24, 18 September 2011 (UTC)[reply]

Discussion

You: "It is definitely not supposed to be a venue for people to advertise their newborn half-implemented projects."

How do you know when a WS implementation is immature? Did you check all? Who's judging?

COI: I am author of a WS test suite used in a couple of projects. The test suite has no inherent interest other than verifying the conformance of implementations. The test suite would not be used by others if it wasn't "neutral". I have started the "Comparisons WS" page, and contacted every single implementation author by email about the "feature status" before setting the page public to get consensus. Meanwhile, a couple of other guys have added their stuff.

Where is the COI you claim I have? This talk of COI seems to suggest that I deliberately published false information to back "my interests". Obviously, I cannot let this false claim stand as is. — Preceding unsigned comment added by Oberstet (talk • contribs) 18:06, 19 September 2011 (UTC)[reply]

You aren't the managing director of Tavendo then? You aren't the lead developer of Autobahn? You don't see a conflict of interest in your being the lead author on an article comparing Autobahn to other WebSocket implementations? You don't see a conflict of interest in us linking to test results of a test suite you developed? You don't see a conflict of interest in you adding a link to your Github repo which advertises your own "Commercial support and services..."? Are you kidding? AlistairMcMillan (talk) 20:11, 19 September 2011 (UTC)[reply]

Look at the edit history of the WS comparison page. The information about most implementations (besides browsers and Autobahn) have been edited by the authors of the respective implementations. For those edits I did myself and have been for implementations other than browser/Autobahn, I have collected the information by email (I can forward those if you have doubts). With regard to the browsers, I am in email contact with Firefox and Chrome developers - and they value the feedback from the test suite .. fixed several bugs in both browsers (again, I have the mails).

Further, you need to differentiate between the WS test suite included in Autobahn, and the WS client/server framework for WS that Autobahn provides. As said, I doubt the test suite would be used by multiple projects if the tests were in any way tailored to favor Autobahn (the client/server impl.). Look at the edit history for Jetty. The author (Greg Wilkens) uses Autobahn for testing. He asked me to publish the results on www.tavendo.de. I did so, even though it's clear from the test results that Jetty (after fixing some bugs Autobahn found), passes all tests and is significantly faster than Autobahn. So this is my conflict of interest. We are publishing test results for a "competitor" on our site.

Really, get over it. This is a community edited page. And that community, including myself, want it to be accurate. Should you find something inaccurate or missing, you're welcome. — Preceding unsigned comment added by Oberstet (talk • contribs) 14:35, September 21, 2011