Talk:Acid3/Archive 3

From Wikipedia, the free encyclopedia
Jump to: navigation, search

This is an archive page for Talk:Acid3

The stable photos[edit]

I think the stable photos should be permanent, i.e. we don't update them when the new version of the browser goes stable.

The pictures should document browsers as they existed at the moment ACID3 was created. This will show the impact of the test, and how it influenced browser development. In fact I think we should add a column for 'Development build as of the release date of the test'.

The Development column should be kept up to date - for now, until all the browsers pass, and then should be removed.

Comments? Ariel. (talk) 03:26, 24 March 2008 (UTC)

Makes sense, but add a "current stable" if an updated stable does not completely pass, but is different than the initial version. There's no reason to put up photos of what it looks like when it passes, however. ffm 17:10, 25 March 2008 (UTC)
It doesn't make any sense to me. There was no "moment of creation" for Acid3. Developers of the browsers were very busy fixing bugs in their browsers for months before Acid3 was officially released. Not to mention that the whole point of Acid3 is to get browser developers to release a stable version that passes the test. To me it serves no purpose and misleads readers of the article as to what the purpose of Acid3 is. How well browsers did before the test was released is of no consequence. I think the table should remain as is, updating the stable and development screenshots as new versions are released, until a stable release of each browser passes Acid3. When that happens, its row should be removed. -- Schapel (talk) 18:34, 25 March 2008 (UTC)
"misleads readers of the article as to what the purpose of Acid3 is" - actually quite the opposite! By showing the sudden improvement in browsers we show how the acid3 test strongly influenced browser development. Look at webkit for example - they would never have made all those fixes if not for acid3. By showing what webkit looked like before acid3 was released we show what acid3 did to browser development. Ariel. (talk) 20:40, 26 March 2008 (UTC)
And exactly how does the current article not show the sudden improvement in browsers and show how the acid3 test strongly influenced browser development? It seems clear to me. What doesn't make sense is showing how browsers did before the test was developed. It serves no purpose. Perhaps after all browsers pass the test, we could show the most popularly used browsers that do not pass and how well they do. That would be useful, to demonstrate that even though all browsers pass, web developers cannot simply use all the features tested, as older browsers that do not pass are still in common use. -- Schapel (talk) 00:01, 27 March 2008 (UTC)
  • I like the idea but I don't think we should have images for each stable revision of a browser, just text to show the progress that was made from the time the Acid3 was released to when it passed.

"History of Acid3 Progress"
• Firefox - 52/100 - 2007/12/01
• Firefox 74/100 - 2008/04/01
• Firefox - 85/100 - 2008/09/01
pattersonc (talk) 03:17, 26 March 2008 (UTC)

  • I'm using Firefox and it passes 53/100. Whats gives? —Preceding unsigned comment added by (talk) 08:22, 3 April 2008 (UTC)
  • I'd just like to point out that the 'Stable Build' for Chrome is not up to date. Chrome's stable release as I am writing this is; and the "Linktest Failed" no longer occurs. --Tindap (talk) 00:44, 16 September 2009 (UTC)
  • I agree with Tindap.. the stable release is and it gets a 100/100, but when I put the image it has been reverted, saying it's a preview release (I don't think so).. and Firefox 3.5 gets 90/100 by the way... Koraiem (talk) 09:26, 16 September 2009 (UTC)

Formal tests[edit]

I propose we make a list article about every major test or test suite for browsers. The Acid tests serve a purpose, but passing formal test suites from W3C is more important. I think we should try to raise awareness and improve WP's encyclopedic value by making a page like I have started sketching out at User:itpastorn/browsertests.--itpastorn (talk) 11:31, 24 March 2008 (UTC)

Rather than create a new article, why not add them to the relevant article:
-- Schapel (talk) 12:32, 24 March 2008 (UTC)
Do both. I think a new page listing all the tests is a good idea, but I don't think you need to create a page for each test (if you were planning it). Ariel. (talk) 13:47, 24 March 2008 (UTC)

No Critique?[edit]

IMHO the article should include negative points about the test as well, like -- (talk) 06:43, 31 July 2008 (UTC)

From what I noticed on the web at the time of that post, it was much regarded as 'whining' from Shaver attempting to push the blame for why Firefox 3 didn't pass Acid3 to the test and it's writers, instead of attributing it to a timing fiasco. And from what I've seen, that has been the only notable article with negative feelings about Acid3. -Pyro3d (talk) 18:17, 6 August 2008 (UTC)

Firefox® 3.1.x Testing Builds[edit]

Hallo, just to report that there's some specific test version of firefox 3.1b1 that strikes 96/100 on acid3 (just tested). Check it and here the screenshot --Gennargentu (talk) 18:34, 18 September 2008 (UTC)

It's now at 97/100 in a build on "Talk like a pirate day" no less... Anyone want to replace the 86/100 screenshot in the "development build" column in the article, as it fits that description, even though it's a special unofficial build, not unlike the WinGogi build Opera screenshot. Mardeg (talk) 21:12, 19 September 2008 (UTC)
I've replaced it with a nightly build. I read that as 87/100 instead of 97/100... I don't think it can be compared to WinGogi as that is at least a build from Opera itself instead of a "third party". --Execvator (talk) 14:11, 21 September 2008 (UTC)
Okay, as requested I replaced it with a screenshot of a build from itself that scores 97/100 Mardeg (talk) 09:44, 3 October 2008 (UTC)
Technically, despite being hosted at, it is not actually an official trunk build. However, since it is built by a Mozilla developer, I think it should be OK. Thoughts? nneonneo talk 03:28, 4 October 2008 (UTC)
Yeah, the official nightly builds (20081007) still score 89/100. We'll see if these changes get integrated. One thing to watch is the Firefox Acid3 spreadsheet linked on the article page. slythfox talk 23:53, 7 October 2008 (PST)
The official nightly trunk build (Minefield 3.1b2pre) still has a score of 90/100 as at 20081015. Though the build referred to is built by Mozilla but it is not the official build. Firefox 3.1 will be released from the official trunk build. I suggest that we should use the screenshot of a official nightly build. Scanorama (talk) 12:00, 15 October 2008 (UTC)
That makes sense. There is a similar problem with the Presto development screenshot; it also doesn't represent how the next release of Opera will perform with regards to Acid3. We should replace Gecko and Presto screenshots with screenshots of the latest official Firefox and Opera builds, which will help to show how the next official releases of those browsers will perform. -- Schapel (talk) 13:14, 15 October 2008 (UTC)
I have replaced the screenshot with the one from the official Firefox nightly trunk build. The latest score is 93/100. Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.1b2pre) Gecko/20081016 Minefield/3.1b2pre ID:20081016033525 Scanorama (talk) 12:24, 16 October 2008 (UTC)

WebKit First to Pass Acid3[edit]

It looks like the WebKit is the first browser engine to pass the acid3 test. Kharri1073 (talk) 20:09, 26 September 2008 (UTC)

Yep, this ought to be mentioned somewhere... Although that version of WebKit is still not released in any stable browser, and Gecko will pass soon too, and Firefox 3.1 might be released before Safari 4, who knows... (Rklz2 (talk) 22:19, 26 September 2008 (UTC))
So, this section of the article is incorrect? --Execvator (talk) 04:39, 28 September 2008 (UTC)
As of build r34278, WebKit produces a smooth animation, and thus passes the Acid3 test[1].

The test might still change[edit]

This might mean that Webkit goes down to 99 again - but it will probably not take many hours for them to patch it again. Heck! They are probably already working on a patch in case the CSS WG decides to clarify the spec and Ian H accordingly changes the test. --itpastorn (talk) 10:23, 28 September 2008 (UTC)

Update. The CSS WG has decided to ask Ian Hickson to change the test.

--itpastorn (talk) 19:51, 9 October 2008 (UTC)

Firefox internal builds at 97[edit]

FYI. Firefox have patches lined up to get to 97.

Probably not stuff for the article yet, though.

--itpastorn (talk) 01:44, 2 October 2008 (UTC)

Chart highlighting[edit]

I'm not sure I agree with the idea of highlighting the table of browsers with red for "fail" and green for "pass," for the following reasons:

  • It is redundant for the first column, since the sentence immediately before the table already states that "every web browser failed the test at the time of its release."
  • No browser is currently shipping a stable release that passes the test, which makes almost all of the table red. It's not clear at all what's being compared until one scrolls to the bottom of the table to see the single green cell, that of Safari 4 DP.
  • No mobile browser at all passes the test, which makes the lower chart completely red.

Simply put, this is unnecessary. Once browsers start shipping stable releases that pass acid3, I propose two sections - one for layout engines that pass and one for those that fail. Cluttering the comparison table, as it now stands, just makes it harder for the reader to make a quick review of where different browsers stand.

I would say something similar about the leftmost column for "score at time of acid3 release" (it just seems irrelevant), but since it's been on the page for quite some time now, the general consensus must be that it is needed.

I won't change the coloring or layout, but I suggest that this is considered.

Thanks, --Luinfana (talk) 21:42, 5 October 2008 (UTC)

I agree with this; additionally, the colours themselves look ugly behind images. I've removed the colours. Warren -talk- 22:11, 5 October 2008 (UTC)

Mobile browsers[edit]

I noticed Warren's edit removing Pocket IE, and I agree -- a screenshot of WM 2003 is hardly new. However, we do need screenshots for Acid3 on PIE, since it is a major mobile browser. I believe that WM6 is the version we want for "Screen­shot of Stable Build at Time of Acid3 Release", and WM6.1 should be the screenshot for "Latest Stable Build Screen­shot". Of course, if someone has access to a development version of WM7, it would be suitable for "De­vel­op­ment Build Screen­shot".

Also, I don't think the Fennec screenshot is indicative of the actual performance of Fennec. It only shows Gecko running on a desktop environment, which is hardly realistic considering that Fennec will likely be pared down and doesn't even seem close to "alpha" status yet. I think that it should be removed, or replaced with Minimo screenshots. nneonneo talk 03:26, 6 October 2008 (UTC)

Yup, if I'm not mistaken, Fennec is actually just a shell running on top of XULRunner, so it uses the same Gecko engine in the XULRunner. Since Gecko 1.9 has yet to be successfully ported to a mobile phone platform (Maemo is not a mobile phone platform), so there's no way to tell how well it will fare in Acid3 when ported to, say, Windows Mobile (as shown with WebKit, on desktop environment it scores 100/100, but after ported to mobile platform it scores lower).
I have removed Fennec from the chart, maybe someone can add Minimo, or wait for a Fennec build that runs on Windows Mobile Ufopedia (talk) 06:33, 10 October 2008 (UTC)
Fennec screenshot on windows mobile added, see Mardeg (talk) 18:14, 11 October 2008 (UTC)

Removing Iris Browser as a "nobody" shows you have no idea what is going on in this space. You can go and download it at, and there are plenty of users. Don't start over the same debate that happened on the WebKit wiki page. It wasn't pretty. Furthermore, Iris Browser has many many many times the number of users that Fennec has. If you are concerned about market share, get some actual numbers, and remove the ones that don't have users. I will add back Iris Browser soon. —Preceding unsigned comment added by (talk) 01:03, 18 February 2009 (UTC)

Should BlackBerry Browser also be added? ("Screen­shot of Latest Release at Time of Acid3 Release" could have a screenshot from a BlackBerry 8800, perhaps.) Meanwhile there's the Hiptop/Sidekick, and the WebKit-based Palm webOS browser... Hexene (talk) 19:16, 1 April 2009 (UTC)

netscape browser[edit]

should netscape 8 or 9 be on the list? i got 35/100 for acid3 and a smiley face that loads better than the netscape 7.2 picture for acid2 (i'm using netscape 8 to test, and you can even scroll somehow...), and (obviously) passed acid1. i'll test netscape navigator 9 tomorrow. Devrit 01:04, 10 October 2008 (UTC)

Netscape had already been discontinued when Acid3 was released. Regardless, Netscape 8 and 9 are based on Firefox, so I don't think they would score any differently than Firefox. - Josh (talk | contribs) 01:22, 10 October 2008 (UTC)

Opera 9.70 Beta?[edit]

I just noticed that the development build for Opera is currently listed as "9.70 b1pre," with the same score and graphic as 9.6, the current stable release.

I can't find a public download or any information about a beta/prerelease of 9.7.

If indeed there is a 9.7 beta, a link should at least be added to the notes section of the chart.

Also, this should not replace the 99/100 screenshot of the Gogi build, which more accurately represents Opera's progress with unstable code.


--Luinfana (talk) 21:56, 16 October 2008 (UTC)

If this is changed back to the 99/100 build from "not the next official version", then Firefox screenshot should also change back to the 97/100 build likewise. Mardeg (talk) 22:25, 16 October 2008 (UTC)
but there is a differeces! the wingogi build is from opera software asa ("an official build"!) --> the firefox build is not from mozilla, it is build by a "fanboy" --> not an official build... so guess... mabdul 0=* 08:37, 17 October 2008 (UTC)
I think you're confusing a separate 3rd party build with the one located at which is built by an official firefox developer. Furthermore the builds are for all 3 supported platforms as opposed to a single platform the wingogi build is for, so a screenshot of it could rightly be seen as more justified to be there than the Opera screenshot, but I'm happy to see both restored. Mardeg (talk) 10:37, 17 October 2008 (UTC)
Yeah you're right, we should bring both dev-builds back! But The Gogi build is also avaible for Linux! ( mabdul 0=* 10:42, 17 October 2008 (UTC)
Why should the development builds be brought back? The Opera dev build screenshot seemed to only cause confusion. Opera users couldn't understand why Opera 9.5 didn't score 100% on Acid3, for example. I didn't want to say anything, because for some reason people think I'm picking on Opera. But after someone pointed out that the Firefox screenshot was confusing, I thought I could speak out without being accused of favoritism. -- Schapel (talk) 11:32, 17 October 2008 (UTC)
If people are confused, then we should adding inks to the article/page where to download the build... mabdul 0=* 17:04, 17 October 2008 (UTC)
I don't see how that would solve the problem. Opera users see that a development build of Presto passes, but it does not represent how the next released version of Opera will perform on Acid3. It's exactly the same problem that was discussed with Firefox. To me that's the whole purpose of the column -- to demonstrate how well future released browsers will perform. It doesn't matter how well a browser performs on Acid3 if it's never released officially and never sees widespread usage. -- Schapel (talk) 17:28, 17 October 2008 (UTC)
I can only agree and have fixed the column headers to clarify their intended purpose (Rklz2 (talk) 21:21, 17 October 2008 (UTC))
It's not about whether you think it will solve the problem of confusing readers, it's about reputable and notable sources. And Opera's Gogi build has many reputable and notable sources. Ufopedia (talk) 11:49, 18 October 2008 (UTC)
I think Wikipedia is all about writing readable and understandable articles. Don't get lost in the minutia of the rules. Use common sense. Just because you're allowed to say something because you can find a reliable source that says it, doesn't mean you should add that information to an article. The information you add to articles should first and foremost make it a better article. Including information about custom Firefox and Opera builds isn't particularly useful for the purpose of this article. What's useful is seeing how the next official release of Firefox and Opera will perform. -- Schapel (talk) 13:10, 18 October 2008 (UTC)
Again, that's not for you to decide, you can spend all your time rationalize what "makes a better article" and what not, but what you think as a "better article" doesn't matter here, and what you think "Wikipedia is about" doesn't matter here. We write the article according to the rules, not according what you think is useful. The rules of Wikipedia say that it should include relevant information from notable and reliable sources, and that's not what I or you say or think, and that's how we should write the articles. Ufopedia (talk) 02:44, 19 October 2008 (UTC)
Quote "The rules of Wikipedia say that it should include relevant information from notable and reliable sources". So the debate is about whether an internal development build passing the test which doesn't even feature a user interface, is relevant to the article. I don't think so. WinGogi is not even functionable as a browser. (Rklz2 (talk) 06:06, 19 October 2008 (UTC))
The article is about Acid3. Is WinGogi relevant to the article? Of course it is. Does it represent the progression in the Presto layout engine? Sure it does. Whether it is a full-fledged fully-featured browser is irrelevant here. And the table is not about browsers, but about layout engines. So WinGogi IS relevant information from notable and reliable sources. Oh, and WinGogi surely DOES feature a user interface, and IS actually a functional browser, just not fully featured. Have you even tried it? I hope you actually go download and try it yourself before saying clearly incorrect things like "doesn't even feature a user interface". And it's quite functionable as a browser, I'm using it to edit wikipedia here Ufopedia (talk) 05:38, 20 October 2008 (UTC)
I agree with Rklz2's changes to the column headers; preview releases are indeed more relevant to the article than development builds. That said, as the headers stood before said edit, I believe the Gogi build was the proper representative for Opera's progress. Although it didn't imply that Opera's next release would score 100/100 (or 99/100), certainly the fact that a browser team can reach a certain score says something about their progress and their dedication to passing. As I said, though - I do think the headers are more clear as they are now, and the preview release cell for Opera can stay at N/A for the time being. --Luinfana (talk) 14:15, 18 October 2008 (UTC)

How about adding a better note or / and a new row to add technical previews? Otherwise we get for a short time editwars and we lost some data... mabdul 0=* 17:36, 18 October 2008 (UTC)

This would just serve to clutter the table more. Most engines would have the same scores and renderings in a "Development Build" column and a "Technical Previews" column, anyway, and it's probably not very clear to most readers what the difference between the two terms is. I believe it's fine as it is now - "Preview Releases" is a suitable title. Let's keep clarity, simplicity, and readability in mind as we make edits and suggestions concerning this article. --Luinfana (talk) 19:00, 19 October 2008 (UTC)
And WinGogi is actually an official preview build from Opera Labs Ufopedia (talk) 05:44, 20 October 2008 (UTC)

What about being pragmatic: put the screenshot of the latest dev build (alpha/beta for IE, weekly for Opera, nightly for Firefox) and for those having specific test/demo build, put a comment under the screenshot saying "There is a test build, called WinGogi, showing 099/100" or "There is a test build from Mozilla, showing 093/100". With a link to it. This will allow to see very quickly what is important: how well the next release will perform, and indicate that work is on-going and give the specific information. (talk) 08:02, 20 October 2008 (UTC)

Test 46 and Firefox Nightlies 'regressing' 1 point[edit]

Test 46 now reports as failing on Firefox and the score has gone down to 92/100, but this isn't being reported on the front page. But the WG has actually changed the specification and have asked Hixie to change the test which has not been done yet. More details here: and here: --Zurtex (talk) 16:11, 5 November 2008 (UTC)

"Origin of the name" section unnecessary[edit]

I do not think that the inclusion of the name origin section is strictly necessary, and even if it is, it contains glaring errors. Since when was "Acid3" short for "Acid test #3"? Google reveals no results, other than the article itself. CHL (aka yse) (talk) 13:48, 7 November 2008 (UTC)

I think it's a useful addition to the article. I for one thought until reading it that the "Acid" name alluded to the substance used while creating the Acid1 test... And how is what you cite a "glaring error"? Acid3 obviously does stand for Acid test number three. (Rklz2 (talk) 00:45, 9 November 2008 (UTC))

It does, but it certainly isn't a shortening of that. CHL (aka yse) (talk) 06:54, 9 November 2008 (UTC)

Konqueror stable screenshot and information[edit]

In "Screen­shot of Current Release" or KHTML(Konqueror), the image shows 76/100, but the label shows "080/100". Shouldn't one of them be updated/fixed to be consistent?Luiscubal (talk) 20:10, 12 November 2008 (UTC)

Took care of it. -Luinfana (talk) 20:16, 12 November 2008 (UTC)

USE="+webkit" (on Gentoo) Konqueror 4.2.3 passed 100 % acid3 test. —Preceding unsigned comment added by (talk) 18:47, 19 June 2009 (UTC)

why is the test not usable without javascript?[edit]

I think that this is not very friendly to people who have javascript disabled. —Preceding unsigned comment added by (talk) 19:30, 14 November 2008 (UTC)

It's mainly a test of how well a browser handles JavaScript and the DOM. How can you test that if you have JavaScript disabled? -- Schapel (talk) 20:41, 14 November 2008 (UTC)
I don't think it's very friendly of OP to disable javascript </useless_comments> (talk) 16:05, 13 July 2009 (UTC)

Wrong Colours, Fails Acid3?[edit]

Reference: Webkit:

Clearly, none of the colours in the image are the same. I am assuming this is because the person who took the screenshot has certain effects turned on in their window manager. Nevertheless, a new screenshot should be created. (talk) 03:08, 22 November 2008 (UTC)

I can only see that the purple is different. the rest seems to be same... --mabdul 0=* 09:57, 22 November 2008 (UTC)
I just checked the colours in Fireworks. The reference has #C0C0C0 for gray (Acid3 text shadow), #F00 for red, #FFA500 for orange, #FF0 for yellow, #0F0 for green, #00F for blue and #800080 for purple. The webkit image has #CACACA for gray, same as reference for red, yellow, green and blue, #FFB300 for orange and #910091 for purple. The Firefox image does the same thing as the reference. I will get a new screenshot. nneonneo talk 17:34, 22 November 2008 (UTC)
Done. nneonneo talk 17:47, 22 November 2008 (UTC)

Opera 10.0 alpha[edit]

Opera 10.0 alpha just released with its new Presto 2.2 rendering engine, and it is the FIRST public browser to nail the acid3 test. YEAH!
Maybe someone should put this on the wikipedia article
20:24, 4 December 2008 (UTC)

Safari 4 has fully passed for months. Opera 10 still doesn't pass the performance aspect of the test yet. That should be clarified in the article. -- Schapel (talk) 19:43, 5 December 2008 (UTC)
Safari 4 Developer Preview, which scores 100/100 in Acid3, was released much earlier than Opera 10.0 alpha 1. And nope neither Safari 4 DP nor Opera 10.0 alpha 1 fully pass Acid3, they both fail the smoothness criteria. Ufopedia (talk) 10:50, 7 December 2008 (UTC)
Safari 4 has fully passed Acid3 since September 25, 2008: -- Schapel (talk) 13:20, 7 December 2008 (UTC)
That's not Safari 4, but WebKit nightly. You can use WebKit nightly with Safari 3 to pass Acid3 too, so does that mean Safari 3 passes Acid3? Clearly no. Thus Safari 4 DP does not pass Acid3 neither.Ufopedia (talk) 12:56, 8 December 2008 (UTC)
Agree with Ufopedia; also it should be noted that neither Opera 10 Alpha nor Safari 4 DP are "public browsers;" they are unstable development releases and should (as they are now) be classed as such in the article. Luinfana (talk) 20:15, 8 December 2008 (UTC)
They are classified as development browsers. When stable versions are released, they will be listed as passing the Acid3 test, in a section that does not exist yet (or is currently commented out). I've tried before to make edits to clarify the situation, and then the table becomes a mess again. Should I try to clarify yet again? -- Schapel (talk) 21:59, 8 December 2008 (UTC)

Opera 10 is again listed as passing, although I think it still does not pass the performance aspect of the test. Does anyone have a source to cite so the claim can be verified? -- Schapel (talk) 20:08, 3 August 2009 (UTC)

Article error[edit]

There's an error in 'Desktop browsers' chart: at the last column of Gecko row, it says 'Firefox 3.1 Beta 3 and...' but Firefox 3.1 are in Beta 1 status as of Today (6 December, 2008) and Beta 2 is expected to be released within a few days from now. I cannot correct the article myself, as i don't know, which browser version was tested... Hardzsi (talk) 13:34, 6 December 2008 (UTC)

I guess it's supposed to mean Firefox 3.1b3pre, but then I think svn snapshots should not be put there as Preview Release anyway? Ufopedia (talk) 11:06, 7 December 2008 (UTC)
The latest nightly version of Firefox is 3.2a1pre for the trunk and 3.1b3pre for the 3.1 branch. Either way, neither require the use of svn and are available by HTTP. --Execvator (talk) 18:23, 7 December 2008 (UTC)
I never said it requires svn, but that those are svn snapshots. And since last time the column was renamed from "development builds" to "preview release", I don't know if those can classify as "preview release" here.Ufopedia (talk) 14:18, 8 December 2008 (UTC)
How is that not a preview (it shows what's going to come)? What exactly do you consider to be the requirement for a release, an announcement? --Execvator (talk) 18:10, 8 December 2008 (UTC)
See this and this, apparently nightly builds are not classified preview releases but development builds here, at least that's what the last column header change implied, since Opera weekly snapshots and WebKit nightly builds were no longer included in the table after the column header change. Ufopedia (talk) 00:28, 9 December 2008 (UTC)
BTW, I guess preview release means this in wikipedia. Ufopedia (talk) 01:52, 9 December 2008 (UTC)
Yes in an article on the browser itself the latest stable beta should be counted as the preview release. Though in the context of ACID3, the latest build that shows what the next stable release of the browser is going to look like would be a better choice, as I see it. Which in the case of Firefox would be the 3.1 branch. --Execvator (talk) 19:19, 9 December 2008 (UTC)
So did you read this discussion here in this page that I have linked? Well, if you think that is the case, then BE BOLD. Like I said, last time the column header changed from development builds to Preview Releases, the Opera weekly snapshots, Opera Gogi build, and Webkit nightly builds got removed from the table, and there was kind of an "edit war" there. Also strictly speaking "the next stable release" of the Firefox browser will be Firefox 3.0.5. Anyway if that's how you see it, feel free to edit it yourself, this is wikipedia after all Ufopedia (talk) 02:42, 10 December 2008 (UTC)
I made that change to "preview release" from "development build". I think the article should show what users can expect in the near future from the browser.
* Firefox -.-.x releases are technically security patch releases, they won't add any new standard support unfortunately so that makes no sense to list them as next/preview release.
* WinGogi was not a fully usable browser and I thought the rendering engine used in it might never be released because Opera was having internal problems, focus on mobile market and it makes not much sense to release a desktop browser. But, Opera 10 Alpha is now released and it's definitely a preview of their next browser and very usable, so the Opera issue could be solved now.
* WebKit nightly is also not usable as a browser, although you can make a stable version of Safari use it, it's not suitable for actual browsing as it crashes very often. Plus, WebKit mainline might include stuff that other browsers that use webkit might never include, for example Chrome explicitly removes some stuff from WebKit, some code is not supported on all platforms, etc.
What it all boils down to is that maybe basing the table on rendering engines instead of browsers was a mistake. If it was browser based, determining what to consider as preview of the next release would be straighforward. Plus, even though many browsers share the same engine, some use older versions of it for stability or coder lazyness or portability issues, so even though WebKit has long been passing Acid3, Chrome users might not see a pass any time soon.
Anyway, some change is definitely needed... (Rklz2 (talk) 17:39, 10 December 2008 (UTC))

Opera 10 Alpha Rendering[edit]

The rendering for Opera 10 Alpha shouldn't show the browser chrome (surrounding menus, user interface), since none of the other images on the chart do. The simplest thing to do would just be to replace it with the reference image, but I won't do it myself. Luinfana (talk) 22:14, 13 December 2008 (UTC)

Done. I just cropped it. nneonneo talk 00:07, 14 December 2008 (UTC)

browsers or engines[edit]

We got a problem ifwe let add everybody his own and favourit browser, which has the same rendering engne.(esp. gecko and webkit!). we should take 3 or 4 product max. of every rendering engine and then stop and lock it! mabdul 0=* 06:50, 15 December 2008 (UTC)

I agree. I'm going to remove from the table every browser that has far less than 1% share, that is, all but Internet Explorer, Firefox, Safari, Chrome, Opera, and Netscape. If someone wants to know how their favorite browser fares, they can simply try the test themselves, or look up the engine the browser uses in the table. -- Schapel (talk) 14:24, 15 December 2008 (UTC)
Why limit it? Why keep away relevant information? The browser that's most interesting these days (in my point of view) is Midori, and it's not even listed. Why not add all browsers? Note that both Chrome and Midori uses Webkit, but Midori scores better. --Alexander256 (talk) 23:18, 1 June 2009 (UTC)
That's a fair point. Littering this article with every browser in existence will be completely unsustainable, but as it is every engine has a relatively complete cross-platform representation except Trident (which isn't cross-platform) and WebKit (for which Midori is the missing link). ɹəəpıɔnı 15:09, 2 June 2009 (UTC)
We will add fthe final versions of browsers that support acid3 in the table, but the development screenshots don't make sence mabdul 23:05, 2 June 2009 (UTC)

Alright, there is much talk about browsers and engines. There is a notation on the main page about the "first release" browser that passed the Acid3 test. I guess "first release" needs to be defined? Anyway, I have a saved image on Photobucket, showing Midori Browser passing the test on Feb 24 2009. That, after I misplaced an earlier image, taken in mid December, I believe. I'm a bit fearful of editing the main page - I don't want to screw it up. The link to the image: I am NOT a developer or anything, I just test a lot of stuff, and I think Midori should be mentioned. Info on Midori at If there is any question about using my saved desktop, I hereby authorize it's use and reproduction according to Wikipedia's TOS and it's cited copyright and copyleft licenses. So, someone snatch it and use it, alright? Runaway1956 (talk) 04:06, 5 June 2009 (UTC)

Ummm, he's right. I just installed Midori 0.1.2 and it passes with flying colors. Interesting. —Remember the dot (talk) 05:23, 5 June 2009 (UTC)

Google V8 vs Apple WebKit JavaScriptCore SquirrelFish Extreme etc[edit]

Okay, so people seem to think Chrome scores differently than Safari because they use different "JavaScript engines". A JS engine has many parts which do:

  • Parsing of the JavaScript code
  • Library which contains implementations of standard JavaScript functions/objects, and a bridge to the layout engine DOM etc
  • Runtime which either interprets the parsed code or turns it into direct machine code and runs it, inserting calls to the Library

Now, KDE KJS/Apple WebKit JavaScriptCore contain all three components, and interpret the code. Google developed V8 which is only a runtime, turning the internal representation into machine code and running it (This is called JIT). This code still calls the functions from the JavaScriptCore library, and should exhibit exactly the same behaviour except for running faster. Then, Apple developed SqirrelFish and later SquirrelFish Extreme which are again only runtimes but don't turn the whole code to machine code but only the parts that get executed a lot. This speeds up the most performance sensitive code while consuming less memory (machine code is longer than JS and V8 thus consumes more memory and is not suitable for mobile phones where WebKit is used). So basically no matter if it's V8 or SFX, the Acid3 score is not affected or any JavaScript behaviour for that matter. The difference in score is caused by different versions of the JSC library and generally WebKit used, even though the User-Agent says the same (525.19?) for both, look at the SVN revision of WebKit which they use. Other differences come from the WebKit port, that is the platform dependant parts of WebKit like font handling, graphics drawing, and so on, which each browser needs to implement on their own and Chrome omits some of the stuff that Safari implements. Look here at the Chrome sourcecode where the JavaScriptCore library from WebKit is included. (Rklz2 (talk) 20:41, 16 December 2008 (UTC))

Acid3 also tests performance so while they might not technically "score" differently in the displayed score/100, they will score differently in the overall test as a direct result of V8/SquirrelFish. ɹəəpıɔnı 03:02, 2 April 2009 (UTC)

Firefox/Fennec development score[edit]

Although Firefox is listed as 93/100, there are currently builds with 97/100, shouldn't this be updated? Also, the Fennec score seems to have dropped. Apparently, it's a Windows Mobile screenshot. In these cases, which version should be displayed? —Preceding unsigned comment added by Luiscubal (talkcontribs) 11:55, 8 January 2009 (UTC)

The builds of Firefox that score 97/100 include patches that will not land for the next version being 3.1 - this was already discussed on this talk page. I've reverted the edit that was a pre alpha 1 screenshot of Fennec which is now up to alpha 2 and scores the same as minefield. Mardeg (talk) 04:37, 9 January 2009 (UTC)

ie 8 rc1[edit]

Why is it that the image for IE 8's RC1 shows 21, but the caption shows 20? Althepal (talk) 06:55, 27 January 2009 (UTC)

It seems like as with previous versions IE8 RC1 stops in one of the tests early without running trough all 100 of them. It stops at a score of 20/100 allthoug looking at the test it seems to have stopped at 20/30 or something. hAl (talk) 07:16, 27 January 2009 (UTC)
In Windows 7, it show 21 score. Matthew_hk tc 22:14, 29 January 2009 (UTC)
Can you upload a screenshot, similar to the ones already on the article? Also provide the user-agent (which you can find from a site like when you do. nneonneo talk 22:58, 29 January 2009 (UTC)
If you hold Shift and click the "A" of Acid3, you will get a full report. It should show that RC1 fails 80 tests. nneonneo talk 22:58, 29 January 2009 (UTC)
I've uploaded a new image from Windows XP running IE 8 RC1. nneonneo talk 22:58, 29 January 2009 (UTC)
matthew_hk that is the IE 8 beta 2, they didn't release a RC1 version yet for windows 7 Redekopmark (talk) 00:57, 30 January 2009 (UTC)

Firefox 3.1 beta 3 pre[edit]

I'm using the absolute latest build of Firefox 3.1 and I noticed that acid 3 "Test 27 failed: e2 - parent element doesn't exist after waiting" bringing its score down to 92. There's supposed to be a beta 3 released today so we'll see... Althepal (talk) 18:30, 2 February 2009 (UTC)

Scratch that, when I refreshed it passed test 27. Althepal (talk) 18:31, 2 February 2009 (UTC)


I'm unsure about the following assertion from the introduction:

"The percentage displayed is based on the number of sub-tests passed. It is not representing an actual percentage of conformance as the test does not keep track of how many of the tests were actually started (100 is assumed)."

Surely if a browser is unable even to start a sub-test, this is a sure sign that it does not conform to the standards that the text requires? I think that the percentage is a percentage of conformance.

Mattus27 (talk) 19:43, 19 March 2009 (UTC)

NO! It isn't there is also a need of pixel - conform - rendering (especial the a in the right corner!) and the speed test! mabdul 0=* 21:51, 19 March 2009 (UTC)

Internet Explorer 8[edit]

Ok so now that IE8 is officially released, it still only scored 20/100 and in "compatibility mode" scored an amazing "13 100" (the slash doesn't even render)

When is Microsoft going to realize IE can't have it's own standards. Rnawky (talk) 03:15, 20 March 2009 (UTC)

wrong place to discuss about... this is wikipedia not a forum ;) mabdul 0=* 08:32, 20 March 2009 (UTC)

IE8 beta 2 score is 21/100 dont remove it! —Preceding unsigned comment added by (talk) 10:38, 15 May 2009 (UTC)

IE8 was released. IE8 beta 2 is an old version. Please do not add IE8 beta 2. -- Schapel (talk) 14:03, 15 May 2009 (UTC)

My machine is running IE 8 (8.0.6001.18702) and I'm seeing a 20/100. Took a while for it to load and it spawned some javascript errors (and a warning wanting me to allow MSXML active X control to run). 22 May 2009 —Preceding unsigned comment added by (talk) 15:38, 22 May 2009 (UTC)

yes, let's all compare our results RIGHT NOW *kills self over state of wiki users* (talk) 17:03, 13 July 2009 (UTC)

Screenshot of preview release[edit]

I reverted an edit updating the version number written under Opera's Screenshot of preview release from build 1139 to build 1491, because of the comment in article source stating:


My revert was re-reverted and now as I look at the version numbers, most are incorrect. Is the above rule to be followed or should the comment be removed? As of now, the screenshots are of the following builds:

Google Chrome                 (article states
Safari 4.0 developer preview (528.1.1)  (article states Safari 4 beta)
Opera 10.0 Build 1139 Alpha             (article states build 1491)
Firefox unkown build                    (article states 3.6a1pre)
(Konqueror is correct)

ɹəəpıɔnı 02:26, 9 May 2009 (UTC)

I think we should just remove all the version numbers. They're incorrect, clutter up the table, need to be updated with each release, and don't really matter as long as the screenshots are taken with the latest version of the browser. -- Schapel (talk) 12:37, 9 May 2009 (UTC)
Good idea. Even if the above comment were "enforced", many would miss it and go ahead and update constantly anyway. I'll get rid of them. ɹəəpıɔnı 20:40, 9 May 2009 (UTC)

Safari 4[edit]

Safari 4 should be discussed here even though it is technically a "beta" (the same way Gmail is beta). Primarily because it is the first (and as far as I know still only) the consumer browser to fully pass Acid 3. I realize that people get zealous about their browsers, but taking this benchmark certainly earns Safari a mention here. Zerocool3001 (talk) 01:59, 24 May 2009 (UTC)

GMail is famous for being the first mainstream "perpetual beta" web application. However, in more recent times "perpetual beta" webapps have declined in popularity and proliferation for various reasons, and most importantly the paradigm has never been extended to desktop apps (excluding RIAs). Safari 4 is NOT in perpetual beta, it is in exactly the same type of beta as IE8 beta was not so long ago and as Opera 10 soon will be.
WebKit was the engine to pass ACID3 on September 25th of last year, so Apple hold the accolade either way. There's no browser or vendor zealotry here, simply a desire for accuracy in an encyclopedia article. No stable browser has ever passed ACID3 to date. The term beta is by definition NOT STABLE (GMail - your example - is not considered stable btw). Safari is not "technically" a beta, it is not "kind of" a beta. It is a beta. ɹəəpıɔnı 16:06, 25 May 2009 (UTC)
The fact that Safari 4 was the first browser to pass Acid3 is mentioned in the article. Am I missing something? -- Schapel (talk) 17:13, 25 May 2009 (UTC)
I disagree. Apple (as with a few other companies) has shown their willingness to use the "beta" title in public releases only to signify stable software that is ready for release but is being held back for inclusion in other major releases (or for similar reasons). Regardless of our feelings on the term "beta" it would be highly inaccurate to not include any mention of Safari 4 in the discussion of the Acid3 standard, not least of which due to the press attention given to the passage of Acid3. To not include it leaves out a major piece of the story.
Schapel. The discussion here is about the inclusion of a single line in the "Development and Impact" section of this article. You're right, it is now included, but some are in favor of removing the reference. Zerocool3001 (talk) 00:37, 27 May 2009 (UTC)
@Schapel - A WebKit nightly was the first browser to pass ACID3, Safari4 uses this engine within it for page rendering.
@Zerocool3001 - Is this your own deduction/speculation, or have Apple actually stated that Safari 4 BETA is a stable release. If the former, I will remove the reference AGAIN (WP:3RR is the only reason it's still there now), if the latter please provide a reference. ɹəəpıɔnı 15:15, 27 May 2009 (UTC)
WebKit is not a browser. It is a browser engine. It looks like a nightly build of Safari 4 was the first browser to pass Acid3 according to the source cited. Why not leave that information in the article? -- Schapel (talk) 15:56, 27 May 2009 (UTC)
WebKit is an open-source project by Apple, Safari is not open-source, similar to the difference between Chromium (open source Google browser) and Chrome (closed source Google browser using the Chromium engine). The WebKit nightly releases of the browser that you can get here are called "WebKit". It is similar in ways to Firefox calling their development builds "Minefield" and "Shiretoko", but even further removed as they are different projects (one open-source, one closed). ɹəəpıɔnı 19:06, 27 May 2009 (UTC)
So the problem with saying that Safari 4 was the first browser to pass the Acid3 test is... what exactly? -- Schapel (talk) 21:12, 27 May 2009 (UTC)

The problem is: it is not released. it is in a special part of a development release (but not the final release!). this is similar with the race to the moon: usa was first; udssr was only first in space... it is simliar! mabdul 21:38, 27 May 2009 (UTC)

"the problem with saying that Safari 4 was the first browser to pass the Acid3 test is..." that Safari 4 is NOT OUT YET!! How can a non-existent browser pass a test? Safari 4 beta ≠ Safari 4 final. ɹəəpıɔnı 22:19, 27 May 2009 (UTC)
Just because the final version hasn't been released yet doesn't mean the browser don't exist. Your argument makes no sense. Safari 4 beta does exist and does demonstrably pass the Acid3 test. We have a source to cite that says so. i will re-add the information. -- Schapel (talk) 23:18, 27 May 2009 (UTC)
Sorry Schapel, I suppose I was a little unclear in my wording. To summarise:
  • Apple WebKit r36882 was the first browser to pass ACID3. This was an unstable build that came out long before Safari 4 beta (September 25th, 2008)
  • Safari 4 beta is an unstable browser build that was released on 24th Feb 2009, 5 months after the first browser passed the ACID3 test
  • No stable release has passed ACID3 yet.
  • Safari 4 will most likely be the first browser to pass it when it is eventually released as Google Chrome 2.0 (just released) doesn't pass, and Opera 10 is not scheduled for release until after August.
Those are the facts of the matter, do with them as you will. I won't engage in an edit war, but the statement that Safari 4 is the first browser to pass ACID3 is false either way. If you say it's the first unstable browser, that's false as it was passed 5 months before Safari 4 was released, if you say it is the first stable browser that's false as it's not stable. ɹəəpıɔnı 04:49, 28 May 2009 (UTC)
Just to note here, WebKit is not, has not, and almost certainly will never be a browser (See the WebKit Blog). Safari 4, although it's a beta, is a browser. Thus, Safari 4 Beta is the first browser to fully pass the Acid3 test. Hope this clears up some confusion on that matter.-Pyro3d (talk) 05:07, 28 May 2009 (UTC)
Ha! Funny, but you (Pyro3d) don't seem to have actually read the page you linked. "When you run, you’re actually running the shipping version of Safari" - by your own logic then (not mine), Safari 3.1.2 was the first browser to pass the ACID3 test. Feel free to add that to the article if you honestly think it's correct.
The fact is, if you insist comparing the term "browser" and the term "browser engine", then only the latter is relevant in rendering ACID3. The UI framework (which is all that distinguished WebKit from being a browser) has nothing whatseover to do with passing. WebKit r36882 was the first browser engine to pass the test. It ran the Safari 3.1.2 UI Framework.
I personally don't think writing "Safari 3.1.2 was the first browser to pass ACID3" would be particularly accurate, do you? It's more accurate than "Safari 4 was the first browser to pass ACID3", but still not really right here. ɹəəpıɔnı 05:29, 28 May 2009 (UTC)
Nitpicking aside, development versions of the browsers Safari and Midori both currently pass Acid3. Correct? Are there any other browsers that pass Acid3? Which was first? Whatever the case, there was a first browser to pass the test, if there are at least two now. Pointing out which was first is simply stating a plain fact. I don't see why it's sparking such argumentation. -- Schapel (talk) 11:46, 28 May 2009 (UTC)
It is already stated in the Development and Impact section which browser engine passed first. Posting that Safari 4 beta was first is posting two contradictory lines in the same section. I don't understand where the confusion is here. Safari 4 was not the first browser to pass the ACID3 test, if you insist on naming a browser (with coupled UI framework) then I can say that Safari 3.1.2 was the first browser to pass the ACID3 test. I will add that to the article if you like. In fact, I think I will now as it will but this silly debate to rest.
Another note, the "reference" used next to this erroneous line - have you read the article it points to (maximumpc)? It's just uninformed nonsense from some non-notable blogger, quite evidently not written by any authority on the subject.
"I don't see why it's sparking such argument" - it's a matter of encyclopedic accuracy. If anyone with any half-decent knowledge of web browsers were to happen upon this article, what they would perceive is an inaccurate statement on Wikipedia. Simple as. ɹəəpıɔnı 15:16, 28 May 2009 (UTC)

I can understand the confusion over the dates. While it doesn't have anything to do with the Webkit release date it does have to do with where Safari 4 was available for download. It was released on the Apple Developer site (for anyone to download) in September 2008, but was only moved to the main site (again for anyone to download) in February 2009. I would be happy to move up the date at which Safari 4 passed Acid3 to its February date, but I think it would highly inaccurate to leave it out of this article completely. It is also, for the reasons cited above, not an unstable release. As for the authority of the article it is a bit non-mainstream, but it was one of many articles I found and simply the one I selected to cite. Additional sources include TGDaily and Mac OS X Tips. Most importantly over all the "Features" page for Safari on Apple's own site cites Safari as the first browser to pass Acid3. It is also telling that although browsers of Apple's site can manually download Safari 3, all pages related to Safari redirect to Safari 4 indicating its officially supported state. Zerocool3001 (talk) 22:25, 30 May 2009 (UTC)

Are you suggesting Zerocool that we painstakingly document the release of every single browser version in the Developments section and comment on whether or not it fully passed Acid3? Safari passed Acid3 in September 2008 - this is already stated in the Developments section.
Saying "Safari was the first browser to pass Acid3 in September 2008" at one point and then saying "Safari was the first browser to pass Acid3 in February 2009" later in the same section is false and confusing. Safari passed the test in September, I don't understand why more info is required. Would you insist on posting "Safari Beta 2 was the first browser to pass the test in July 2009" if that version were to (hypothetically) come out then.
Incidentally, please stop adding the link to - it does not pass WP:SOURCES. ɹəəpıɔnı 23:22, 30 May 2009 (UTC)

No one is asking for two references two Safari. The first references is to Webkit and an unnumbered "development" version of Safari, which is technically incorrect and misleading. I would be happy to consolidate the two references for the sake of compromise. However I think leaving only the first reference does not accurate reflect the situation. As for the article cited, see my above comment.Zerocool3001 (talk) 03:48, 1 June 2009 (UTC)

The February release of Safari 4 was a development version - no different to the September release. How is the statement about WebKit in any way "incorrect and misleading"? Stating this does not make it so - it is perfectly correct and perfectly clear, it is only the addition of your erroneous line that is making the whole issue of when Acid3 was passes misleading. The fist publicly released browser to pass Acid3 was released in September 2008. Stating otherwise is false, will mislead Wikipedia readers and erode the perceived reliability of Wikipedia content. Please remove the line AGAIN, I know you will just put it back again if I do it, despite WP:3RR.
I see you also put back the "reference" to maximumpc. Please take the time to read WP:SOURCES. ɹəəpıɔnı 05:04, 1 June 2009 (UTC)
What I, and others are trying to tell you that WebKit is not a development release of Safari. Sure, they're primarily develped by the same company, and Safari uses WebKit as its renderer, but WebKit is not limited to Safari. Just because the WebKit team releases their binaries as a wrapper for Safari does not mean that WebKit has magically become equivalent to Safari. There are other projects, such as QtWebKit, that use Webkit, or the multitude of Linux browsers that use it too. Adobe AIR, Palm's webOS, and mobile phone browsers are out there too. WebKit != Safari, and we should keep the difference clear to the readers. -Pyro3d (talk) 20:14, 1 June 2009 (UTC)
Joe Bloggs says to himself "I wonder when Acid3 was first passed..." - checks Wikipedia, reads down, says "Oh, February 2009. Ok." - leaves, ill-informed. ɹəəpıɔnı 20:26, 1 June 2009 (UTC)
Joe Bloggs would either a) Heard of Acid3, know a little about it, and read more of the article to learn more, and see that WebKit was the first engine to pass, and passed in September; b) Have no idea what the hell Acid3 is and read the whole article, or c) Have never heard of it, like most of the web, and not be there in the first place. If we leave it as is he'd still leave ill-informed, thinking that it was Safari, not WebKit that passed first. -Pyro3d (talk) 22:34, 1 June 2009 (UTC)
Your (a) example is patently incorrect. I know a good deal about Acid3, so I'd say I fit into that category. If I were reading this article and trying to find out the first browser to pass I would either (a) think it was Safari 4 in February (false) or (b) be extremely confused by the fact that the article contradicts itself (not how a Wikipedia reader should feel). ɹəəpıɔnı 04:55, 3 June 2009 (UTC)
If you know a good deal about it, then you wouldn't fit in to a), which specifies little knowledge of Acid3 ("know a little about it". I suggest you read my post before contradicting yourself. Note that "a little" != "a good deal"). I left those who know a lot out, because they obviously would not be looking here for information, they'd be contributing it. The difference in terms, Browser vs. rendering engine, browser engine, engine, etc. would not necessarily confuse the reader, because, like most of Wiki, there would be a link to the corresponding pages, thus making the difference clear. Saying Safari passed in September is equally, if not more false, because saying that Safari passed first has some truth to it in a browser frame of reference. Currently it misleads users into thinking that WebKit is a development version Safari. -Pyro3d (talk) 06:22, 3 June 2009 (UTC)
Whether I know "a little" or "a good deal" is pedantics - they are relative/subjective terms. The point is, if I were reading this article looking for that particular information (because I would do that, regardless of my level of knowledge - I consider Wikipedia a good source of information and reference it a lot), that is how I would interpret what I read (incorrect facts, or simple confusion). Maybe it would not be how you interpret it, but it would be how I would, and how I believe most would.
I could download a browser in September 2008 and use the browser to browse the web (as browsers do). That browser passed the Acid3 test. What you prefer calling that browser is a subjective issue - some would call it the WebKit browser, some would call it Safari 3 - neither is 100% correct, neither is 100% incorrect. What is inarguable is that it's a browser.
The article currently states that the first browser to pass Acid3 was released February 2009 - this is false and misleading.
As a sidenote, the Midori browser passed Acid3 in December 2008. There is/was no question as to the name of this release - it is/was called the Midori browser and could be downloaded and used to browse in December 2008. ɹəəpıɔnı 05:36, 7 June 2009 (UTC)
I should mention that Midori compiles using whatever version of WebKit is installed; it doesn't come bundled with either the source or binary of any version of WebKit. In fact, WebKit developers specifically claimed that the GTK+ port, which Midori uses, was not claimed to have passed Acid3; only relatively recently (less than 3 months ago) did WebKitGTK pass Acid3, so Midori, in December 2008, could technically be considered to not have passed the Acid3 test, since WebKitGTK at that time didn't pass Acid3. On a similar note, only a version of Safari that shipped, bundled with a passing version of WebKit, should be considered to have passed. I have absolutely no idea which version that is, though. CHL (aka yse) (talk) 06:18, 7 June 2009 (UTC)

I see that someone got a little overzealous and added Safari to the end of the list. I know that it can get a little confusing so I propose that we consolidate the references. I've written a small paragraph that we can place at the end of the list. It doesn't fit with the timeline, but I don't want to make someone searching through all the references to find which browsers passed and which was first. I've put the paragraph here so I can get your feedback before I modify the article:

Safari 4 is the first, and only, browser to pass Acid3. On 25 September 2008, a development version of Safari was the first browser to produce a smooth animation on the reference hardware, and thus was the first to fully pass the Acid3 test.[1][2]On February 24 2009, Safari 4 was the first fully functional build browser to pass Acid3.[3] On 08 June 2009, Safari 4 was released to the public as an official browser.

If you have an ideas, let me know. If this doesn't meet with your approval, we could always add another section titled "Passed Browsers".Zerocool3001 (talk) 21:51, 8 June 2009 (UTC)

That's a good idea, introduces a bit of clarity into the article.
My only slight issue, is that I don't really know what a "fully functional build browser" is. Nor will the reader I should imagine as it's not defined in the article that I can tell. Don't forget to state the WP:OBVIOUS and know your WP:Audience.
Otherwise the paragraph seems excellent. ɹəəpıɔnı 09:35, 9 June 2009 (UTC)
I really think there should be a "Passed Available Browsers" section, immediately under the contents. "fully functional build browser" means it's not available to the public yet, so I don't think it should be listed. I was thinking about adding this myself, but decided to post here first. Something like:
Passing Available Browsers
Desktop browsers: Safari 4.0.2
Mobile browsers: Iris Browser 1.1.7
Darxus (talk) 00:53, 13 July 2009 (UTC)
This section should list the first version that passed the test so we don't need to keep updating version numbers. Also, the previous sections should be for browsers that do not pass yet. There's going to be no reason to give the screenshot in all browsers when they all pass. -- Schapel (talk) 12:21, 18 July 2009 (UTC)
Good idea. Added "(oldest version)" to the section's title. Darxus (talk) 19:15, 18 July 2009 (UTC)
What does it mean for a mobile browser to pass? Iris Browser gets 100/100 and lays out correctly without rendering errors at 100% zoom but I don't know of any hardware that it passes the performance test on yet. The problem is that the performance test is designed for PCs. I think the meaning of the test is a bit undefined for mobile browsers. —Preceding unsigned comment added by (talk) 21:49, 20 July 2009 (UTC)

Opera Scored 106/106, first to find an easter egg in the acid3 test[edit]

opera is the first browser to have scored 106/106 in the acid3 test. this should be mentioned the wikipedia article. see the link below for more details. and snapshots. —Preceding unsigned comment added by Fatal eyes (talkcontribs) 15:16, 29 May 2009 (UTC)

Ha! That wasn't an easter egg, it was an April Fools joke (note the date on the article). Perhaps it can go in one of Wikipedia's articles on April Fools. ɹəəpıɔnı 15:52, 29 May 2009 (UTC)

IE Mobile?[edit]

How is there no mention of Microsoft's mobile browser? The user agent of my phone is "Mozilla/4.0 (compatible; MSIE 6.0; Windows CE; IEMobile 6.12) Sprint:PPC6800". I purchased the phone before Acid3 was released, and have not upgraded the browser / OS, so it's perfect for "Screen shot of latest release at time of Acid3 release". The result is "JS ?" (didn't get to 1/100). Darxus (talk) 02:34, 14 July 2009 (UTC)

There are also newer versions for trident on Mobile. There are many WM 6.5 ROM's floating around showing a newest development version of the new WM browser version (currently ROMs based on WM build 23024) which could be used to show development versions. It seems the Zune HD also has an upgraded mobile browser version (not designated as IE thus probably "Zune Browser" would be a good name). However it is unlikely they will perform very good on ACID3 as MS does not see this test as a very relevant test. They probably consider the 7000+ test they submitted to W3C for CSS2 much more relevant for their browser support. hAl (talk) 12:20, 11 August 2009 (UTC)

Opera 10 does not pass Acid3[edit]

I've added Opera to the list of public, stable browsers that pass the Acid 3 test. I noticed on the history that there was a bit of debate about whether or not Opera qualifies, so I decided to justify adding it.

Opera version 10.00 was released today (1 September 2009) as a public, stable release, and there is already a screen shot showing that it passes. I have also personally verified that it passes the test by downloading the new version of Opera and running the test myself. Passed test with 100/100 and perfect clone of reference image; tested with both Windows and Linux versions. Ean5533 ( View! / Talk!) 13:40, 1 September 2009 (UTC)

First, does it pass the performance aspect of the Acid3 test? Second, your checking that Opera 10 passes is original research; we should cite a reliable source instead of relying on original research. Please provide a citation to a reliable source before adding that Opera passes Acid3. -- Schapel (talk) 13:43, 1 September 2009 (UTC)
In terms of performance, yes, it passes the test (smooth animation) as far as I can tell. However, I understand that my original research does not qualify as evidence. I'll wait until ars or one of the other tech blogs mentions it. Thanks. Ean5533 ( View! / Talk!) 14:33, 1 September 2009 (UTC)
Follow-up: This press release from Opera claims that Opera 10.0 passes the test. I assume this qualifies? Ean5533 ( View! / Talk!) 14:47, 1 September 2009 (UTC)
What do you mean by "as far as I can tell"? What is the report given when you click on the "A" in Acid3? If you get any lines stating "Test XX passed, but took XXms (less than 30fps)", that's a fail. The press release you link to says it gets a 100% score. That is different from passing all three aspects of the test. The score is only one aspect of the test. -- Schapel (talk) 14:52, 1 September 2009 (UTC)
Didn't know about clicking the A, and you're right, it does report that one of the tests took too long. I assume that no blogs will report the browser passing completely, so it's fair to leave Opera off the list. However, this brings me to a new question: should the Iris (mobile) browser be removed from the list as well? The reference listed only says that it produces a "pixel perfect" rendering of Acid 3, not that it passes all the conditions of the test. Ean5533 ( View! / Talk!) 15:00, 1 September 2009 (UTC)
If you throw a powerful enough machine at Acid3 on Opera10, it will pass all tests including the performance tests, so therefore it's a pass. Just because it doesn't pass on your machine does not mean it's not compliant. Likewise if I run Safari4 on my 486DX, it will also fail the performance aspects of the test. So does that mean it should also be delisted? There is no critera for what machine you need to run the test on therefore the performance aspects are meaningless. -- (talk) 11:21, 2 September 2009 (UTC)
There is a criterion for the reference platform for the performance test. Please read the article, and you will find out exactly what it is. In any case, we need a citation from a reliable source before we add information to an article. -- Schapel (talk) 11:27, 2 September 2009 (UTC)
You can find discussion of whether Iris passes a few screenfuls above. Someone made the suggestion that the performance aspect doesn't count for mobile browsers, because they cannot run on the reference platform for the performance test. I suggest you discuss the topic up there. -- Schapel (talk) 15:08, 1 September 2009 (UTC)

I've been doing a little more research on the performance of Opera on the Acid3 test (I get bored at work). One thing I noticed is that, in the Wikipedia article, it's specifically mentioned that Opera does not pass the performance aspect of the test. However, there is no citation given to justify that claim, and I haven't been able to find any recent websites that confirm it (plenty of stuff from 2008 talking about alpha versions not meeting performance, but that's old news now). On the other hand, I did find this, which is from July '09. This site specifically mentions that Opera does render the image smoothly (and that in fact Safari does not, but that's not the point). They very clearly announce that Opera passes every aspect of the test.

There are also dozens of sites (like Opera's press release) that discuss its passing score. They do not mention specifically that Opera passes the performance aspect, but I believe it's implied. As I mentioned, I cannot find a single website claiming that Opera does NOT pass the test, but there is at least one that claims it does. As such, I can't see any reason to keep Opera off the list. At the very least, we should remove the statement specifically claiming that Opera fails the performance part of the test without some sort of citation to back it up (posts from 2008 about alpha versions being slow wouldn't qualify). Ean5533 ( View! / Talk!) 13:59, 2 September 2009 (UTC)

The first link you give is to a personal blog, so it is not a reliable source. Furthermore, that site says that Firefox passes the performance aspect of the test, but it does not. The author clearly did not know how to interpret the results of the test. The second link you give again says that Opera attains a 100% score on Acid3, but we all know that the score is only one of the three aspects of the test. Certainly, the Opera developers would say that Opera 10 fully passes if it did pass Acid3. I will add tags to any claims that Opera 10 does not pass the performance aspect, so that editors will know to look for reliable sources. -- Schapel (talk) 14:23, 2 September 2009 (UTC)

Really I would like to understand, clearly and with simple words what do you mean by Opera pass the performance aspect of Acid3. Please, make the test yourself, each one. And show me a pixel that is different from the reference picture. Because I can't picture why Opera 10 does not, and why Safari 4 does. There's no difference. —Preceding unsigned comment added by Youknowvictor (talkcontribs) 16:11, 2 September 2009 (UTC)

The rendering is not the issue. The issue is that the test specifies that each of the 100 subtests must complete in 33 ms on the reference hardware. That is the performance aspect of the test. Please read the article, which explains each aspect of the test in great detail. In particular, the second paragraph of the section "The test" describes the performance aspect. -- Schapel (talk) 17:04, 2 September 2009 (UTC)
I just downloaded and installed Safari for running acid3 on the same machine as opera10. Opera10 failed 3 tests with regard to performance (25 - 281ms, 65 - 78ms and 69 - 6 attempts). Safari failed two tests with regard to performance (26 - 77ms and 69 - 17 attempts). Opera ran all tests in 1.28s, Safari in 1.95s. The PC used to run them is fairly outdated hardware. IMO, under these circumstances, it's simply not fair to say that Safari passed and Opera didn't.
On the other hand, I looked a bit into the test code, and test 69 waits for some resources to download. This means the test is biased by how close the test machine is to, how fast the connection is, and whether you get to access the server on one of its good or less good days. This, coupled with the acid3 test author's specification of the HW reference platforms (first of all, IMO there's an issue with different HW platforms for different browsers), make me think that a notice on the wiki page saying that there are serious issues regarding the performance statements would be appropriate. I know I may be countered with the so often repeated statement that this page is about acid3 restults, not about discussing potential issues with acid3. But IMO wikipedia's purpose and goal is to inform and educate readers. Leaving such a comment out, given the issues, is IMO manipulation of the truth, in two ways: first of all, by not noting the issues related to performance which the acid3 test has, readers are led to believe there are none, second, again, by not noting those issues, browsers favored by the very generous specification of the reference hardware are made to look better than other browsers, even if the resulting classification is debatable.
I found some more up-to-date information in a thread on the Opera forums. It shows that Test 26 takes more than 33 ms. Of course, no one mentions what hardware they're using. The reference platform is specifically a laptop, and a new desktop should run faster than any laptop. -- Schapel (talk) 03:31, 3 September 2009 (UTC)
I downloaded Opera 10.00 and running now, got a result that took about 200ms for subtest 26 and some attempts to pass subtest 69). Tested with a machine - Intel P4 CPU 2.66GHz. --Naturehead (talk) 03:38, 3 September 2009 (UTC)
From the citation given, the reference hardware is "... whatever the top-of-the-line Apple laptop is at the time the test is run." Looking at Apple's current selection of MacBook Pros, that's some fairly serious hardware to be using as a reference. Even the lowest-end macbook pro has a 2.26Ghz dual-core and 2 gigs of ram. The top-of-the-line model has a 2.8Ghz dual-core and 4 gigs of ram, or a 3.06Ghz dual-core and 8 gigs of ram if you choose to customize it. In any case, I wouldn't take those specs lightly. When I tested it on my 2.00 Ghz dual-core desktop, I got 46ms. I would speculate that a 2.8Ghz processor would be pretty close to 33ms, but without a credible 3rd party source we can't say either way. Ean5533 ( View! / Talk!) 13:09, 3 September 2009 (UTC)
Apple laptop? This means the test is HUGELY biased in favor of Safari. Who is making these decisions anyway? (talk) 16:08, 3 September 2009 (UTC)
That would be Ian Hickson, the author of the test, and Apple laptops work well with numerous web browsers. —Remember the dot (talk) 16:42, 3 September 2009 (UTC)
Opera 10 for Windows completes Test 26 in over 50 ms on my new 2.4 GHz Intel Core 2 laptop. That is the best time out of many tries. If top-of-the-line mobile chips have comparable performance, a 3.6 GHz mobile processor would be needed for Opera 10 to pass the test. Perhaps Opera 10 will pass after Core i7 mobile chips come out later this month. -- Schapel (talk) 17:56, 3 September 2009 (UTC)
It's interesting that we're seeing such a wide gap for Opera's results on test 26. (Everything from 46ms to 200ms) Purely out of curiosity's sake, have you run Safari 4 on your machine and seen it pass? I'm not going to try and use the answer as evidence of anything, I'm just genuinely curious. On my work machine (2.33 Ghz Core 2) I get around 46-47ms for Opera and 50ms for Safari. I wonder why the results vary so dramatically? Ean5533 ( View! / Talk!) 18:47, 3 September 2009 (UTC)
I tested both Safari 4 and Opera 10 on Windows, and Test 26 seemed to take more time on Safari 4 than Opera 10. The actual reference platform is an Apple laptop, so perhaps Safari runs faster on Macs than on Windows, just as Opera runs faster on Windows than Linux (on Linux, Opera always took longer than 80 ms for Test 26 when I tested). -- Schapel (talk) 14:11, 4 September 2009 (UTC)

Changing format of "Browsers which pass" section[edit]

Following-up a lot of talk about browsers (not) passing the test, I'm going to be WP:BOLD and change the way we format the "Browsers which pass" section completely. Hopefully this will remove any ambiguity from the section and stop us from arguing over semantics. Even if you don't like the new format, please discuss the change before reverting it! Changes should be up in ~10 minutes. Ean5533 ( View! / Talk!) 15:19, 4 September 2009 (UTC)

Update: Changes have been applied. Please list suggestions for improvement or discuss why the change is not valid. Ean5533 ( View! / Talk!) 15:35, 4 September 2009 (UTC)

The date for Opera[edit]

Why is Opera listed as passing the two first test on 1 September 2009 ? Opera 10 had score 100/100 long before that. In february 2009 and probably before. And for the "pixel perfect test", I don't remember but I'm sure it was passing the test before the final release. Therefore, why putting the date of Opera 10 final release ? Safari 4 final was not released on 24 february 2009. And for the performance test, to this day, on a windows XP 3 GHZ and 2 Gigabyte of ram, neither Safari nor Opera pass the test. I know, the "reference hardware" is a top-of-the-line Apple laptop, which mean that no other browser than Safari will never pass that test. Apple can modify the hardware, the browser and the OS. The others browsers (I'm not talking about IE 8, of course) can only the modify the browser. p!Ral (talk) 13:46, 6 September 2009 (UTC)) —Preceding unsigned comment added by (talk)

Agreed about the date for Safari. It looks like someone already changed the date for Safari passing the test to June 8th 2009, and I think that's a good change.
However, regarding the comments about Safari passing the test because the reference hardware is an Apple laptop -- here isn't the place to complain. Believe me, I agree that using an Apple laptop specifically as the reference hardware is incredibly stupid. In my eyes it makes the performance part of the test completely invalid. However, the guy who created the test outlined those criteria. If we want to use the Acid 3 test as any kind of benchmark, we need to respect the test author's rules. Ignoring certain parts of the test because we don't like them isn't a valid response. You can, however, write to Ian and ask him to provide an alternative hardware spec. Ean5533 ( View! / Talk!) 12:55, 8 September 2009 (UTC)
I'm guessing he chose it because the name is static and to eliminate confusion seeing if he chose a Dell 12345 or an HP 54321, those names could easily change or the model could disappear completely. Plus, if he did choose those platforms, would the test then be biased /against/ Safari? Windows can still be run on a Mac, while OSX cannot (legally--debatable, I know) be run on a non-mac PC. So, I don't see the logic behind using a MBP as the reference hardware makes "the performance part of the test completely invalid". However, to be completely fair (and consistent), he should outline the specs, like, say a 2.6GHz proc and 2GB of RAM, as the reference hardware. -Pyro3d (talk) 19:54, 8 September 2009 (UTC)
I think you already answered your own question. The reference hardware shouldn't be either "an apple laptop" or a specific model; rather, it should be a specific set of hardware specs like you suggested (X.XXGhz processor with X cores, XGB ram, etc). This allows the test to be accessible to all platforms without having to do things like buy a $2500 apple laptop and install Linux on it just to test Linux browsers on Acid3. (I doubt many Linux users are going to want to do that) Ean5533 ( View! / Talk!) 11:31, 9 September 2009 (UTC)

Midori Image[edit]

Erm... even if both Chrome/Chromium/Midori use WebKit, why should the Midori image be instead of (talk) 08:02, 11 September 2009 (UTC)

I've never used Midori, but to me it seems like Midori itself might have interfered with the content, by auto-zooming or something... --Zirrozify (talk) 13:12, 11 September 2009 (UTC)

How useful is it to add yet another Webkit based browser to the list? Not to mention that its market penetration is zero? Add now bit by bit all the QT/GTK/wxwindows/Haiku ports? Rather not. (talk) 14:19, 21 September 2009 (UTC)

We have two options than continue reverting:

  • change paragraph to: "Engine which do not pass", and list only one row per engine version
  • keep the paragraph say: "Browsers which do not pass", and list all notable browser, even if they use the same engine

Notable is not over 5%, take note that only two browsers have a market share of more than 5%.
--Efa (talk) 23:42, 29 September 2009 (UTC)

Three things: The mobile browser section is still not consolidated, perhaps also due to the wild growth in the market. Second: I wrote .5%, and this share is certainly debatable. I proposed this figure because it's roughly the threshold used in for being listed under "other" or for being listed separately. Third: I think it makes sense to mention the main browser of each engine, that's why I'd include Konqueror, even though its market share is quite low. Let me ask the other way around: Where do you see the benefit of adding all the browsers that make use of the webkit engine (or gecko or trident)? Which of the browsers listed in and would you want to see included, and why? Oh, just in case that you'd argue that then Chrome shouldn't be included either, as it's also based on webkit: Chrome's javascript engine is different, and acid3 tests javascript conformity. (talk) 22:52, 30 September 2009 (UTC)

--some note-- chrome has only a different js interpreter, not an own js engine! the same engie, but faster results by optimization! mabdul 23:51, 30 September 2009 (UTC)
that's incorrect, see (talk) 09:32, 1 October 2009 (UTC)
you don't read the article, or? mabdul 00:33, 3 October 2009 (UTC)
So indeed V8 is a separate engine, the techniques employed for obtaining the speedup are also employed in JSC, see (talk) 23:24, 4 October 2009 (UTC)

I'm not understand if you are arguing that other components of the browser make the difference in rendering over the engine alone. If so, the title "Browsers which do not pass" is correct and everyone can add the browser he want. If you say that the engine is enough to pass the test, we are going to remove the "main browser of each engine" and leave the engine alone. --Efa (talk) 23:42, 30 September 2009 (UTC)

You must be kidding. You are in favour of adding the roughly 15 webkit browsers, roughly 15 gecko browsers and hell knows how many trident browsers? Rather not, I hope. Same engine, same Acid result - further no market share -> not relevant. (talk) 08:59, 1 October 2009 (UTC)

I'm not in favor of adding 15 ... browser as YOU say, I never say that. I'm only in favor of complete neutral and right information on this page for the community. So up here I spoke about two situations and a solution for each. I asked to you that deeply know all browsers, if the html engine is the only thing that determine the test pass. Now you've said technically yes, every browser that use a particular version of an engine for you score the same. So you are saying that the browser is not the important part for the test, but only the engine matter, right? --Efa (talk) 23:11, 2 October 2009 (UTC)

Two different browsers using the same http engine can have different scores, as seen in the mobile browsers and safari vs chrome.
I like the way we have it now, the five flagship browsers of the five engines + an exceptional circumstance in chrome, which is its engine's market leader, but not flagship product.
--AlexTG (talk) 01:12, 4 October 2009 (UTC)
While Chrome uses the same html engine, the javascript engine is different, that's what makes in primarily useful for having it in the list. So I agree, it makes sense to keep the list as is. (talk) 23:24, 4 October 2009 (UTC)
Efa, so you want to add not all 15 webkit browsers, but just Epiphany. What is neutral about this random pick? What is special about Epiphany, compared to the other Webkit-based browsers that share a close-to-zero market share? The list of the other browsers is and (talk) 23:24, 4 October 2009 (UTC)

No Lynx score[edit]

Why in both this article and the other two Acid tests, there is no mention of Lynx? --ÆAUSSIEevilÆ 13:27, 28 September 2009 (UTC)

Because Lynx does not support graphics, while the tests require graphics, thus Lynx is inherently outside of the intended targets of the tests? (talk) 01:21, 22 October 2009 (UTC)

Opera Mini[edit]

The Opera Mini section needs some attention. Server version needs to be added since this is what really matters when it comes to web standard compliance. The client has very little influence on the end result of this test. Himasaram (talk) 16:19, 30 September 2009 (UTC)

Gave it a go but could not reproduce anything near the 80% of the currently published screenshot. See File:Acid3 Opera Mini 4.2 server 4.12.png. Himasaram (talk) 19:24, 30 September 2009 (UTC)

The trick with Opera Mini is to repeatedly click the "Acid3" title text which causes a JavaScript popup to appear; upon dismissing the popup, the display will update. This can be done repeatedly until the test completes (which is evident because the popup text and appearance changes). I'm using an Opera Mini simulator, so it would certainly be nice if someone used the browser on a real phone. I've uploaded a new version of the screenshot and included the server version (4.12.998). nneonneo talk 02:34, 1 October 2009 (UTC)
Thanks a bundle - looking good now. Himasaram (talk) 18:19, 2 October 2009 (UTC)

Scored 100/100[edit]

There is no point having the 100/100 column in the browsers that pass if we're only including browsers which have a pixel perfect rendering anyway. Obviously if it has a pixel perfect rendering then it will have 100/100. So we need to either inculde chrome in browsers which pass, or eliminate this colomn. --AlexTG (talk) 04:11, 4 October 2009 (UTC)

That's nothing — the reference "rendering" is not a raster image, so it's impossible to determine what is and isn't a pixel perfect rendering. ¦ Reisio (talk) 04:19, 4 October 2009 (UTC)
I understand what you're saying about the pixel-perfect rendering automatically implying a score 100/100. However, what I do not understand is why you think including the score 100/100 column means that Chrome should be added to the list. Chrome is not on the list because it doesn't produce a perfect rendering. The 100/100 column has absolutely nothing to do with it. (talk) 7:21 am, Today (UTC+11)

Because chrome gets 100/100, and 100/100 is the first column and hence logically should mean inclusion. Nowhwere is it stated on the article that pixel perfect rendering is what makes a browser get into the list. Can't just make up rules as we go along, if 100/100 is in the table then chrome should be included.

--AlexTG (talk) 23:02, 5 October 2009 (UTC)

Either that or Opera should be removed from the list. --Execvator (talk) 23:41, 6 October 2009 (UTC)