Jump to content

Talk:Acid3/Archive 1

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

This is an archive page for Talk:Acid3

pictures

[edit]

How about to add the "reference-picture" in ACID2 and add the other browsers status? —Preceding unsigned comment added by 84.169.243.167 (talk) 14:10, 6 February 2008 (UTC)[reply]

Plesese, someone use a standard pixel and upload to common. Matthew_hk tc 11:46, 10 February 2008 (UTC)[reply]
Do we really need a new picture for every nightly build of webkit or any other engine? If someone could help me improve the article instead that would be so much better.--itpastorn (talk) 08:01, 12 February 2008 (UTC)[reply]
Yes, if the current nightly has improved test results over the previous.--toehead2001 (talk) 08:09, 12 February 2008 (UTC)[reply]
No. It is recentism. These screenshots are not notable over time, nor verifiable. --Fenring (talk) 09:54, 12 February 2008 (UTC)[reply]
You make a good point; I agree.--toehead2001 (talk) 17:03, 12 February 2008 (UTC)[reply]
The screenshots are easily verifiable – just download the software and run it. Whether or not we want to include screenshots of beta software is another issue. —Remember the dot (talk) 17:59, 12 February 2008 (UTC)[reply]
But still not of encyclopedic value. Neither is it of any such value to update them every day.--itpastorn (talk) 22:13, 12 February 2008 (UTC)[reply]
No need for an everyday picture update, but having a small reference about how browsers do on the test is quite useful IMHO. --HeffeQue (talk) 02:47, 13 February 2008 (UTC)[reply]
Verifiable means the claim has been published by a reliable source. Downloading the software and testing it is original research. --Fenring (talk) 10:52, 16 February 2008 (UTC)[reply]

Reliability of the pictures

[edit]

I just opened the test in FFox 3 because I wanted to source my Ian Hickson quote in the article. 2.0.12 got 52 - not 50. This illustrates how pictures really should be coming from someone with formal testing experience, an not from us random Wikipedians. (Random? Yes, I do know quite a bit about this subject, but I've not done any formal tests.)

Improve the article by finding such pictures, from stable browser releases, would be very helpful, of encyclopedic value and not original research.--itpastorn (talk) 13:19, 16 February 2008 (UTC)[reply]

I am now reverting

[edit]

Any addition of images from non-stable browser releases. Some people are adding them and seemingly do not read read this talk page. Until arguments in line with WP policy are put forward that demonstrate the encyclopedic value of these images, I will continue this habit.--itpastorn (talk) 20:12, 17 February 2008 (UTC)[reply]

The test is final - confirm scores (and replace images)

[edit]

This is what I get on Win XP. Please confirm or comment my scores:

Opera 9.26
47/100
no boxes at all
MSIE 7.0.5730.13
12/100
And that horrible, unreadable image
Firefox 2.0.12
50/100
Decent images only gray boxes

I have not tested in stable Safari (on Mac) since the test went final, Safari 3 beta on Windows gets 40/100 and no gray boxes

Neither have I tested Konqueror 4.0.1 yet, but I have no reason to distrust 61/100 as stated in the article.--itpastorn (talk) 17:48, 4 March 2008 (UTC)[reply]

Howdy, I uploaded the Acid3 Test screenshot with the current score of 52/100. Not sure how you wanted to link it up. Image:Acid3 ff20012 score52.png MarsInSVG (talk) 17:40, 10 March 2008 (UTC)[reply]

Attitude of layout engine teams

[edit]

If someone could verify the following - which is my OR - it would also perhaps be of some benefit. It really seems that Dave Hyatt/Apple/Webkit see the Acid3 as a way to advertise themselves. The fury that they have shown in fixing bugs, compared to Mozilla - mimicking the behavior of both teams for Acid2 - is somewhat telling. And even though Konqueror has supported all CSS3 selectors for quite some time, so was it not until those selectors showed up in Acid3 that Webkit started implementing what was lacking.--itpastorn (talk) 13:19, 16 February 2008 (UTC)[reply]

Good for them! All these means is that the Acid tests are helping promote web standards and browser innovation. Masterhomer 22:25, 3 March 2008 (UTC)[reply]
It is not a comment I wrote as criticism or endorsement of any team. It was about the impact of the test. It was about trying to improve the article. I would like to write about it, but have no sources. So how can I do it without it being original research?--itpastorn (talk) 17:52, 4 March 2008 (UTC)[reply]

Dave Hyatt, et al, has started blogging about their efforts to make Webkit pass Acid3. Posting the link here until I rework the article text. Surfin Safari: Acid3 --itpastorn (talk) 18:36, 8 March 2008 (UTC)[reply]

adding images from beta releases

[edit]

There is no reason not to add late beta images. Acid3 isn't even complete, they are still adding 16 more tests to it, so all of the images will be wrong in two-three weeks anyway. So if we follow the rational of not adding beta images, we should get rid of all of the images except for the test rendering.Austin512 (talk) 20:42, 17 February 2008 (UTC)[reply]

Your argument is precisely a good reason for not adding more non-perennial content to this article. I think a timeline (as in Acid2) would be more appropriate than screenshots. And afaik, the test is in a "final review" state. --Fenring (talk) 23:20, 17 February 2008 (UTC)[reply]

Also, Acid3 renders a 50 on Windows machines running FF 2.x, on Mac's running FF it renders a 52. Austin512 (talk) 20:42, 17 February 2008 (UTC)[reply]

I think the screenshots should stay, since they're fairly stable and give a feel for how various browsers perform. They aren't finalized yet, obviously, but they are still useful. —Remember the dot (talk) 05:18, 18 February 2008 (UTC)[reply]
This is an answer to Austin512's comment when he (she?) removed all pictures with the following explanation: "Seeing as though Acid3 isn't finalized, and there are still 16 tests to be added, the screenshots aren't of encyclopedic value because hey will soon be meaningless."

Technically that is part true and part untrue. The lasts 16 tests have been added, and I have put information about their contents in the article. All 100 tests are in place, but they are not officially declared stable. Having images - with appropriate disclaimers - is therefore of some encyclopedic value (IMHO). It will give a glimpse of how the test will look in a browser of today. What I am against is updating those images with every new release of WebKit nightly or Minefield. Since:

  • The exact score is not certain (indeed I got a 52 on FFox on Windows once).
  • The results are not guaranteed to come from a professionally conducted and scientifically controlled test.
  • Even though I prefer open source software, it will give misleading impressions. We have no idea what Opera or microsoft are doing behind the scenes. Daily updates are therefore misleading.
  • Daily updates is about tracking progress. That is recentism. Daily updates of this kind are an appropriate topic for a blog, but not for an encyclopedia.
  • And yes, the test may change as well. Probably only slightly, but it may affect the score.

Summary: Keeping the images from stable software is a reasonable compromise between the no images and every possible image camps.--itpastorn (talk) 09:06, 18 February 2008 (UTC)[reply]

HI : french and new on wikipedia... : I would like to add my picture which is not one of the Konqueror beta-version. You can find it at http://rom1dep.ifrance.com/desk/score.jpg I would like to know wether or not it's correct to publish it on this page... thanks

Images are best uploaded to Wikimedia commons. See Wikipedia:Images for more instructions. If you have a screenshot taken in a controled environment I think we will allow it among the rest. It is still formally original research though. But until Acid3 is finalized this is the consensus.--itpastorn (talk) 20:44, 2 March 2008 (UTC)[reply]

adding betas

[edit]

I agree that adding betas and prereleases are recentism and shouldn't be added. Adding misrenderings of stable browsers aren't recentism because in a few years the browsers will still be in use, and the screenshots will still be valid representations of misrenderings. Current betas and prereleases won't be in use in a years time, and so adding screenshots from these releases don't have encyclopedic value.

The fact the the test isn't finalized doesn't matter much (I know I'm flip flopping here) because the images are only examples of how it could be misrendered with respect to the example rendering. Even if the final release of Acid3 produces slightly different renderings, the current examples will be close enough, and will become slightly outdated as the browsers become updated anyway.

Austin512 (talk) 07:26, 20 February 2008 (UTC)[reply]

Proposed solution for betas

[edit]

As the test is final, I think we should have some table (and perhaps screenshots) of available development releases. But in order for this to have encyclopedic value I think it should be from nightlies from the same date the test was released.

Under the heading compliant applications I suggest a table like this:

Date Browser Type Score Notes
2008-03-03 MSIE 7 Official Release 10 Public betas not available
2008-03-03 Konqueror 4.0.2 Official Release 64 What is the score from the latest developer build? Notes on the rendering...
2008-03-03 Opera Build 9815 Non public build 65 Descent rendering. Does not download font. Does not implement text-shadow. Top left iframe ("FAIL") not hidden.
2008-03-03 Minefield (Firefox) Non public build 67 Descent rendering. Does not download font. Does not implement text-shadow.
2008-03-03 Webkit 525.12 Non public build 87 Descent rendering.
2008-03-05 MSIE 8 beta 1 Beta 17 Has parsing and major rendering issues [1][2][3][4]--itpastorn (talk) 20:59, 5 March 2008 (UTC)[reply]

Perhaps we could add Official releases and public betas (as of 3 March 2008) as well. But again, that info is in the images above (provided we put in correct alt attribute text for visually disabled readers).--itpastorn (talk) 20:48, 4 March 2008 (UTC)[reply]

I found this source as to today's "standings": http://ivan.fomentgroup.org/blog/2008/03/04/acid3-and-konqueror/--itpastorn (talk) 22:29, 4 March 2008 (UTC)[reply]

I think a table should be added however, I would change the order of the headings slightly, group them by: Browser Name/Build, Release Type, Rendering Engine, Date Last Checked, Score, Notes. Essentially we are only interested in the most recent scores for stable/beta/alpha builds, we are not looking to document the date each browser made a step forward. Fair enough documenting 100% completed tests. --80.6.152.199 (talk) 00:58, 8 March 2008 (UTC)[reply]

Personally I prefer to have the standing on the release date of the test, as in my suggested table above. Maybe we should have the very latest results as well - I do not see the encyclopedic value in that though. I do agree that we should not add a new row for every new point scored in a nightly build. That would be ridiculous. Rearranging and renaming the columns is OK by me. However, I think it might be of some value to have them in sync with the table in the Acid2 article.--itpastorn (talk) 16:36, 8 March 2008 (UTC)[reply]
It makes even less sense to list how well development releases do on the test, but not update them as they are developed. If there is any sort of table, it should have one row per browser (which lists the score on the latest stable release) or two rows per browser (one for the latest stable release and one for the latest preview release). Alternatively, we could have two tables, one for stable releases and one for preview releases. Currently, all this information is in the article, and personally I think it would be more readable if it were presented in table form, instead of awkward paragraph form for Compliance efforts and just screenshots for Non-compliant applications. At the very least, if all the information is left in the article, information about stable and preview releases should be put into the same form, whether that form is a table or screenshots. -- Schapel (talk) 16:49, 8 March 2008 (UTC)[reply]
Steve, the reason not to mention preview realeases scores is that the informations need more updates. And it's against wp's guidelines which say informations must be notable over time. The risk is that users don't update the values at each preview release. They do now (despite the comments), because Acid3 is in the spotlight. But will they keep doing this ? See, e.g this table is not updated very often. We don't want an article with obsolete informations. However, I think we won't prevent users to add preview scores, for the moment. So your proposition (one table, two columns) makes sense. --Fenring (talk) 17:54, 8 March 2008 (UTC)[reply]
@Fenring: Hear, hear! @Schapel. There is a reason why there is nothing but screenshots so far, and that reason is that no one has bothered to write any encyclopedic text! I will stub-mark the sections. Seems I get to do it as well (right now I consider myself the main contributor to this article). I don't mind but I've not got 26 hours to the day.--itpastorn (talk) 18:29, 8 March 2008 (UTC)[reply]
You may have misunderstood me. Instead of "encyclopedia text" we should put the information about stable releases in a table. If we decide to include information about preview releases also, that information should also be in a table in the same format, if not in the very same table. Perhaps I should put all the information in a table, then we can discuss whether it's readable and whether it will be kept updated. -- Schapel (talk) 18:40, 8 March 2008 (UTC)[reply]
I will proceed with the plan to put the stable and development releases in a consistent format in one table, instead of having the compliance effort and non-compliant applications in two different sections with inconsistent formats. If anyone has any unaddressed concerns, please speak now. -- Schapel (talk) 13:39, 11 March 2008 (UTC)[reply]
Go for it. Don't be dissuaded or distracted by people who don't understand the difference between "recentism" and "up-to-date information"; our readers are interested in these details, and they're the ones we're here to serve. -/- Warren 13:49, 11 March 2008 (UTC)[reply]

Minefield 3.0b5pre hits 70 points. --Ilhanli (talk) 21:49, 11 March 2008 (UTC)[reply]

I suggest grouping the rows by layout engine. And one column for stable screenshots, and another for latest... --Fenring (talk) 23:17, 11 March 2008 (UTC)[reply]

Okay, I'll try it. I'll order the table by increasing score for development version. -- Schapel (talk) 23:42, 11 March 2008 (UTC)[reply]
I think that table you've started is useful. However I will duplicate the image I had as an illustration to how the test works and put it back up. I'm sure one of the webkit fanboys soon will replace the image in the table with an image with the current score (90)...--itpastorn (talk) 07:51, 12 March 2008 (UTC)[reply]

Minefield some times hits 69, and sometimes 70 points (on my PC). What do you thik, Will we write 70? --144.122.250.130 (talk) 01:09, 14 March 2008 (UTC) Sorry, not looged in --Ilhanli (talk) 14:48, 14 March 2008 (UTC)[reply]

It's the same with FFox 2.0.12 (50/51/52), Minefield (69/70) and in my experience webkit nightly. I downloaded the latest version yesterday on a Mac and only got 89. I think it would be better to have a paragraph saying that the exact scores are not 100 % reliable. Currently the images are not sourced at all. We do not know if they were made during laboratory like circumstances or not. The screenshots only give, quoting from an earlier discussion, an "in the ballpark idea" about the current status.--itpastorn (talk) 08:13, 14 March 2008 (UTC)[reply]

Acidtests are linking back

[edit]

I hope this is in recognition of our work (probably mostly on the Acid2 article). But it is also a responsibility. How can we improve the article? More editors are welcome!--itpastorn (talk) 09:07, 27 February 2008 (UTC)[reply]

Acid 3 finished?

[edit]

Maybe betas shouldn't be added. OK. According to the article, the Gecko development team have both a bug and spreadsheet to track progress with the test. A link from the BugZilla page shows the test has been completed.[5] And just my two cents:

Imnotyouok (talk) 17:22, 27 February 2008 (UTC)[reply]

That link points to a blog which generally is a good resource, but in this case it was wrong, as stated further down in the discussion.--itpastorn (talk) 20:42, 27 February 2008 (UTC)[reply]

According to this[6] the test is now officially released. 77.125.91.97 (talk) 22:26, 3 March 2008 (UTC)[reply]