Test drive 104 new Professional-grade tests for web standards
The Windows Internet Explorer Platform Preview build marks an important milestone in the development of the next version of Internet Explorer. A significant part of the platform preview is focused on new or proposed World Wide Web Consortium (W3C) web standards.
The goal of industry standards is actually “interoperability.” For HD television standards, that means that multiple TVs can render the same rich content in a consistent way. In the web case, that means the same HTML, script, and formatting markup work the same across different browsers. Eliminating the need for different code paths for different browsers benefits everyone, and creates more opportunity for developers to innovate on great web content.
Developers have said they want to use the same markup across all modern browsers. Each proposed standard needs a comprehensive test suite to make this happen. The test suite helps resolve ambiguities in the specification by testing any implementation (aka browser) built on that specification.
The test suites also need to thoroughly exercise everything in the specification and not just test a handful of things sampled from various parts of the spec. During the IE8 project we submitted 7201 test cases to the W3C’s CSS 2.1 Working Group in an effort to help the industry develop a comprehensive test of the CSS 2.1 specification. It was exciting to see other companies join in and provide additional cases as well. As a result, the count of CSS 2.1 test cases grew to 8777 as it moved from pre-alpha to Alpha on January 27th, 2010. This will help every browser vendor build a more consistently behaving browser on which web developers can finally start to depend.
As we look through the test suites for other web standards working groups, we’re finding areas in which there is a need for more thorough testing. As a result, we’re expanding the working groups into which Microsoft is submitting test cases. We also listened to feedback about the volume of test cases and the frequency in which we submit them to the working groups. In conjunction with our first preview release, we’re submitting 104 test cases across six different proposed or recommended web standards. These include:
| Specification | Number of new cases submitted by Microsoft |
Test Suite |
Test Case Feedback Link |
31 |
|||
33 |
CSS 3 Backgrounds and Borders Test Suite (no suite yet) |
||
16 |
|||
9 |
|||
10 |
|||
5 |
|||
104 |
Until these all make it through the W3C test publishing process, you can find the cases on the Internet Explorer Testing Center.
I want to thank the members in the W3C working groups that helped us develop these tests, the community members for providing valuable feedback on the tests, as well as the engineers on the IE team that made these possible.
Jason Upton
Test Manager, Internet Explorer
Comments
Anonymous
January 01, 2003
Your article was very good, I like to visit you want to emptyAnonymous
January 01, 2003
@sane Just ignore the Firefox/Opera/Apple fanboys. Nobody cares about them.Anonymous
January 01, 2003
Just ignore @Hans Nobody cares about him - MSFT fanboys are worse than those that try to push openness and standards.Anonymous
January 01, 2003
Ok well done microsoft for bringing a browser that appears a lot closer to the rest out there. But please ease up on the Bl st, save it for marketing. The people who read this blog are developers interested in helping make the new IE a good browser, do not insult their intelligence.Anonymous
March 17, 2010
The comment has been removedAnonymous
March 17, 2010
Microsoft's IE9 standards tests vs. reality http://my.opera.com/haavard/blog/2010/03/17/microsoft-svg-tableAnonymous
March 17, 2010
@Jason Upton [MSFT] 2 spelling mistakes in DOM 3 events tests: The link "Event object property: preventDefault" should be "Event object attribute: defaultPrevented" because this is how the attribute being tested is spelled: www.w3.org/TR/DOM-Level-3-Events/#events-event-type-defaultPrevented 2- It's stopPropagation, not stopPropogation See "Event object method: stopPropogation" regards, Gérard TalbotAnonymous
March 17, 2010
The comment has been removedAnonymous
March 17, 2010
The comment has been removedAnonymous
March 17, 2010
www.codedread.com/svg-support.php. Enuf said.Anonymous
March 17, 2010
@read: If you read the Opera post, you'll see that they don't disagree with the fact that they fail the tests. Microsoft clearly states what they're testing and what the results are. @SamIAm: as IE8 doesn't try to support these features, i'd assume that ie8 fails them all.Anonymous
March 17, 2010
I think this is fantastic work. The work work you did with CSS 2.1 was fantastic and I think SVG sorely needs something simmilar. I think it's a little premature to be comparing just your tests which you just published to other browsers, let's have a more holistic comparison please!Anonymous
March 17, 2010
The comment has been removedAnonymous
March 17, 2010
Thanks. Keep up the good work :)Anonymous
March 17, 2010
The comment has been removedAnonymous
March 17, 2010
The comment has been removedAnonymous
March 17, 2010
There are several possible other issues with some of these tests. I came across this list http://underdark.nl/dump/ie9-debunk/fieldset.htm It seems mostly interpretation (as usual W3C standards have room for some interpretation)Anonymous
March 17, 2010
The comment has been removedAnonymous
March 17, 2010
Don't forget about text-shadow property... Don't forget about memory-eat by IE... Don't forget about fix fade bugs of effects using jquery... (PNG IMAGES) well i think is that, good luck.Anonymous
March 18, 2010
@Gérard Talbot This is great feedback. Can you please start the case-by-case feedback as threads at the appropriate W3C mailing lists though? That's why I provided links. Having debates among the membership in the official forum is the right place to discuss any syntax issues or test assertions. Thanks!Anonymous
March 18, 2010
@sane > What the page DOES reveal is the current state of IE9 versus the other browsers, in the context of those tests. Personally, I think that
- providing such a small batch of testcases (104)
- spread over, scattered in 6 different specs
- with 3 of them still not CR, "proposed web standards" dixit Jason Upton definitely does not reveal the current state of IE9 versus other browsers. Even if you absolutely want to consider only those 104 testcases, there are still problems... Both Mozilla (Firefox, Seamonkey, Camino, etc) browsers and Safari 4 browser support border-radius (and have been supporting border-radius for quite some time) but with the vendor-prefix version. So why list all the red cell "Fail", then count the score and then add a small asterisk (*) saying that they do not support the non-prefixed version? 3 testcases in "DOM Level 2 Style" are wrong. Period. Any testcase which does not declare a strict DTD and has validation markup errors is not "Professional-grade". " The goal of industry standards is actually “interoperability.” (...) In the web case, that means the same HTML, script, and formatting markup work the same across different browsers. " That's only true for valid markup code, valid CSS code and ECMAScript and DOM compliant code. As soon as the markup is not valid, then there is no reliability, no certainty on the rendered layout. " So many people lose sight of the fact that Microsoft is made up of developers just like us. " http://validator.w3.org/check?uri=http%3A%2F%2Fblogs.msdn.com%2Fie%2Fdefault.aspx&charset=%28detect+automatically%29&doctype=Inline&group=0 reports 1086 markup errors http://validator.w3.org/check?uri=http%3A%2F%2Fblogs.msdn.com%2Fie%2Farchive%2F2010%2F03%2F17%2Ftest-drive-104-new-professional-grade-tests-for-web-standards.aspx&charset=%28detect+automatically%29&doctype=Inline&group=0 reports 331 errors Microsoft should start practising what it preaches in those IE blog posts and in all other Microsoft-controlled websites. regards, Gérard
Anonymous
March 18, 2010
The comment has been removedAnonymous
March 18, 2010
The DOM 3 event support in IE9 is awesome. I hope you guys keep mouseenter/mouseleave in IE even if those events get axed from the DOM 3 candidate rec.Anonymous
March 18, 2010
@ Gerard If you're trying to help, why not consolidate your information and send it? If you're not trying to help, then why are you here again?Anonymous
March 18, 2010
Just wanted to let Microsoft know that not supporting Windows XP is a big mistake they are making. XP is a legendary OS, one that's seen the most success ever than any Windows operating system will see and is as of 2010 as old as it might be, the world's most popular operating system. It's shameful really that the OS maker is the first to drop support for its most successful Windows product when there are millions using it. You can't ignore the truth that XP as of 2010 is your most used and popular OS and one that is going to remain the most popular until at least 2012-13. You might as well support one last revision of IE which has critcal improvements to web standards rendering on your most popular OS ever. I swear as an XP user, even if I do upgrade my OS ever to Windows 7, I will not use IE.Anonymous
March 18, 2010
Everyone - you can spread the word. The unofficial public tracker for Internet Explorer is now up and running at GoogleCode. http://code.google.com/p/openbrowserissuetracker/issues/list I am posting this comment twice - one for this post and one for the newest post. Note that if it does not receive the proper attention after a certain amount of time, I will cease to maintain it.Anonymous
March 18, 2010
Hi Jason, are you planning to move to XP also?? as a web developer, I would love to test my pages in IE9, but it-s impossible for me to upgrade to vista/7. Thanks MarceloAnonymous
March 18, 2010
This IE8 CSS issue is still present in IE9: https://connect.microsoft.com/IE/feedback/details/408759/overflow-scroll-causes-element-to-always-have-its-max-height Same for this one: https://connect.microsoft.com/IE/feedback/details/442805/visibility-visible-applied-to-inline-element-doesn-t-override-inherited-visibility-hidden-value-applied-to-it-s-block-level-parent#details