Inside The CSS Working Group


The IE team is active on several  W3C working groups such as SVG and HTML. As one of our three regular CSS Working Group (CSSWG) representatives, I wanted to follow up on the latest face-to-face meeting the group held at Apple in Cupertino at the end of last month by sharing some of the work and progress being made. While it aims to be representative of the three-day meeting, the list below is not exhaustive.

  • CSS2.1 and the CSS test suite: the working group discussed many of the remaining open issues. Elika Etemad (fantasai), a CSSWG Invited Expert who consults for Mozilla, is now also working with Microsoft on the completion of the CSS2.1 test suite.
  • Vendor prefixes: the CSS specification requires browser vendors to add a vendor prefix to property names when a feature is either:

    This convention avoids name collisions between standard and proprietary features; it also enables browser vendors to gain implementation experience and gather valuable feedback from early adopters without impeding design progress on the specification. In practice, this can also result in authors writing multiple declarations of the same property. A common example today would be:

    -moz-border-radius: 10px;
    -webkit-border-radius:10px;
    border-radius:10px;

    Following up on feedback submitted to the www-style mailing list, the WG has begun discussing whether this convention needs to change to ease the authoring burden and support an accelerated pace of standardization for CSS features.

  • Transitions and Animations: as well as a number of improvements on the current Transitions and Animations specifications, the CSSWG also followed up on mailing list feedback and proposals that aim to bring both features together both in terms of their capabilities and syntax.
  • Fonts: with better cross-browser support for CSS3 Fonts and web font formats, work has begun to expose advanced typographic features through CSS.

On behalf of the CSSWG, we very much welcome the feedback and insights of web designers on www-style.

I am looking forward to the next CSSWG meeting at Opera in Oslo this summer. Regular engagement with the W3C, developers of other browsers, spec editors and experts invited from the broader community is not just exciting; we strongly believe it is necessary to advance the state of the web for all users.

Sylvain Galineau
Program Manager

Edit 11:26am: correcting the link to proprietary extensions definition. 

Comments (17)

  1. Anonymous says:

    « file:///C:/Users/sylvaing/Documents/Downloads »

    Whoopsie?

  2. Wurst says:

    I am excited about the enhanced support for many OpenType font features!

    I wonder if in 10-20 years time a common web browser will be a viable alternative to QuarkXpress, InDesign, Publisher, LaTeX and Co.

  3. Transitions and animations making it in to IE9? Now that would be pure icing on the cake! Custom fonts via CSS3? That would make me drool too!

    I think it’s important that vendor prefixes exist. First and foremost as I’ve been implementing CSS3 properties for all browsers that support them and I’m heavily relying on object detection to determine which browser is which so even if the browser is spoofed it’ll load the correct compiled CSS3 style sheet. There are some differences between how certain properties have been implemented such as one (can’t remember offhand what it is, maybe gradients?) between Gecko and WebKit. As each standard (or in the case of CSS3 each module) becomes a final specification (or stable enough as far as the industry is concerned) then yes, I’d prefer to see those properties become standardized in browsers. For those who would like to see vendor prefixes disappear altogether forget it! There just isn’t any justification for dumbing down the web and making our lives more difficult.

    Awesome post though you know I’d really like to see the W3C open a forum, mailing lists aren’t my thing.

  4. s5s5 says:

    请问可以快点出新版本?

    安装和升级可不可以快一些,比如一周一个新版本发布!

  5. @Sylvain Galineau [MSFT]

    CSS 2.1 and the CSS test suite

    ==============================

    IE8 fails over 150 testcases submitted in the CSS 2.1 test suite (alpha 2, build 20100316). What I’d like to know is

    – if these failures are fixed in the current IE9 platform preview

    – or if these failures will be fixed for the next IE9 preview

    – or if they are being tackled so that they will be fixed within the IE9 beta development.

    I am not referring to CSS 2.1 current issues or spec clarifications here but to undisputable, unquestionable CSS 2.1 testcase failures.

    Several testcases (say, about 100)  submitted by Microsoft to the CSS 2.1 test suite have inaccuracies or are not well/best written. Will they be corrected? Are they being worked on?

    Vendor prefixes

    ===============

    Both D. Glazman and you missed -khtml-border-radius used by Konqueror.

    Personally, the sooner browser vendors adopt the non-vendor-prefix for property names, the better it is for everyone involved.

    Transitions and Animations

    ==========================

    To me, this is just more "bells and whistles" which can be misused, abused, overused and which is often [unneedlessly] user system resources demanding. Content is and will still remain king.

    regards, Gérard

  6. Gaurav says:

    I think that despite having to retype some css rules a few times, vendor prefixes are a good idea and shouldn’t be messed with. Right now, there are differences in syntax with the webkit and mozilla implementations of linear gradients and a common prefix would have created problems. The experience we are having today with having to type the same rule three or even four times is much preferable to having to use some css hacks to target browsers with different implementations of CSS3.

  7. carlos says:

    @Sylvain Galineau

    "The IE team is active on several  W3C working groups such as SVG "

    IE9 preview fails in 72% of SVG ( the other browsers fails 15-20% ).

    Would you share with us the IE team goals regarding SVG support in IE9? Do you target 40% of compliance? 50% 70% ?

    Thanks in advance

  8. infinte says:

    I think seamless interops with Silverlight will be great.

    And <canvs> or <sence>!

    And NO MORE COM !

  9. hAl says:

    @Gerard Talbot

    Where is the list of those 150 failed tests?

  10. hAl says:

    @carlos

    As you probalby read on this blog before the SVG implementation scheduled for IE9 is not yet completed and is significantly more than is currently available in the preview.

  11. Victor says:

    Thank you for sponsoring WOFF. I just want to share my desire for WOFF fonts to be available in IE9. You’re doing nice.

  12. kotonfajer says:

    My suggestion for vendor prefixes: add something like

    vendor-prefixes: ms, webkit, moz;

    use-vendor-prefixes-for: border-radius, border-style;

    to CSS

    / so if there is "border-radius" in CSS -> browser will try all vendor prefixes

  13. Francis Hemsher says:

    @ "…the WG has begun discussing whether this convention needs to change to ease the authoring burden and support an accelerated pace of standardization for CSS features."

    I surely hope all browsers include a default "px" for style size/locations. This perticular point is the biggest problem for the migration of legacy sites to standardization. Also, It will remove the mind-numbing process of including "px" for web developers in the future.

  14. Wurst says:

    @Francis: Are you serious? *eyeroll*

  15. Randall says:

    Maybe it would help if the W3C said it’s OK for vendor A to support vendor B’s proprietary CSS extensions.  Then if websites are using -moz-border-radius but not border-radius, IE 9 can still parse it.

    Of course, if the -moz-border-radius implementation differs in small ways  from what’s finally standardized in CSS3, Microsoft would have to choose whether or not to try to parse the old syntax.  It probably depends, case by case, on the value of compatibility to authors and users versus the complexity of implementing it.

    The downside for authors is that -moz- no longer means a declaration is guaranteed to have no effect on non-Mozilla browsers.  You can imagine an extended CSS attribute being "ruined" by having drastically different effects on different browsers, forcing browser-specific hacks.  

    But these extensions are, by definition, nonstandard and experimental.  If authors want predictability, they should stick to the finalized standards.  Microsoft or any other vendor shouldn’t feel constrained from parsing a CSS attribute if it typically improves their renderer’s fidelity to the author’s intent.

  16. Wurst says:

    « If authors want predictability, they should stick to the finalized standards. »

    Will they? That is the question. And what has history told us?

  17. FrankV says:

    I still have no clue on the question "Will webpages look great in IE9 when these pages are marked up in CSS3". The results from FindMeByIP ( http://www.findmebyip.com ) are an embarrassment for IE8 and Microsoft.

    The second remark: when will IE9 be released and what can we expect from it. Instead of focussing on "new" things like video, most webpages are basically made out of text and image boxes (with borders and the like). So what can we what can we expect from IE9 on those isues. Do I need hacks or will plain html and css be all.