Share via


Additional Tests Submitted to the W3C CSS 2.1 Test Suite

It’s been just over five months since the MIX08 conference and IE8 Beta 1. One of the things I remain committed to is the furthering of web standards through a comprehensive test suite for each standard. This is necessary to eliminate ambiguities or differences that cause implementation differences between user agents (aka browsers). Those differences create frustration for web developers who are just trying to build web sites that interoperate.

The IE team has been actively working on Internet Explorer 8 Beta 2. In parallel with the CSS 2.1 implementation in the upcoming beta, the IE Test team has been developing test cases against the CSS 2.1 specification. Today we’re happy to announce that we’ve submitted an additional 2524 more test cases to the W3C for inclusion into the CSS 2.1 test suite. This brings the test suite much closer to the necessary breadth needed to ensure that web sites will interoperate. These tests are available on the IE Development Forum until they are fully reviewed by the working group and accepted into the official test suite.

I also want to thank everyone that provided great feedback on the tests we submitted back in March 2008. Based on the feedback on the W3C’s CSS 2.1 Working Group’s mailing list and my March IE Blog post on the subject, we made corrections and design changes to 28 of the 702 test cases we submitted in March. We also deleted 5 cases that became redundant through the other 28 changes. These updated tests are also available on the IE Development Forum until the W3C integrates them. It is this collaboration with the web development community and the W3C that will really make these web standards more reliable and able to create a more predictable web development experience.

This brings Microsoft’s contribution in this suite to 3221 test cases and the entire W3C CSS 2.1 test suite to 3708 test cases. We, the IE team, will continue to work closely with the CSS working group on these tests and listen to any feedback you provide.

In addition to the CSS 2.1 standard, IE8 is supporting the new Accessibility Rich Internet Applications (WAI - ARIA) draft standard in development by the W3C. It provides a way to create web sites that are accessible to people that need Assistive Technologies to help them live and work. We’re using some of the existing test suite to validate our implementation. We also just submitted our first tests to the working group for inclusion into the test suite. They are also available for download on the IE Development Forum until they get included into the W3C test suite. As with the CSS suite, we will continue to work closely with the WAI – ARIA group.

Thanks,

Jason Upton
Test Manager
Internet Explorer

Comments

  • Anonymous
    August 19, 2008
    PingBack from http://hubsfunnywallpaper.cn/?p=1406

  • Anonymous
    August 19, 2008
    "The IE team has been actively working on Internet Explorer 8 Beta 2." So do you have an estimate for release better than "soon" or "in August"? Is it still coming in August? Inquiring minds want to know! :)

  • Anonymous
    August 19, 2008
    There are twelve days left in August... I'm sure they're busy...

  • Anonymous
    August 19, 2008
    i'm going to slit my rists if beta2 dosen't come soon.

  • Anonymous
    August 19, 2008
    Microsoft I hope IE8 Beta 2 is going to be worth the wait!!! Don't let us down!

  • Anonymous
    August 19, 2008
    I'm a little worried about the restrictions I've heard about in IE8's implementation of data: URLs. Are they really ONLY allowed for image URLs in CSS? Will data: URLs in HTML/XML be ignored? That would be a big disappointment.

  • Anonymous
    August 19, 2008
    The comment has been removed

  • Anonymous
    August 19, 2008
    now that the win dev team is open to suggestions why dont they make a better download manager for win. 7

  • Anonymous
    August 19, 2008
    The comment has been removed

  • Anonymous
    August 19, 2008
    I test my sites on browsershots.org and IE8 has consistently broken tables leaving gaps. IE8 is the ONLY browser that does this. Please see browsershots.org for a complete list of browsers.

  • Anonymous
    August 19, 2008
    This is excellent work - thousands of tests are a real indication of a hopefully solid browser release. Thank you for making these public.

  • Anonymous
    August 19, 2008
    IE8 beta 2 for the next week ? I hope for that !

  • Anonymous
    August 19, 2008
    I'm glad to see Microsoft are doing all of these to help us see that they care about web standards now. I can't wait to see IE8's new GUI. I here that it will be better.

  • Anonymous
    August 19, 2008
    Jason, lots of hard work evident there! Well done.

  • Anonymous
    August 19, 2008
    That's sweet! Reading the testcase titles, one can almost feel Beta 2. I mean, some seem to indicate a known bug was fixed. Hm, that's evil. Last time Beta was released before the testcases were! But today's wednesday, so I'm still feeling ok. Hope to hear from you again soon.

  • Anonymous
    August 19, 2008
    @Gérard: Deprecated: "A deprecated element or attribute is one that has been outdated by newer constructs. Deprecated elements are defined in the reference manual in appropriate locations, but are clearly marked as deprecated. Deprecated elements may become obsolete in future versions of HTML. "User agents should continue to support deprecated elements for reasons of backward compatibility."

  • from http://www.w3.org/TR/html401/conform.html#deprecated. IE must continue to support elements that are deprecated, if not obsoleted, if the document type supports it (HTML 4.01 Transitional does, Strict does not). The declared XHTML 1.1 inherits HTML 4.01 Strict. Secondly, the purpose of the test is to ensure that the HTML 'align' attribute is being overridden by CSS and in that regard is a suitable test. It could be fixed easily by changing <caption> to <tr><td>. The document does not validate only because of the deprecated attribute (not in HTML 4.01 Strict) and because <table> should have at least one <tr>, and <tr> at least one <th> or <td>.
  • Anonymous
    August 20, 2008
    This brings Microsoft&#8217;s contribution in this suite to 3221 test cases and the entire W3C CSS 2.1 test suite to 3708 test cases. We, the IE team, will continue to work closely with the CSS working group on these tests and listen to any feedback y

  • Anonymous
    August 20, 2008
    where Internet Explorer 8 Beta 2????

  • Anonymous
    August 20, 2008
    @Brianary: No, DataURIs may be used for more than just images in CSS files, although from a performance perspective, that's pretty much the only place their use makes sense. See http://code.msdn.microsoft.com/ie8whitepapers/Release/ProjectReleases.aspx?ReleaseId=575 for an explanation of the actual enforced restrictions on DataURIs; the restrictions are primarily for security reasons.

  • Anonymous
    August 20, 2008
    While I applaud the IE team for bringing IE8 into better compliance with the CSS 2.1 spec, I have to wonder at the lack of attention to CSS 3.0. Although IE8 will provide limited support for CSS 3, I'm concerned by the lack of specific features such as CSS 3 pseudo-classes and RGBA color notation (see CSS Compatibility and Internet Explorer - http://msdn.microsoft.com/en-us/library/cc351024(VS.85).aspx) Firefox, Opera, and Safari continue to provide increased CSS 3 support with each new version and interstitial release. Based on the previous IE release model, I don't expect to see  significant 8.1, 8.2, 8.3, etc. releases, so getting as much CSS 3 support into the IE8 release now would set the stage for the next few years.

  • Anonymous
    August 20, 2008
    Great work to the team! Not the IE team but the team at mozilla involved in creating a plugin so canvas is supported in IE. http://digg.com/software/Mozilla_Drags_IE_Into_The_Future_With_Canvas_Element_Plugin far from ideal but I guess it stops you guys from continuing to hold back the web. Just how many requests did you get for CANVAS and SVG support?

  • Anonymous
    August 20, 2008
    @Mike: I'm guessing thousands of requests, but why would they support SVG and CANVAS properly, when they can push SilverSlieght, VML and other proprietary stuff? I've been playing with the test suite... but I was wondering if you had published stats on IE's coverage? e.g. I viewed many of the tests in IE6...  lots of fails... in IE7 (ever so slightly less fails)... and IE8Beta1 (well, I'm running batches of tests by chapter... and certain chapters crash IE8Beta1). Will you be publishing how many tests IE8Beta2 (or whatever you have internally) pass? How about IE8Beta1? thanks!

  • Anonymous
    August 20, 2008
    the "mozilla canvas" add-on isn't done, is not signed, is not available for download, has scary security implications, provides incomplete support for a non-accepted RFC, and uses functionality for which Apple has filed patents.  See http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-March/010129.html epic fail.

  • Anonymous
    August 20, 2008
    Ted: yeah, because the company who doesn't deign to implement such features in the first place (or at least the finished ones) isn't the one "epic failing".

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    what part of "Apple holds a patent on CANVAS" do you fools not understand?

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    Well, it's great to see this effort towards standardization coming from Microsoft. Everybody's benefiting from that.

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    Morten : They implement an unfinished ACCESSIBILITY draft agains an mainly UNUSED finished standard to concentrate on USED CSS 2.1 and get it right.

  • Anonymous
    August 20, 2008
    daniel: Glad (but skeptical) to hear it.  Cite your sources, please?

  • Anonymous
    August 20, 2008
    @Ted, First the HTML 5 draft by the WHATWG (http://www.whatwg.org/specs/web-apps/current-work/multipage/) which states on its cover: "© Copyright 2004-2008 Apple Computer, Inc., Mozilla Foundation, and Opera Software ASA. You are granted a license to use, reproduce and create derivative works of this document." I don't know where the W3C documents are, but the HTML Charter of March 2008 (http://www.w3.org/2007/03/HTML-WG-charter#patentpolicy) states: "This Working Group operates under the W3C Patent Policy (5 February 2004 Version). To promote the widest adoption of Web standards, W3C seeks to issue Recommendations that can be implemented, according to this policy, on a Royalty-Free basis." The publication of Working Drafts, AFAIK, also pushes the Companies to talk about patent issues, yet Apple did not react. I think the conclusion is that canvas is no longer a problem patent wise.

  • Anonymous
    August 20, 2008
    Spike: Are you claiming that the css opacity is an unused property? Do you really think the filter is the way to go for opacity? (last I heard from the IE team was that they are likely going to introduce a third and new proprietary method for opacity, so not even filters will work!)

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    @Gérard Talbot: It's great that you collect so many bugs, but don't you think reporting them all in the bug database would make much more sense than reporting them in the overpostes comment section of a blogpost? Of course, the questions about the testcase sumission still fit here.

  • Anonymous
    August 20, 2008
    "The publication of Working Drafts, AFAIK, also pushes the Companies to talk about patent issues, yet Apple did not react." ...but Apple DID act.  They specifically claimed IP in a public forum for the working group, as per the url sent previously. now, Apple seems to currently be open to RF licensing (http://www.w3.org/2004/01/pp-impl/40318/status#current-disclosures) but as I read the Patent Policy, they can revoke that promise at any point up to 90 days after the spec becomes a recommendation.  So, we have quite a while left before we know whether or not Canvas is encumbered.

  • Anonymous
    August 20, 2008
    Oh, I need to take a closer look into this. Though I doubt, Apple will really encumber canvas as this would really hurt the companie's image.

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    but then again the google canvas has the same downfall. as for mozzerella foundations project i could not find any ActiveX pluging anywheere for it. so it looks like the projects a flop!

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    >"Dean, GM of IE Team even had the gall" Lie. >"Canvas standard will hurt them far more than trying to enforce the so called "patents"" Your use of the phrasing "so called "patents"" suggests that you know very little about either the legal system or the business world.

  • Anonymous
    August 20, 2008
    @Ted_the_EPIC_FAIL Go complain on /. with all your buddies instead of wasting space here with your vitriol

  • Anonymous
    August 20, 2008
    Whatever the case, Mozilla is sort of wasting their time, seeing as IE would have gotten Canvas support in a few years. I cannot wait to see IE8; it is sort of like the runner who was last until the home-stretch when he put on the speed. He hasn't caught the other players quite yet, but he will soon. If not IE8, IE9 will be one hum-dinger of a browser. Just wait and see!

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    http://en.wikipedia.org/wiki/Canvas_(HTML_element) "Apple later disclosed the patents under the W3C's royalty-free patent licensing terms[8]." Looks like Ted just made a fool of himself, no doubt not doing any proper research.

  • Anonymous
    August 20, 2008
    Microsoft doesn't even support CSS3, SVG and other standards in IE8, what makes anyone think they'll catch up to Firefox or other web browsers or be a hum-dinger of a browser? Inferior Trident engine and lazy devs who didn't do any work between 2001 to 2006 is not gonna make a browser any good.

  • Anonymous
    August 20, 2008
    IE8 Beta 2 is about to come? I am waiting for that.And thanks a bunch to you all~ .Don't let us down. Come on !!!

  • Anonymous
    August 20, 2008
    Hii.. According to the one of the website MS is going to release IE8 Beta2 on Aug 28 and Final version on Nov 2008. So.. only 7 days to go..

  • Anonymous
    August 20, 2008
    Hii.. According to the one of the website MS is going to release IE8 Beta2 on Aug 28 and Final version on Nov 2008. folks who already have IE 8 Beta 1 installed will get Beta 2 via Automatic Updates So.. only 7 days to go..

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    @Nikhil: Who said that? And where? Pleas tell me the source! I'm so sad, I hoped IE8b2 would be released this week... Wish they'd written September instead and released it in the first week of that month. Now I've had a sad August.

  • Anonymous
    August 20, 2008
    @Daniel chk the link.. http://blogs.zdnet.com/microsoft/?p=1543 But y u sad if it Release on late August.. as MS already told that IE8 Beta2 will be drop by the end of August..

  • Anonymous
    August 20, 2008
    @Ted_the_EPIC_FAIL Thank you for your insight.

  • Anonymous
    August 20, 2008
    The comment has been removed

  • Anonymous
    August 20, 2008
    @Nikhil: Thanky for the Link, I didn't know of that article. I'm a bit sad or disappointed because I'm currently on vacation. It wasn't intended to cross the release of IE8b2, but when I found out I thought that this could be a chance to do some hitting on the software. I'd like to improve the CSS support as much as possible.

  • Anonymous
    August 20, 2008
    @Daniel I m Hoping alots of features, improvment & great UI. that mke IE8 a great browser ever seen.. I mke my finger cross...

  • Anonymous
    August 20, 2008
    @Bob Holt:  I believe the reason for limited CSS3 support is that IE8 intends to fully master accurate rendering of CSS2.1 first.  This is no small task.  Firefox 2, for example, came out well before IE7 and had vastly superior CSS2.1 rendering compared to IE6 offering many more features.  Dispite that superiority the FF2 rendering of CSS2.1 was not complete.  FF2 contained a very serious float wrap bug and failed to implement display:inline-block.  Even with the many more obvious failings of IE6 it did, at least, get those correct. In other words FF3 and Opera are attempting to offer 95-98% accuracy for CSS2.1 and 30%+ CSS3 while IE8 intends to offer 100% CSS2.1 and only 10% CSS3.  I don't know if that will happen, but it seems while others are offering a wider variety of CSS features IE8 wishes to become the beacon by which all things CSS2.1 specific are measured.

  • Anonymous
    August 20, 2008
    @austin cheney: Actually, Mozilla and Opera also aim for 100% CSS 2.1 and every release improves this. For example, Mozilla's CSS 2.1 support currently only lacks the display:run-in property and some @page rules. I guess they'll finish CSS 2.1 soon. But yeah, Microsoft promised 100% CSS 2.1 support for IE8 and looking how far they already are, I don't doubt they can make it. After all, MS is way bigger than any other browser vendor. And the released testcases indicate some details they've looked into.

  • Anonymous
    August 21, 2008
    austin cheney: Firefox 2 was actually released just after IE7, but maybe you were confused with alpha/beta versions.

  • Anonymous
    August 21, 2008
    Oy, there... That Microsoft submitted a bunch of test cases is good. Great, even. Problem is, they need review, as Gérard pointed out:

  • use of XHTML 1.1 with text/html MIMEtype (explicitly set as 'incorrect' by W3C)
  • faulty HTML markup may cause varying results in different browsers
  • the cases themselves could have been created in a way that is easier to use. Gérard has kept his database up to date for years now, and has advertised it on this blog ever since IE7 beta 1 came out. The fact that the IE team doesn't make use of it his their fault. In fact, that's where tests like Acid1/2/3 are interesting: they hit where it hurts, where usage has shown where browsers not only contradicted each others, but also went against the standard - and in some cases, where the standard was incoherent. About the 'canvas' patent: true, it's not fully resolved. FUD-provoking though, I don't think so:
  • CANVAS is supported in Safari
  • Safari uses Webkit
  • Webkit is under the Lesser GPL 2.1, which includes a provision about patents. In short, AT THE VERY LEAST LGPL-compatible softwares may implement canvas without fear of running afoul of the Adobe patent; as such, whoever develops an LGPL (or, as is the case, MPL) plugin for canvas support in IE has it well covered. Since the LGPL is a bit more lax with software distribution in binary form (if an unmodified library is the one being distributed, no source is required, only a copyright notice) and doesn't enforce license propagation as much, it would be possible for Microsoft to build a canvas plugin based off Webkit's code for canvas, distribute it with IE as a compiled library, as long as that plugin sources are available somewhere they'd be both covered by the Adobe patent and free to distribute the plugin. @Stifu: Firefox 2.0 was released after IE7, true; but Firefox 1.5 had pretty much the same engine and several of Fx2.0's features were already present. @Daniel: supporting CSS features is one thing; supporting them correctly is another. For example, IE7 didn't add THAT much over IE6 (selectors, mainly); but what it prominently brought was bugfixes. Let's just hope IE 8 won't need IE 9 to correct as many bugs as IE 7 did to 6...
  • Anonymous
    August 21, 2008
    The comment has been removed

  • Anonymous
    August 21, 2008
    WEB STANDARDS should be your number one priority here. Stop worrying about adding all these new features without first developing the core function a browser, which is to display a site properly based on the standards everyone agrees upon. You have certainly made progress in this field, but not nearly enough and, frankly, I'm tired of it. It is no longer a joke to say "oh, that's just IE for you." You guys have to step up.

  • Anonymous
    August 21, 2008
    @Stifu: You're right, I though I don't have to mention it speparately. @Xavier: Even if Beta 2 will be declared "layout complete" bugfixes are still possible. IE7's layout complete Beta was way more buggy than the RTM version. So I think when B2 is release we should test as much as possible to ensure it'll really have a great CSS 2.1 support!

  • Anonymous
    August 21, 2008
    @Daniel Hoping for 100% CSS2.1 is not the same as making it a mandatory development priority.  FF and Opera have always striven to achieve superior and wider CSS implementation, but that does not directly infer a decision to apply 100% CSS2.1 rendering to the defined standards.  This is the difference I must not have made clear previously.

  • Anonymous
    August 21, 2008
    @austin cheney: Yeah, that's the difference, I see. However, the others aren't just hoping. 100% CSS 2.1 is great for them as well^^

  • Anonymous
    August 21, 2008
    Well, I'm looking forward to this Beta! Having to work daily with IE, it is sure good to know standards supports has been greatly improved. I sure hope more people could juste let IE6 go away and switch to another browser.

  • Anonymous
    August 21, 2008
    all the people say ie 8 beta 1 will be updated automaticlly are wrong because when i tested ie7 on xp the made us remove the beta & run win update to get the RTM build

  • Anonymous
    August 21, 2008
    as many people know ms alot of times now delays releasing things to thier downloads server on time

  • Anonymous
    August 21, 2008
    @8675309 T'was two years ago, probably they've learned a bit how to do a good release? ;)

  • Anonymous
    August 21, 2008
    what about xp/vista SP's beta's

  • Anonymous
    August 21, 2008
    will ie8 beta 2 be released on or before august 28th, that was promised by microsoft???

  • Anonymous
    August 21, 2008
    do microsoft know that there is already a newer version of css going to be released? i am not sure. are firefox 3 and safari3.1 in the future or is the internet explorer 8 in the past?

  • Anonymous
    August 21, 2008
    travis: MS promised "sometime in August."   meggs: the CSS 3.0 proposal isn't done yet, and no one has implemented it fully, as that's not even possible yet.  As Daniel observes above, FF doesn't even have full CSS2.1 yet.

  • Anonymous
    August 21, 2008
    @meggs. I wouldn´t say that IE 8 is in the past, it´s just a different approach and regarding to CSS 2.1 IE is even more than a step ahead, so it looks like.


I can´t wait for the new beta release as well, since I have to get familiar with it. But since it´s still beta there will be some scratching the surface, only. It comes to a showdown and when IE 8 alpha (I am consciously not saying stable;)) will be released and then will see where FF stands. Anyway, for webprogramming you have to deal with all browsers.

  • Anonymous
    August 21, 2008
    "Microsoft is preparing to make Internet Explorer 8 Beta 2 available for download by August 28." http://www.ditii.com/2008/08/20/ie8-beta-2-by-august-28-and-ie8-final-in-nov-08/

  • Anonymous
    August 22, 2008
    maybe a bit off topic, but interresting in testing selectors in IE8: http://www.css3.info/selectors-test/

  • Anonymous
    August 22, 2008
    The comment has been removed

  • Anonymous
    August 22, 2008
    Jeremy W: FYI, Firefox 3.1 dev builds also pass this test perfectly.

  • Anonymous
    August 22, 2008
    PS: IE8 Beta 1 results, without crashing (tested using IETester on Vista) From the 43 selectors 14 have passed, 4 are buggy and 25 are unsupported (Passed 335 out of 578 tests)

  • Anonymous
    August 22, 2008
    Where is IE8 beta2? I want to download it.

  • Anonymous
    August 22, 2008
    Doesn't crash here either, but I wouldn't care to report it. B1 is half a year old, and B2 is almost there. I'll start testing again when B2 is available.

  • Anonymous
    August 22, 2008
    Please people, while it's understandable that you want support for new features like CSS3 and HTML5, just remember that they're not standards yet.  Whether or not Microsoft supports them is their choice until they become standards.  So don't give them grief for it.  Just as Mozilla dn Opera and Apple choose to support some features of draft standards, so too does Microsoft have that right.  

  • Anonymous
    August 22, 2008
    i agree since ms did write in 802.11n draft support for xp/vista & there server versions

  • Anonymous
    August 22, 2008
    while were waiting for them to release v2 of the beta the wheel in the sky keeps on turning!

  • Anonymous
    August 22, 2008
    @Mitch 74 If you attach patent rights license to code it will only apply if the code containing the patented technology is used unchanged. So if some technology patent is transferred using a OSS license on code it does not mean that it releases the patent in general.

  • Anonymous
    August 23, 2008
    @Jeremy W. IE8B1 has just finished css test. Results: From the 43 selectors 14 have passed, 4 are buggy and 25 are unsupported (Passed 335 out of 578 tests) 4 points better than IE7.

  • Anonymous
    August 23, 2008
    pkubaj: I already pointed that out above, and also, 335-329 = 6 in my books, not 4. :p

  • Anonymous
    August 23, 2008
    Jason  & IE Team, This deserves proper recognition.   I have a few small requests:

  1. Proper titles. For example, instead of:- "Chapter 4"
  • the following:- <a title='Grammar and Parsing Rules'>4. Syntax and basic data types</a>
  • would be much easier to understand and would provide a much clearer outline of what is being tested.
  1. Not automated.The CSS tests require clicking through. They're not automated. The only way I can think of automating tests would require an application to launch the browser, take snapshots, and communicate with a server program.
  2. Not enough edge cases. I've noticed that the BODY element's display property doesn't work right in IE8. Tests should test edge cases like these.
  3. Clearer results. I don't mind viewing the source code, but I think the tests could be a lot more obvious as for PASS/FAIL result. If you're creative, you can think of ways to craft tests with more apparent expected results. Finally, I have feedback for the larger picture of things. The CSS2.1 example/demo pages are useful, but only part of the picture. It would be a very good idea for Microsoft to continue in the spirit of these tests, and to make automated tests for Ecma-262r3, HTML 4.01, and w3c DOM (HTML, Events, et c). These are important standards that are partially implemented and buggy in MSIE. It would not be nearly as hard to develop an automated test runner for these standards. @mystere - I agree with you about not rushing to implement untested working drafts. I can see evidence of Microsoft implementing parts of CSSOM views, and parts HTML 5 which are questionable, at best. Rushing to implement poorly designed and changing features can be harmful to the web. Test cases can reveal design issues that the draft may not reveal, and should be an integral part of developing such standards before implementing them. Writing test cases for existing standards would seem to be at least as important. Garrett
  • Anonymous
    August 23, 2008
    @EricLaw: Sure that the most obvious reason to use data: URLs for performance. There are other reasons to use data: URLs than performance, though. Other motivations include: as a fallback when other options are not available (comment attachments, blog post images without separate image hosting support, simple steganographic passing of data out of an oppressive government or corrupt agency to public forums), or as a way of serializing multi-file content into a single file ("Save as a complete HTML file"). I'll have to do some testing to see what happens with multipart compressed data and some other corner cases. Thanks for the link!

  • Anonymous
    August 23, 2008
    One big question that I would love to be answered by the IE team - Why now? FF, Opera, and Safari have been supporting web standards for years and IE is just now taking them seriously? Microsoft is even a member of the W3C and has not cared about the standards they put out until now. Someone, I forgot who, in the IE team was even quoted after being asked, "what makes IE a great browser" as answering "we make the standards." So if IE was so set on making their own web standards, why the change?

  • Anonymous
    August 23, 2008
    So is beta 2 gonna be really good enough to hold one off until the full comes out? My IE7 really acts up and I am waiting fo r IE8. But I want to know if I should wait for the full to come out.

  • Anonymous
    August 23, 2008
    one more week. five more working days. can microsoft deliver?

  • Anonymous
    August 23, 2008
    Does anyone here exprience Google problem? I have is shopping cart on google search result.

  • Anonymous
    August 24, 2008
    From my last comment: Note that I would really love to see IE finally be css compliant as I am a web developer. I would love to just tell people, "update to the newest version" instead of "we can't do much about that" or thinking to myself, "I really don't want to waste time recoding that in javascript (e.g. the :focus css pseudo class that IE7 doesn't support). And if I do, what is the point in keeping :focus in my styles?" So I would love to see IE8 bring itself into full compliance...I just want to know why now? I just wanted to make that clear that I am not against Microsoft and in fact love their OS's, Visual Studio.net, asp.net, vb.net (seriously one of my favorite programming languages), and other technologies and software they have made.

  • Anonymous
    August 24, 2008
    @Gérard Talbot & @ Mitch 74 I will answer some of your questions.  I’ve also asked Arron Eicholz to answer some as well as he’s more familiar with the details of specific cases.  Arron is Microsoft’s Junior Peer on the CSS 2.1 working group (http://csswg.inkedblade.net/test/css2.1/review).  I also have some questions for you. > 1- Why choose XHMTL 1.1? Why was that necessary? >- use of XHTML 1.1 with text/html MIMEtype (explicitly set as 'incorrect' by W3C) See this statement, “The baseline format for CSS2.1 tests is XHTML 1.1 in UTF-8.”  This is from the test suite test page requirement for submission to the W3C.  Here is the link:  http://csswg.inkedblade.net/test/css2.1/format >- faulty HTML markup may cause varying results in different browsers Yes, serving any pages as application/xhtml+xml may break some applications, which is why we used text/html. >- the cases themselves could have been created in a way that is easier to use The cases were written to be browser independent and something that can be automated with a test harness, like the one the working group is designing.   >2- Why so many javascript functions and handlers in such CSS test suite? … All of the other known CSS test suites do not, did not rely on javascript support. There are no jscript functions in the test suite itself only the harness used to drive them.  This is consistent with other test harnesses, including the one the Working group is considering.  Again, we’re trying to align as close as possible to the Working group’s guidelines here. >var arrDocs is a single javascript array variable that is 167504 bytes long in size: does that make sense?? It is simple and straightforward but not necessarily elegant.  We erred on the side of simplicity vs. elegance in some cases. > Here's an nth good example, …  all of the intra-testsuite navigability would have been elegantly solved without javascript. These tests needed to work with as many user agents as possible, not just those that supported specific features. > The testsuite, despite its number of tests, is not even exhaustive either… Correct.  Every test suite can always have more tests.  There is a balance between complete and necessary in every good test suite. I believe the CSS 2.1 suite needs more tests as well.   > Visit my webpage and see tests #37, #38… What’s the URL? > To navigate to the previous or next test, the webpage uses a submit button… Where are you looking? > When the document contains inline script (event handler) attributes… It might be a bug.  At which page are you looking?

  • Anonymous
    August 24, 2008
    @Jason Upton: thank you for your answer. Since the test cases must be submitted as XHTML 1.1, that also means that the files can't bear the suffix .htm (nor .html for that matter) - they should at least be named .xhtml, and let the server deide on the mime type required (strictly speaking, providing an XHTML file also means enable application/xhtml+xml mime type; since most servers decide on the type to send from a file's suffix, get ready to do a mass rename). Gérard's URL is in his comment: click his name. Adding Jscript to the page to improve compatibility is superfluous, as a browser should be able to pass without any crutch (for example, adding document.createElement("abbr"); doesn't make IE 6 really support 'abbr'...); if you really want to provide test cases, those should be:

  • valid XHTML 1.1
  • semantically correct. However, I guess the test cited by Gérard refers to point 6.4.4 in the CSS 2.1 specification - in which case the test is basically correct. It is also true that the W3C doesn't ask for a table to contain anything - a caption could be the only child.
  • Anonymous
    August 25, 2008
    Jason Upton, I hope you're not the best MS can offer, as we're all doomed if you can't find Gerard's website, given it's linked to where his name is displayed and should be required reading for you.

  • Anonymous
    August 25, 2008
    Jason Upton, I hope you're not the best MS can offer, as we're all doomed if you can't find Gerard's website, given it's linked to where his name is displayed and should be required reading for you. http://blogs.msdn.com/ie/archive/2008/08/19/more-tests-submitted-to-the-w3c-css-2-1-test-suite.aspx#8882089

  • Anonymous
    August 25, 2008
    I applaud the pace as which the IE team is playing catch-up. I'm simply hoping that with the massive amount of resources that MS has at their disposal, that there will come a time [again] when IE leads the pack in terms of implementation of specs, and is no longer the fat kid in gym class we all have to wait for to catch up.

  • Anonymous
    August 25, 2008
    what's microsofts answer to tracemonkey

  • Anonymous
    August 25, 2008
    Re: Test Suites. Anyone that has ever written one should know that automation is key. The original set by Jason Upton I presume is just a draft set.  It would have been better to build the set with the automation built in, but let bygones be bygones. Since the test suite would suggest that Beta 2 has better support than Beta 1, can you indicate anything about Beta 2 in terms of what we can/should expect? Thanks.

  • Anonymous
    August 25, 2008
    The comment has been removed

  • Anonymous
    August 25, 2008
    @Jason Upton 1- > "The baseline format for CSS2.1 tests is XHTML 1.1 in UTF-8."  This is from the test suite test page requirement for submission to the W3C. The CSS 2.1 Conformance Test Suite indicates XHTML 1.1    XHTML 1.1 tests sent as application/xhtml+xml www.w3.org/Style/CSS/Test/CSS2.1/current/ If the Microsoft tests are adopted, their XHTML version will be served as application/xhtml+xml while an HTML 4 version of such tests will be served as text/html. As it is, the current 3221 tests are not served as application/xhtml+xml. We were right regarding such concern and in pointing out such issue. 2- samples.msdn.microsoft.com/csstestpages/Chapter_6/Rules/HTML-attribute-001.htm should be fixed, corrected. Just changing <caption> to <tr><td> should be sufficient. 3- >>- faulty HTML markup may cause varying results in different browsers > Yes, serving any pages as application/xhtml+xml may break some applications, which is why we used text/html. Why not fix faulty HTML markup?? 4- > These tests needed to work with as many user agents as possible, not just those that supported specific features. The CSS tests themselves should not depend on the support of specific browser features. Agreed. As far as I know, only Safari (among non-IE browsers) can not make use of a <link>-ed related Site Navigation toolbar. Anyway, whether a few, most or all browsers can make use of a <link>-based/related Site Navigation toolbar should not refrain an author from defining intra-site <link>s for navigation purposes. In a small|avg ordered set of inter-related webpages, <link>-ing webpages is generally useful, helpful, significantly relevant. w3.org/QA/Tips/use-links Microsoft could implement a Site Navigation toolbar (turn off, hidden by default or as an available add-on). I made such suggestion 3 years ago and I was not the only one proposing such feature. 5- > What's the URL? Where are you looking? At the test harness webpage. @ Mitch 74 > the W3C doesn't ask for a table to contain anything - a caption could be the only child. A table must include at least 1 <tr> ... which must include at least 1 <td>: <!ELEMENT TABLE - -     (CAPTION?, (COL*|COLGROUP*), THEAD?, TFOOT?, TBODY+)> <!ELEMENT TBODY    O O (TR)+           -- table body --> etc Regards, Gérard

  • Anonymous
    August 25, 2008
    @Garrett > to make automated tests for Ecma-262r3, HTML 4.01, and w3c DOM (HTML, Events, et c) Automated tests for HTML 4.01 is probably very difficult to do. Anyway, 3 tests sites on HTML 4.01 are already available: Ian Hixie HTML 4 tests hixie.ch/tests/adhoc/html/ Robin Lionheart HTML 4 Conformance tests robinlionheart.com/stds/html4/results Draft HTML 4 test suite w3.org/MarkUp/Test/HTML401/current/tests/index.html Automated tests for DOM 2 HTML are already available at w3.org/DOM/Test/#releases And acid3 test has several tests on DOM HTML. Regards, Gérard

  • Anonymous
    August 25, 2008
    The comment has been removed

  • Anonymous
    August 26, 2008
    new features and submitting new tests are great, how about passing some first?  Acid3 status?

  • Anonymous
    August 26, 2008
    @standardsguy: They pass Acid1 and 2 because they've worked on CSS 2.1 support, as said, they plan to support it 100%. Acid3 concentrates on JavaScript and the DOM, an area that the IE team only slightly improved in IE8. The big overhaul of this section is expected for IE 9.