CSS 2.1 Implementation Reports


I want to provide an update to my last blog post about the W3C process that we follow to develop and finalise Web Standards. The working group published the Release Candidate (RC) of the CSS 2.1 Test Suite on September 17. The next step is for the major browser vendors to submit their implementation reports using the working group’s template within one month from publication of the test suite. The group has set a deadline on October 18.

While each of the browser vendors works through the test suite there has been a flurry of activity on the CSS working group’s mailing list. As a result, the group updated the test suite at the weekend to incorporate the feedback so far. On Wednesday, Microsoft submitted the CSS 2.1 implementation report for IE9 Beta showing that IE9 passes 18960 of the 19403 tests (97.7%). You will be able to find the reports for other browsers here after they are submitted.

It’s taken a lot of hard work in partnership with others in the working group and the community to get to this point. Many people in the IE team have been involved in this project and we’re excited to be almost at the end. It feels like a long time since we submitted our first batch of CSS 2.1 tests alongside IE8 Beta 1.

While we’ll be glad to see CSS 2.1 finished, the real excitement comes from the knowledge that the CSS working group can focus attention solely on the new CSS3 modules that bring richer graphics to the web. Each of these modules has a dependency on finishing CSS 2.1.

Look out for the implementation reports coming from other browsers. These reports will be used to move CSS 2.1 to Proposed Recommendation status providing developers with a stable and interoperable CSS platform. I will blog again when we have more progress to report.

Adrian Bateman
Program Manager

Comments (37)

  1. Arieta says:

    Does this mean that CSS2.1 will get updated in IE9 so it passes the rest of the tests?

  2. Andrew says:

    How does 97.7% compare to 'competitors' such as Gecko (Firefox) and Webkit (Chrome/Safari)?

    I've run all the versions of IE 9 so far and they are much better than IE 8 and previous versions BUT getting past (or completing implementation and passing with high marks) CSS 2.1 is not a halting point for the other browsers. Rather, these browsers implement some new CSS 3 features or wannabe HTML5/CSS 3 features, such as Apple with their <canvas>. I question IE only because it never implemented any of these, not even text-shadow (which both of these engines support without a prefix such as -moz). That was CSS 2.1.

    Where do you want to draw the line? Other browsers want no such line and will fix bugs as time goes along with patches. Although I do not use IE other than for testing as a designer, I still feel very limited in capability in IE and I do not see what would be so bad as to release a patch for IE 8 that made text-shadow available, or opacity, or simple ones like that. We already know how degrading filter CSS can be.

  3. CvP says:

    @Andrew:

    "You will be able to find the reports for other browsers here after they are submitted."

    "…..after they are submitted.'

  4. war59312 says:

    @Arieta Let's hope so but for some odd reason I doubt it. 😉

  5. Socrates says:

    How ofter do you check the feedback on Microsoft connect?

    I posted a bug a week ago and still no comment, has someone checked it out?

    connect.microsoft.com/…/vertical-scroll-bar-on-right-is-duplicated-in-silverlight-app

    This app works in all other browsers except in IE9.

  6. Brön says:

    @Andrew: text-shadow is CSS 2.0, not 2.1

    I thought IE8 was pretty complete in its CSS 2.1 support but seems like IE9 will be even completer. 😀

  7. Adrian Bateman [MSFT] says:

    @Arieta – The results we've submitted are for IE9 Beta. We're still working on bug fixes and I expect IE9 to improve further.

    @Andrew – We're not stopping with CSS 2.1 but it is an important foundation for the CSS3 modules. The various CSS3 drafts depend on CSS 2.1 getting to Recommendation before they can become final standards too. You can see examples of some of the CSS3 features in IE9 at http://ietestdrive.com/.

  8. Mitch074 says:

    passing a bunch of compliance test is one thing – one _very_ neat thing, mind you.

    However, in IE 8 and (to a lesser measure) IE 9, the problem is more of breakage in specific situations, such as:

    – not recognizing hover events if a CSS '-ms-filter' is applied to a block (IE 8)

    – list-style-position:inside broken if list uses image for marker (IE9pre3)

    Can we hope that in IE 9, when such bugs are discovered, they'll be fixed without having to wait for IE 10?

  9. Peter Beverloo says:

    Do you intend to pass the current RC of the testsuite before shipping the final IE9?

    WebKit currently passes 95.8 percent of the suite, although they might decide to fix some larger issues before submitting their report. Presto (Opera) and Gecko (Firefox) will publish their results as time passes :)

  10. Meni says:

    Warning: Totaly off-topic (sorry OP)

    Google's answer to IE9's fishtank demo:

    http://www.youtube.com/watch

  11. Taciturne says:

    Google: -1 for non-standard, -2 for deliberately rendering in Quirks mode

    A little-less off-topic, the beta's "Send feedback" is a nightmare

    No back button

    No mandatory fields

    99% odds of duplicates

    Live! sign-in assistant fiasco

    Just send them over to Connect directly, like you did with platform previews

  12. SG says:

    Socrates, I also doubt they look at the reports (why were asking for them though??).

    Mine is over here: connect.microsoft.com/…/iimgctx-fails-to-draw-after-installing-ie9-beta

    Same story – not a word in response, not even any sign of life..

  13. Wheret says:

    SG, you should wait at least a month before complaining, this is their average response time.

  14. Eeny says:

    @Meni

    That's a Youtube video, not a working demo. I tried the link to the actual demo and it doesn't even work in the latest Chrome Beta (7 beta). That demo will be relevant when they have it working in their latest beta (not some random nightly or trunk build) and on a wide variety of hardware with consistent results. If your browser demo starts requiring a $100+ dedicated graphics card then it's really more a test of your graphics system and not your browser.

    Besides, WebGL isn't even a W3C spec (yet they rip off the W3C's document template to make it appear as such). WebGL is the antithesis of open standard–it's directly tied to one specific technology. If it were an open standard then we'd have a implementation-agnostic 3D canvas API…..not OpenGL ported to the web masquerading as a "standard."

  15. Mike Dimmick says:

    SG: Considering you're using an internal IE component that doesn't appear to have been intended for public consumption – never having been documented in MSDN Library in the 12 years that it has existed – it's not surprising you haven't had an answer. Microsoft are interested in bugs in rendering HTML and in the usability of the browser itself, not the knock-on effect on other applications using components that weren't ever meant to be used by anyone else.

    'Works by fluke' is not a sensible software design strategy. Use a component that was meant to be used by third parties.

  16. Damien says:

    You actually did better than you are reporting in this article. After looking at your submitted results, you only failed 383 of the tests for a total of 19343 tests. So the final numbers should be 18960 of the 19343 tests (98.02%)

  17. Aethec says:

    Damien >> The results aren't only "pass" and "fail" – there are also 60 "?" 😉

  18. Damien says:

    Ahhh…in that case I count 63 "?" for a total of 19406 tests…either way good job.

  19. SGX says:

    Mike Dimmick, you're not in IE team, do you? If you are really not – thanks for "defending" them (I doubt they need it though) and all the teaching words, but I'd really like to hear something from them, not from other people, even very talented and experienced ones.

    So, if you know somebody from IE team, or know how to reach them, that's where I'm asking for your help. Basically all I wanted is their confirmation that this component is not going to work in IE9 anymore. It is really that simple – no strategies, no philosophy, no long discussions.

    Thank you.

    Bye.

  20. John Le Drew says:

    Just wondering what the IE Team's response is to this blog post? css3wizardry.com/…/ie9-is-the-ie6-of-css3

  21. CvP says:

    @John Le Drew

    Where do you get links of this kinda shitty blog post? I call the writer of that post a _retard_ who knows *** about web.

  22. A Tester says:

    Any idea when you guys will update the VPC Images?

    Those currently avilable for download at: http://www.microsoft.com/…/details.aspx have an expiry date of 01 October 2010…

  23. David S says:

    Does IE9 have a limit on included CSS files like previous versions? I think older IEs stop respecting any CSS in a file after the ~288kb point. Other browsers are capable of including one big file instead of breaking them up into smaller pieces.

  24. Aethec says:

    @John Le Drew

    How could this guy have "the latest Beta 4 of IE9" ? He's an anti-MS fanboy…what he's using is the PP4…

    Besides, comparing Safari and Chrome, two WebKit-based browsers, is pointless.

  25. johnnyq3 says:

    @John Le Drew

    If one were to take a look at the code for the CSS, it would probably have the webkit prefix on the selcetors.

  26. Harry Richter says:

    @ A Tester

    Just go to Download Center on the MS Homepage and download the VPC images that have an expiry date of  January 11, 2011

    http://www.microsoft.com/…/details.aspx

    Harry

  27. A Tester says:

    @ Harry Richter

    Thanks. It appears that the images, and related web page, were updated shortly after my original post, as the link you provided is the same as mine… Anyway, appreciate the reply!

  28. Bert says:

    @ Aethec

    Isn't IE9 supposed to be all about CSS3? If that's so, then why wouldn't you want to compare Firefox, Webkit or Opera's CSS3 support to IE9? Sounds like you're the one with a sacred cow. From what I'm seeing, IE9 has great backend support for under the hood CSS3 stuff, but not much for the fancy CSS3 visual stuff that the other browsers have been doing for a while. That means all those sites with the cool CSS3 features don't work on IE9. Of course you could do some browser sniffing and give IE9 a canvas/SVG version, something I'm considering.

  29. Bert says:

    @ Aethec

    Isn't IE9 supposed to be all about CSS3? If that's so, then why wouldn't you want to compare Firefox, Webkit or Opera's CSS3 support to IE9? Sounds like you're the one with a sacred cow. From what I'm seeing, IE9 has great backend support for under the hood CSS3 stuff, but not much for the fancy CSS3 visual stuff that the other browsers have been doing for a while. That means all those sites with the cool CSS3 features don't work on IE9. Of course you could do some browser sniffing and give IE9 a canvas/SVG version, something I'm considering.

  30. Aethec says:

    @Bert

    Comparing [some other engine]'s CSS3 support to IE9 is a good thing – but showing the same page in Safari and Chrome is not a comparison since they both use WebKit. (or maybe he's using -webkit- attributes just to bash IE9…)

    And if that guy wants his "test" to be useful, he should give away the source code, and open a bug on Connect about the unsupported features.

    Doing nothing but complaining is useless.

  31. Me says:

    Well, since IE has fallen apart in terms of standards and recent reports showing that its usage has also dropped to 50-60 per cent (depending on country) our company has decided to build Websites following the W3C standards with a sprinkle of HTML5 features for video/audio, advanced forms (input types and placeholder) and CSS3 for presentation like text-shadow (still missing in IE 9), box-shadow (has been in Firefox, Chrome, Safari for about 2 years and finally arrived in IE 9), border-radius (has been in Firefox, Chrome, Safari for about 2 years and finally arrived in IE 9), linear gradients (still missing in IE 9) and serving a scaled down user experience to those people still using Internet Explorer. Pretty much the same behavior Microsoft showed for about ten years when you are using their Exchange and SharePoint products with a non-MS browser.

    Our development and testing time has dramatically decreased after focusing on W3C standard compliant Website development with some minor exceptions for non-standard browsers like Internet Explorer and our management is very happy with it. Oh, and we rolled out Google ChromeFrame (and enabled our Websites to use it) to get an even better user experience without waiting for IE 9 (and going through the upgrade hassle of Windows XP to 7).

  32. Gérard Talbot says:

    @Adrian Bateman [MSFT]

    In the IR, it's written:

    # Internet Explorer 9 beta – Windows 7.

    Is it beta 1? beta 2? How about indicating the build number?

    I can not find IE9beta result for background-color-176 testcase in your Implementation Report. It was in the /other directory as it had the HTMLonly flag.

    margin-left-001 : we reported this one during IE8 beta development (well before IE8 was released) and today, IE9beta (still) can not pass this testcase.

    In

    lists.w3.org/…/0030.html

    , I explained why vertical-align-115 and vertical-align-116 are invalid testcases and today they still are IMO.

    97.7% is an excellent score… but there are several dozens of testcases from Microsoft which need to be corrected, improved, rejected or recoded or reworked, etc.

    regards, Gérard Talbot

  33. hAl says:

    I see results "pass", "fail" and "invalid" in de testreport

    What does the result "invalid"  mean in the testreport

  34. Adrian Bateman [MSFT] says:

    @Gérard – thanks for the feedback on the implementation report. We will update the report after the next test suite update and at that time we'll include the IE9 build number. That's a good idea. I think there was a problem with background-color-176 not being built but that should be fixed in the next update of the test suite, which will most likely be at the end of this week.

    It's probably best to follow-up on the mailing list to discuss which tests are invalid or not. Vertical -align-115 and -116 are on our list for review and we plan to get to them this week. I expect that all the comments made to the list regarding corrections, improvements, rejections and recoding will be addressed this week.

    @hAl – invalid means that we believe the test is invalid and are following up in the working group.

  35. hAl says:

    I can see only the Safari and IE9 reports on the location listed in this article

  36. Richard Fink says:

    +1 to Gerard Talbot's call for a Beta version number. Is there going to be any kind of auto-notification within the first beta when the next is released?

    In the meantime, what does this mean for bug reports. Should we just consider it an Alpha and wait?

  37. albert says:

    Where is the submitted data of Mozilla, Opera, Google, any Linux distro?

    Does this "deadline" of Oct 18 have any consequences if not met or is this just an excercise for the W3C?