Standards Documentation for IE7 and IE8

Over the last month, as part of Microsoft’s commitment to interoperability, we’ve published information for Internet Explorer 7 and Internet Explorer 8 describing variations from certain web standards. Today we published another set of documentation which includes information about how IE extends these web standards:

  • [MS-CSS21E]: Internet Explorer Extensions to Cascading Style Sheets (CSS) Level 2 and DOM Level 2 Style Specifications
  • [MS-DOM2CE]: Internet Explorer Extensions to the Document Object Model (DOM) Level 2 Core Specification
  • [MS-DOM2CEX]: Microsoft XML Extensions to the Document Object Model (DOM) Level 2 Core Specification
  • [MS-ES3]: Internet Explorer ECMA-262 ECMAScript Language Specification Standards Support Document
  • [MS-ES3EX]: Microsoft JScript Extensions to the ECMAScript Language Specification Third Edition
  • [MS-HTML401E]: Internet Explorer Extensions to HTML 4.01 and DOM Level 2 HTML Specifications
  • [MS-ISO10646]: Microsoft Universal Multiple-Octet Coded Character Set (UCS) Standards Support Document
  • [MS-ISO8859]: Microsoft 8-bit Single-byte Coded Graphic Character Sets Standards Support Document
  • [MS-XMLSTYL]: Microsoft XML Associating Style Sheets with XML Standards Support Document

In addition to publishing these new documents we also refreshed the previous content based on your feedback. We have received feedback both through the blog comments here and privately and I’m very grateful for the time spent to so thoroughly review the information. Today’s release completes the documentation for the relevant final-approved web standards from W3C, ECMA, and ISO and I’d like to encourage you to send me feedback to help us make further improvements. You can leave comments here or post questions or comments to the Documentation on Standards User Forum on MSDN.

While most of the interoperability information we have provided relates to final approved web standards (for example a Recommendation from the W3C), CSS 2.1 is the exception. We have discussed in the past how important we believe interoperable CSS 2.1 is to the web development community. Having a comprehensive test suite available from the W3C is one aspect of ensuring that interoperability and publishing documentation about where product behaviour varies is another. Because CSS 2.1 is still a W3C Candidate Recommendation we will update our documentation should the specification change and as the test cases in the test suite become officially approved by the working group. We will do this until the standard reaches final approved Recommendation status.

Adrian Bateman
Program Manager

Comments (46)
  1. Anonymous says:

    Posting 11 minutes after I did, not having a link to your own site, not posting a last name, and stating no purpose; you are a troll. Keep sitting here hitting F5 while eveyone else gets real work done.

  2. Anonymous says:

    Visiting the IEBlog to whine about the MSNBot is considered "wasting time."

    Suggesting that IE8 somehow "messed everything up" is considered "trolling."

    Thanks for playing.

  3. Anonymous says:


    I said MSNBot and Googlebot do not make the *declaration* which has nothing to do with actual support for XHTML.

    Do your comments have any purpose or are you just here to troll and waste time?

  4. tim says:

    This one is my favorite! String.prototype.blink( )

    Return the result of WrapWithHTML(this value, "BLINK").

    IIRC, IE doesn’t honor the old Netscape blink tag. 😉

  5. billybob says:

    "Microsoft has patents that may cover your implementations of the technologies described in the Open Specifications."

    Can someone from Microsoft confirm which, if any patents are covered by the extensions mentioned in this blog?

  6. ie9user says:

    Please bring back InLine AutoComplete in IE9. Shift+Entier is not the same as Inline AutoComplete. Do requests like this even get read once by MS let alone taken notice of? Safari, Firefox, Chrome all have InLine AutoComplete today. I don’t understand why it can’t be added when it is an option. In Firefox, a simple tweak in about:config does it. Eric Law blogs that it was his favorite feature and yet you continue to ignore user requests. Please keep it optional in place of the smart address bar but add it back.

  7. boen_robot says:


    This feature does exist as an option in IE8, right?

    If that’s the case, wait for an IE9 *beta*. This is currently a *preview* for developers, not for end users. Inline AutoComplete is not a developer feature – it’s a UI feature – so it’s not in the preview. It’s still questionable whether it will be in IE9 or not. My guess is it will be there, as an option, off by default, same as in IE8.

  8. EricLaw [MSFT] says:

    @boen_robot: Actually, no, Inline Autocomplete was removed from IE8 to the great lament of some. I wrote about this here:

    @ie9user: As always, thanks for the suggestion.

  9. obvious failure says:

    the website icons were also removed from the location bar in IE8 and so now it isn’t half as useful as it was in IE7 or other browsers like Chrome and Firefox. is this fix coming in IE9 for sure? it was most sad to not see it fixed in final IE8.

  10. CvP says:

    Hi EricLaw.

    Please make sure IE9 supports open-search properly.

    IE8 has some kinda problem with it.

  11. EricLaw [MSFT] says:

    @CvP: If you have specific details, I’d love to hear them. Thanks.

  12. WinWear says:

    Hey, great documentation for web developers.

  13. ao says:

    Great work guys on IE9 preview!

    But have a look at Hopkins collected a bunch of bugs on IE8, where most of them aren’t fixed in IE9 yet.

  14. Adrian Bateman [MSFT] says:

    @billybob: In general, we make licensing information for Open Specifications available within 45 days of publication. We will provide this information (including a patent map) for the extensions described in the IE documents no later than May 10.

  15. Jeff says:

    I’m very encouraged to see Microsoft’s commitment interoperability on the web with Internet Explorer. Extending this slightly further, does this mean IE9 will come with out of the box support for open video and open audio formats (such as Theora and Vorbis) in HTML5 to make it trivial for web users to make full use of major sites such Wikipedia?

  16. Matt says:

    Jeff, no, they’re not going to implement "open" formats with known IP problems that would get them sued. As the Eolas case proved, trolls will sue Microsoft first.

  17. Greg says:

    Please address the lack of standardization compatability levels in W3C html/css.

    One cannot find out what a web browser included in a phone can support without actually using the phone.

    This is already done with great success for the Mpeg4 video standard.  

    Manuracturers can only benefit from standards that can be easily displayed to the user in the advertising/packaging/specifications of a phone or other internet ready device.


     level 1: basic html tags, tables and non-animated images

     level 2: adds basic css, animated images, basic javascript

     level 3: advanced html tags, advanced css

     level 4: advanced javascript

     level 5: plugins (silverlight, flash, acrobat, shockwave, etc).

    A longer term benefit of this will include easily obsoleting outdated html/web tags.  This hasn’t been done and is a primary contributor to the messy current standard.

  18. XP says:

    Why do you keep releasing IE app compat VHDs now? You have caused the most pain by deciding to not release IE9 for XP.

  19. hAl says:


    You said: "But have a look at"

    That buglist contains a text decoration bug claim (bug 48) which is in fact incorrect.

    The underlining of images in a text-decoration span is a bug in other browser as the CSS 2.1  spec specifically states that text-decoration like underlines only apply to text and not to images. IE does this correct and several other browser fail on this.

    The added testcase:

    is incorrect.

  20. jaycee says:

    I’m very encouraged to see Microsoft’s commitment interoperability on the web with Internet Explorer. Extending this slightly further, does this mean IE9 will come with out of the box support for open video and open audio formats (such as Theora and Vorbis) in HTML5 to make it trivial for web users to make full use of major sites such Wikipedia?

  21. 淘宝点评网 says:

    what is <meta http-equiv=”X-UA-Compatible” content=”IE=EmulateIE7″>??

  22. AlexMadjar says:

    Extending web standards means first implementing existing ones properly.  Will IE successfully pass the acid3 test anytime soon?

  23. mike says:

    I get excited every time I read about IE implementing new standards but then I go to "When can I use…" and have to have a little cry.

    from the bottom of:

    The reality is that as long as we still have to support IE6 (has 4% support of standards) we can’t progress.  IE7 is only slightly better at 11%.

    Worst of all, IE9 (in dev) is currently only getting a 45% which is still about half of what all other browsers are currently shipping!

    We appreciate every effort – but it is so disheartening to have to start/end every developer conversation with:

    Dev: We can do this really awesome, slick thing…

    Mgr: And it will work everywhere?

    Dev: Yes in all browsers… well except in IE of course.

    Ugh! – it is utterly frustrating developing anything that has to work in IE too.

  24. ao says:


    I get excited every time I read about IE implementing new standards but then I go to "When can I use…" and have to have a little cry.

    from the bottom of:


    You’re talking about Candidate Recommendation and  about Working Draft.

    They are no standards (yet)!

  25. hypotheekrentes7 says:


    It is great that you will update your documentation and should the specification change and as the test cases in the test suite become officially approved by the working group.


    <a href= ” ‘>”> best mortgage deals </a>

  26. Richard says:

    The chart ao links to may be a little misleading: it shows IE6 and FF3.0 both as "distant past", whereas in fact IE7’s release coincided with FF2.0’s, while IE8’s release predates FF3.5’s – yet in the chart, FF3.5 is past while IE8 is present/near future.

    Yet there is some sense in this representation, because FF users are much better at upgrading (and Mozilla is much better at facilitating their upgrades and at making those upgrades available to users with older versions of Windows). Developers have a huge numbers of people still using IE6 to deal with, while FF2 is a tiny issue (and of course, FF2’s standards support is better than IE6’s).

  27. @mike and @ao,

    At the top of the page

    you can uncheck "Candidate Recommendation" and "Working Draft" checkboxes and you will get a summary with the following scores:

    Near Future (early 2010)

    IE8.0: 9%  Ffox3.6: 75%  Saf4.*: 72%  Chrome4.0: 72%  Opera10.5: 97%

    Future (late 2010 or later)

    IE9.0: 44%  Ffox3.7: 97%  Sf4.*: 72%  Chrome5.0: 72%  Opera10.*: 97%

    It’s clearly, utterly convincing that IE9 still has not catch up (and is not predicted to catch up) where the other browsers are right now.

    I wish Konqueror 4.4 would have been evaluated in that page.

    regards, Gérard

  28. Robin says:

    First of all, great work guys!

    And i don’t mean to load you with anything more, but i have a bit of a request.

    With everything you’ve already implemented, could you please extend support to one, just one Array "extra"…

    We would love to see a native implementation of


    A lot of our sites depend on it, and for IE, it has to be emulated in an ugly way. Looping thru every array member can be time consuming, esp when it comes to comparing ALL members of a sparse array with large gaps.

    We try to work with our developers to avoid those cases in production code, but oftentimes things like that do slip in, esp with data generated on the fly.

    With all the performance improvements that have been put in place for IE9, lack of array.indexOf should really be looked at, as a source of REAL WORLD slowdowns

  29. @hAl,

    > The underlining of images in a text-decoration span is a bug in other browser as the CSS 2.1  spec specifically states that text-decoration like underlines only apply to text and not to images. (…)

    hAl, did you report this to James Hopkins?

    What should happen wrt ‘text-decoration: underline’ if the image is not rendered (due to internet traffic or miscoded url or user pref setting disableing image download) and the alt text is rendered? Shouldn’t the text-decoration apply to the substituted alt text then? After all, alt text is text…

    regards, Gérard

  30. hamed says:

    html5 audio not working in platform preview , also videos have many problems , many web sites will use this features soon pls fix it

  31. EricLaw [MSFT] says:

    @hamed: As mentioned by DHach in the keynote at MIX, IE9 Platform Preview #1 doesn’t yet support either the HTML5 Video or Audio elements. These elements will be supported in a future platform preview.

  32. hAl says:

    @Gerard Talbot

    I wrote him more than a month ago (even before you) but he did not react.

  33. hAl says:

    When looking at the

    page I see several very selective choices by the page creators.

    For instance on CSS3 I see some new working draft on 3D transformations listed but the long time standing candidate recommendations on CSS for Basic User Interface, Ruby, Namespaces and Marquee not listed.

    Not showing candidate recommendations but listing newish draft specs makes the percentages a joke.

    This preselection on which spec to show already shows deep bias in the selections on that page.

    That page is not really ment to show support on standards but only on some selected parts of the standards that fits the purposes of the page creators.  

  34. ao says:

    @Gerard Talbot

    The page is a joke. You know it and so does the athor, since he removed the "tests".


    WellI also noticed that he’s not very active on his site 🙁

  35. @ao

    > The page is a joke. You know it

    I replied to what you originally said regarding "Candidate Recommendation and  about Working Draft". And I quoted you.

    > and so does the athor, since he removed the "tests".

    There are test pages available related to the %tage of results and they are linked from that page: is one.

    I agree that all of the test pages should be made available, accessible so that anyone/everyone can verify the tests results and the tests themselves.

    What is not a joke and what I know for a fact is that Mozilla, Apple and Opera have been into supporting and implementing HTML5, CSS2.1, SVG, CSS3, application/xhtml+xml, each and all DOM 2 interfaces (DOM2 Core, DOM2 Events, etc), MathML, acid2 test, acid3 test a lot sooner than Microsoft and that their browsers (Firefox 3.6, Opera 10.5, Safari 4.0.5) achieve more support and better (more correct) implementation than IE8. Regarding CSS 2.1, it is true that IE8 has a full support of all CSS 2.1 properties but IE8 does not have an impeccable (correctness) implementation. What is questionable here is that, during the IE8 beta development, I reported clearly defined CSS 2.1 bug reports (with reduced testcases, relevant excerpts of spec, etc) and they were confirmed and… postponed… but when IE8 came out, IE Team (and several MSDN pages) boasted loud and clear that IE8 was fully CSS 2.1 compliant and passing CSS2.1 test suite.

    Even with regards to HTML 4.01, Firefox 3.6 does better at Robin Lionheart’s HTML 4 Conformance tests than IE8.

    I am convinced that Firefox 3.6 is a more UAAG compliant browser than IE8.

    I know that Firefox 3.6, Opera 10.5, Safari 4.0.5, Konqueror 4.4.2 achieve better at DOM 1 Core test suite. And I have listed the scores I got. Anyone can verify for him/herself.

    regards, Gérard

  36. lord_potter says:

    Dear MWIE-team

    Internet Explorer 9, dont behave well with World of Warcraft!

  37. ao says:

    @Gérard Talbot

    Just want to point out that most of the tests are based on standards still in development.

    Also most of the ‘complicated tests’ are not based on real-world scenarios. Most of the times there are easier ways to make things work that work in any browser. And remember "No Web browser fully supports HTML 4.0 yet, though it became an official standard as far back as 18 December 1997."

  38. I’m not sure how to get in direct contact with the folks running MSNbot though I’d really appreciate it if someone from the IE team could pass on a small request on my behalf.

    I’m not sure if MSNbot understands XHTML served as application/xhtml+xml though I can not serve pages with the correct mime unless the string "application/xhtml+xml" is present inside of the HTTP ACCEPT header. The problem is that IE8’s compatibility mode only serves "*/*" and well, we all know that IE9 supports the mime now (WOOHOO!) though IE8 does not, we can’t trust the user agent, and the goal is to serve the page correctly on the initial load. If the MSNbot team could include the mime in the HTTP ACCEPT header it would allow XHTML websites to serve their pages the same to bots as they do to browsers.

    I’d be happy to also discuss this with them if this would help or to clarify anything. Any way keep up the awesome work on IE9!

  39. Matt says:

    John: The premise of your question isn’t clear. Why do you think it’s important to send XHTML to the MSNBot?

    Since browsers that don’t understand XHTML outnumber (in prevalence) those that do, you’re closer to accomplishing your goal if you simply serve it HTML unless it specifically asks for XHTML.

  40. @Matt

    1.) XHTML is ten years old, it *SHOULD* be supported by all software now.

    2.) Since IE8’s compatibility mode messed everything up with the pseudo */* declaration XHTML pages can no longer reliably be served unless the string "application/xhtml+xml" is declared.

    3.) My site is XHTML, it is not HTML. Forcing me to serve my content as text/html forces me to cloak the page which means different content to different user agents.

    4.) As a web designer and web developer I can only adapt so much until I’m making crap-shots. I already made it explicitly clear *why* the string must be part of the HTTP ACCEPT header.

    5.) My site should not have to be given permission to be what it is. "Who said you were allowed to be Matt?" I’d have the same reaction to such a question posed to me too.

    Yahoo currently does this correctly, MSNBot and Googlebot do not.

  41. Matt says:


    <<<XHTML is ten years old, it *SHOULD* be supported>>>

    Utterly and completely irrelevant. There are many standards which are decades old that are not broadly supported (or even ever implemented at all).

    <<<IE8’s compatibility mode messed everything up with the pseudo */* declaration>>>

    Uh, no. IE has sent */* forever, and that’s completely legal anyway. Go read what "ACCEPT" really means. A browser is free to send whatever it wants. Any version of Internet Explorer is perfectly willing to download an XHTML file and that works just fine.

    <<<Forcing me to serve my content as text/html >>>

    Which has nothing at all to do with the MSNBot.

    <<<Yahoo currently does this correctly, MSNBot and Googlebot do not.>>>

    Suddenly, you’re certain what MSNBot and GoogleBot support? How’d you go from "Someone please help me understand how search engines work" to "I know how each search engine works and I don’t like it." Rather odd change up in just a few hours.

  42. Mark says:

    @John B. – coding to XHTML before all browsers supported it was your choice.  There was a risk, you took it and you are now dealing with the ramifications of that choice.

    As for the MSN bot accessing your content – seriously who cares?! As long as the Google Bot can see your stuff you are fine.

    If the Google Bot can’t see your stuff, then go chat to Google – no one else seems to be having the same issue as you, and the Google Webmaster Tools will help you build a sitemap file to aid in crawling your site.

    On my site 92% of search traffic comes directly from Google. I’m not wasting a second of my time on 8% if they can’t handle what Google can handle.

  43. Will Peavy says:

    @John – what’s the benefit of serving XHTML? Why not use well-formed HTML?

  44. @Will Because you’re not supposed to cloak (e.g. serve different content or content differently) to different user agents be it browsers or spiders. Server resources shouldn’t have to be used to deal with different user agents either. Few other reasons too though I don’t have endless time to reiterate everything.

    Also I’ve confirmed that Speedy Spider ( and Sogou web spider ( send the HTTP ACCEPT headers in addition to Yahoo’s Slurp which I mentioned earlier.

  45. Will Peavy says:

    @John – but why not just send well-formed HTML to all browsers? What benefit do you get from serving XHTML?

Comments are closed.

Skip to main content