Documenting Standards in IE

This post discusses some of the work we’re doing on the IE team to fulfill our commitment to document our support of web standards. A good starting point is Microsoft’s interoperability principles, something we’ve written about here on this blog before, and a principle that’s easy to see in action in our products, like IE8.

The essence of interoperability in this context is that the same web page markup works the same way across different browsers. There are many challenges in getting to this goal. Even with the best intentions, as an industry we are still learning and working through how to do this well. You can look at how different tests run even today in modern browsers (for example here at 19:57). You can look at how standards evolve, like how quickly CSS2 became CSS 2.1, or the process to finish CSS 2.1 and make it a final Recommendation, or what happened between XHTML and HTML5. You can look at the challenge of delivering interoperable products while specifications are under construction (as in the case of 802.11 wireless). There are many challenges, and the web standards process, primarily at the W3C and similar organizations, is an important means to get the different communities involved to a consensus agreement.

The work in developing a public CSS 2.1 test suite and contributing it to the W3C, our recent work on different aspects of HTML5, and the improvements in IE9’s interoperability we showed at PDC are all examples of our principles in action. You can try out some of the tests yourself, in different browsers and on different operating systems.

As part of our commitment to interoperability, we’re going to make more interoperability information available about IE and keep it up-to-date. Today we’re publishing the first pieces of documentation here. These documents are drafts, and are not final. We will post more here on the IE blog about interoperability documentation (e.g. how we engineered creating this documentation, the process for keeping the documentation up to date).

Thanks –
Dean Hachamovitch

Comments (94)

  1. Anonymous says:

    Thanks for these articles, I enjoyed them!

  2. Anonymous says:

    Thanks for sharing. eine Seite ist entworfen worden, very nice Glückwünsche;)

  3. cazaresulina says:

    Very interesting

    I will wait for iexplorer 10

  4. Anonymous says:

    @Mark – Rebecca’s logic makes perfect sense.  Just apply a little thought, she’s correct.

    @AntiTroll – you’re sound like your antithesis.  Todd was expressing his opinion, I believe something he is allowed to do in a Democracy.  You however didn’t assert your opinion, you make a smart comment, hence you’re an idiot.  Whether Todd’s right or not about IE 9, IE is losing share generally and I believe within the next 18 months – 2 years will be below 50% as Safari, Opera and Firefox continue to improve (especially Opera, it rocks the desktop right now).

  5. Tom says:

    Will you guys have a post with on the changes made in the json update made avaible on 2-23-10 or could you use some pull and talk to whoever posts on the jscript blog to write up a post on the changes made?

  6. Ivar Vasara says:

    Could you please publicly commit to supporting SVG ? This post contains a link to the SVG working group update, and Microsoft will be presenting SVG at MIX, but there has been no official statement of upcoming browser support for the standard.

  7. mogden says:

    Every web developer knows that the open source browser projects do a far better job than IE with interoperability.  I don’t really like the attitude that seems to be expressed in this post that "standards are hard".  That’s probably true, but why does the leading market share browser do the worst job?  Surely it is because of inadequate focus and resources applied to the problem, relative to the amount of pain that it causes for everyone developing web sites.

  8. Huh? says:

    "Every web developer knows that the open source browser projects do a far better job than IE with interoperability" — data please. Not about the 2001 browser either, unless your data is about interop with other 2001 browsers.

    IE8’s implementation of CSS 2.1 has thousands of tests behind it. Not sure where this post says "hard" — it is realistic about what’s involved in it.

  9. blankblank says:

    >>>>"standards are hard"

    Great lets fire the IE team and get new people in, cause they can’t get the job done!

  10. @mogden, IE 4/5/6 blew Netscape 4 away back in 2001 and at some point Netscape went from closed to open source.

    PDF? Are you guys serious? You’re trying to make the web faster and then force me to wait  for Adobe’s can’t-load-in-under-thirty-seconds-without-an-SSD-and-thousand-dollar-CPU software? Please stick to (X)HTML.

  11. blah says:

    The rest of the world has already documented the frustrations and pitfalls of ‘standards’ in IE. Forget Acid; people have had so much collective trouble getting the simplest of stylesheets working right, myself included.

  12. EricLaw [MSFT] says:

    @Tom: I’ve passed this feedback to the JScript team and hopefully we’ll be able to publish information about this shortly.

  13. Eric says:

    This is probably some sort of sin on this blog but here goes…

    I would love to see IE use Webkit.  Before I get flamed to death, let me at least back up my intentions here by saying I am a Microsoft fan.  My preferred OS is Windows 7 and it is my absolute favorite OS. I have so few complaints about the entire OS that I’m astounded.  Even all of the Apple fans I know have a lot of respect for it.

    That being said using Webkit would allow Microsoft to focus more on creating native features for its browser to be able to pull ahead in the unending browser war.  The benefit to the web developer community would be _massive_ and would certainly put a squeeze on Fire Fox.  Microsoft shouldn’t be worrying about open standards it’s not going to impress anyone creating yet another rendering engine and IE is notorious for being the slowest to implement new features… *cough CSS3 *cough*…

    I’m stuck worrying about "elegant degradation" day in and day out if I want to be leveraging the best of what is available to me.  Does anyone care to add up the number of wasted hours in this world spent creating style sheet hacks for IE6?

    Prove the modern conception of Microsoft wrong and adopt a third party technology.  Relinquish the control in a web rendering engine in favor of developing OS native features that _nobody_ else can touch and push IE back into a very good position.

    Thanks for reading 🙂

  14. Tom Stack says:

    @EricLaw[msft]- thank you for the update about your json update.

  15. Will Peavy says:

    @Eric – are you familiar with Chrome Frame?

  16. Eric says:


    I am, yes, but that’s merely crutch and doesn’t allow Microsoft to benefit from not having to focus on continuing development of a rendering engine.  Using Webkit would free up those resources and developers to work on something innovative.  As it is they’re not even talking about CSS3 meanwhile the entire world has already started using CSS3 and is having to worry about still making it look good in Internet Explorer.

    Switching to Webkit would win over the hearts of tens of thousands of developers across the planet.  The entire web design / development community would praise Microsoft.  If you want to win a browser war; fulfilling a collective dream is an excellent place to start 🙂

  17. bye bye IE6 says:

    YouTube dropping IE6 on March 13!

    about time!

  18. @Eric I’d love for people to stop suggesting IE use WebKit. As much as I like the fact that WebKit is stuffed full of CSS3 goodness it’s better in the long term that there are many, not fewer, interpretations (rendering engines/browser vendors) of standards. No single rendering engine in example is bug-free on my site which reinforces in a way one of the bits that Dean posted about. If it wasn’t for Microsoft then how would the web be different from today? Consider things such as AJAX and favicons when composing your answer.

  19. Foo says:

    @ Eric — what webkit? there are so many to choose from, with so many differences:

  20. J. King says:

    I can’t agree more with John Bilicki: Why PDF?  Besides the irony of publishing documentation about the Web in a non-Web friendly format, the pagination of the PDFs is such that examples (and far worse, version information) are broken up by soft page breaks and the page footers distract heavily from the actual document content.  

    Personally I would find releasing documentation with such careless pagination totally unacceptable, and given the audience far better to just use a continuous display medium like HTML displayed on screen.  By all means attach a print stylesheet, too, but the current documents are just too damned hard to read.

  21. Eric says:


    It’s easy to post a hypothetical "what if" and say we wouldn’t have technology A… well that’s true but the need was there and a solution was inevitable. It may not have been AJAX but it would be something similar. Let’s not pretend that multiple rendering engines were responsible for that.  Your argument is based on a classic logical fallacy.

    I’m not saying there shouldn’t be additional choices here.  What I’m saying is that Microsoft’s implementation is lagging behind and it’s at a cost that the entire market bares… Microsoft isn’t competing with their rendering engine they’re sandbagging the industry.  Maybe I’m slanted because I have to deal with this mess nearly every day of my life but there’s a reason people are recommending this course of action.

    How much money has IE6 cost this industry as a whole in development time? The cost is MASSIVE and it’s arrogant to shrug that off.  This fact has left a very bad taste in the mouths of web developers everywhere, there’s a lot of discontent towards Microsoft… Why do you think everyone is celebrating the death of IE6?  They’ll do the same with IE8 if something doesn’t change…

  22. marc says:

    I will say it again: please IE if you will replace Trident, please choose Gecko, not Webkit.  Gecko has proven maturity and is totally supportive of web standars.

  23. rick says:

    Eric’s right. There’s no real benefit to the web community from having an IE rendering engine. I don’t actually care if it’s Webkit or if they upgrade the IE engine, but it’s silly that I can do simple things like rounded corners and shading with a line or two of CSS for Chrome/Safari/Firefox/Opera, but because IE fails to support those CSS3 calls I have to tweak images or try to make some Javascript hack work to get that effect in IE, even IE8.

  24. Shinu says:

    @Eric I couldn’t agree more with you. I’ll take a look at your blog, it also looks interesting 😉

  25. Rob says:

    You aren’t "documenting standards"! This is just Microsoft’s bug report! Microsoft is NOT any standards committee so how can you be documenting it?

    By documenting them, does this mean "won’t fix"?

  26. Eric says:

    @Rob: the phrasing in the first sentence is more correct: "to document our support of web standards".

    The IE team only makes changes to standards support in new browser versions.

  27. Ryan says:

    >Why do you think everyone is celebrating the death of IE6?

    Uh, because it’s very nearly 10 years old?

    Firefox 1.0 is only half as old and web developers aren’t happy to target that browser version either.

  28. hito says:

    It’s enough already.

    What we want is a standard conformance implementation.

    Just support the DOM level 2, and 3, and XHTML, SVG, HTML5.

    I can write code that works perfect in any other browser.

  29. Mitch 74 says:

    Why open source browsers do better on interoperability – or any other browser, really.

    There is one, very simple, reason.

    They talk.

    Let’s look at current browsers: all of Mozilla’s (Firefox, Seamonkey, Fennec etc.), Safari, Chrome/Chromium, Konqueror, Opera.

    What engines are in use?

    – Mozilla: all use Gecko.

    – Safari, Chrom(e/ium) use Webkit.

    – Konqueror uses KHTML, which is the basis for Webkit (and backported some stuff from Webkit)

    – Opera has its own.

    If you look at these browsers’ mailing lists, you’ll see that developers for each browser confer with their upstream developers or with each others on quirks in their implementations, and corner cases in the specs; and they will then confer with W3C members (when they are not members themselves). Trial implementations (snapshots, developer builds, private builds) are often shared to test these quirks and various ways to harmonize them.

    Yes, Opera also does that – they don’t provide the source code, only frequent ‘test’ or ‘beta’ builds.

    And that dates back to 2001. In 2003, when KDE started working on KHTML, they did it “like Gecko”. When Apple decided on a rendering engine for their browser, they forked KHTML and built Webkit – and (reluctantly) started discussing with KDE developers.

    What were Acid tests again? Well, they were graphical demonstrations on how browser quirks could kill a page’s layout. Version 2 and 3 hit browsers where it hurt. Please note that these didn’t even try to target IE – it hit other browsers, only because IE was too easy a target…

    During that time, IE developers were… doing nothing. Well, almost: a pop-up blocker was made for IE 6, and IE needed a monthly security fix.

    So, how can they catch up on nine years of backlog in quirks and corner cases programming?

    By doing the same thing others did. By communicating.

    What are we seeing now? Well, starting with IE 7, there was a blog with a post once a month or something, and sometimes a comment. There also was a bug tracker.

    Both were compared with competition, but it was a vast improvement over before – when there was nothing. The results? IE 7 was much less of a pain to program for than IE 6.

    Then, IE 8 entered the pipe. The bug tracker was more or less fixed, and there were enough blog posts and details on how it was developed, and more reaction on comments…

    IE 8 improved steadily between beta 1 and RTW. Now, IE 8 is _relatively_ painless to program for (it’s the first IE release that didn’t require a single CSS fix from me). It still sparked a lot of controversy, that had an outcry for more active W3C participation (cross-domain AJAX, for example, but also the rendering engine choice) – and, instead of MS saying “we’re just gonna do what we want here, you suck it up” once every 6 months (I’m exaggerating, but that’s how it felt at the time), there is now active participation from MS in W3C.

    The results? IE 9 is shaping up to be “back in the game”: better Javascript speed, SVG support (I find myself hoping for DOM 2 events, but I betcha it will still be sorely lacking). Why?

    Because IE developers stopped developing stuff all alone in a vacuum.

  30. TheLudditeDeveloper says:

    @Eric Most of the entire world could not care less about CSS3, most in fact don’t even know what CSS is (invented by Microsoft by the way – and then ‘standardised’ by others).

    Most complaints against IE refer back to a browser standard set in 2001, when it WAS the only game in town.  At the time nobody was making any complaints.

    Only a complete idiot would try to compare todays open source standards HTML5, XHTML, CSS3 and SVG to the features of a browser developed over eight years ago.

    That said, the real frustration is the technology lag.  Developers are being presented with new features nearly every day, but the users (note THE USERS) are are not following as quickly as the developers would like.

    Solution don’t support those users who insist on using old browsers.  Write code only for the latest browsers.  That would mean no more hacks for IE6 or even IE7.

    For those using operating systems like windows 2000 and older and who can’t update their IE6 browser tough luck, move to another browser.

    For those who have applications that can’t use modern browsers, talk to your application developers.

    Even television broadcast have gone digital today.

    I am particularly minded of some managers in the UK NHS who are talking about upgrading from IE6 to IE7.  If they are still stuck in IE6, but are considering an upgrade they should definitely upgrade straight to IE9 when it becomes available.

    The problem today is not old browser versions it is users and IT managers who live in a fast changing world, want all the benefits of change, but just don’t want change.

    Live with it. Anyone still complaining that IE6 does not adhere to standards which came into effect long after IE6 was released is simply not too bright.

  31. Rob says:


    Most of the world doesn’t care about HTML either but most web developers do care about CSS3.

    Since when is HTML5, CSS3, SVG and XHTML an "open source standard"? They are standards we all follow despite what Microsoft tells you to believe.

  32. mike says:

    @John re: "interpretations… of standards."

    That’s exactly the issue! If IE used WebKit we wouldn’t need to worry about IE "Interpreting" the standards incorrectly.

    The whole point of standards is to AGREE on how something should work.  Each vendor can internally do whatever they want, but if the web page says make this with a 5px radius, then it should have a 5px radius. End of story.

    Where IE fell down and lost the respect of developers was when IE totally failed to implement things correctly.

    .getElementById(id) was a ridiculously simple spec yet IE managed to double-bork it.  It wasn’t case sensitive and it matched on attributes that were not the "id" attribute.

    .setAttribute(name,value) was the same disaster. A total fail on the implementation level.

    Even when IE did get things right when they invented their own properties/methods e.g. .innerHTML they still messed it up!  It is an awesome property to get/set the inner HTML but there were 2 major flaws.  1.) It didn’t work on all elements and the ones it failed on were the most desired. 2.) the ‘get’ part was horrible returning the absolute worst version of spaghetti HTML markup you can imagine.

    I seriously hope that IE9 has fixed this and returns clean, exactly as defined HTML.

    As for AJAX and favicons – yes they were a good idea.  The initial implementation for AJAX via ActiveX was a huge mistake (later fixed) and the favicons was good… but auto-downloading when a link is specified elsewhere is an implementation fail… and not supporting PNG favicons is also a major fail.

    PS the CAPTCHA is still a major fail on this blog. I guarantee the number I typed was correct.

  33. John says:

    @mike: If you don’t understand how Webkit also *interprets* standards, then you don’t really understand much about browser development. You should go back and read the posts by the IE team about how they’re engaging with the SVG working group about ambiguities in that spec. The differences in implementation between the "standards compliant" browsers is sorta shocking.

    The same applies to CSS2.1; read the standard, then read the hundreds or thousands of ambiguities that exist in the standard.

    No standard is perfect. That’s why the W3C requires both 2+ interoperable implementations AND a test suite to have a final standard.

  34. mike says:

    @John (ps I was originally pointing to John A B the 3rd)

    I understand there are ambiguous parts and they need discussing – no qualms with that.

    However the black&white issues that leave no imagination to the implementation – they have to be rock solid.  IE8 started shaping up here (only if you run in standards mode), but previous to that IE was a mess.

    However since I haven’t had a chance to test IE9 I’ll save my comments until there is something real to test.

    I do want to submit a bug on the IE dev toolbar though.

    When you are in Script debugging mode (possibly other spots) double clicking a word in the code highlights the entire word (very basic understood and expected text highlighting implementation) however if you double click a word in single quotes (e.g. 99% of your JavaScript string values, it happily highlights and selects the closing single quote.

    Not only is this useless, but it breaks the established behavior. Please fix.

  35. Don McClary says:

    Why was the logic for the security warning to see secure and un-secure content together changed to ‘no’ instead of ‘yes’ as in previous versions? You’re totally cutting against user familiarity and experience to your product.

  36. Adrian Bateman [MSFT] says:

    @John: the documentation we published is available as PDF (for people who wish to download it and read offline or print) and as browsable HTML content in the online library. Occasionally we do make this type of document available first in PDF to get it to you sooner while we do the extra work to format for the MSDN publishing system. All the documentation we published yesterday is available online in HTML as well as PDF.

  37. Lenny says:

    @EricLaw / @Don

    I just read the blog post and re-familiarized myself with the new dialog.

    I "get" why the change.. but there is a new failure added, in that the question prompt is now a double negative.

    "Do you not hate Britney Spears?"

    [Yes] [No]

    It takes everyone (regardless of their stance) twice the time to figure out what the question is *actually* asking where as the following is dead simple.

    "Do you like Britney Spears?"

    [Yes] [No]

    So, in the new HTTP/HTTPS dialog not only is the button choice different (e.g. was yes, and is now no) but the user is even more confused by the question.

    Granted it isn’t an easy question, but the new version is actually worse than the original.

    The question should be something like:

    Title: Security Warning

    Message: This page contains insecure HTTP content.  Do you want to display the insecure content.

    [Yes] [No]              (more info)

    [_] Always allow insecure content from

    The "more info" should be a hyperlink (not a button) indicating that it can be opened and viewed before making a decision.

    The default button should be [No] if you want to enforce the high security level, but the user should be able to "dismiss" the warning for a site they trust. (e.g. if my CRM app includes an insecure Google Map, I’m A-OK with that)

    Don’t not never confuse your users with double or triple negative logic – it always not never fails.

  38. @Adrian Bateman [MSFT] Thank you, I understand how a single self-contained file would be more desirable in certain situations.

    @Eric Consider the context of what I am talking about. In example IE’s CSS implementation (IE4/5/6) at the *time* was vastly superior to Netscape 4’s, any one who consider’s IE6 a nightmare to deal with should spend a month trying to get anything to work with just CSS in Netscape 4. Zip forward to today and yes IE6 is an utter nightmare and even after I cleaned it up for the 29th version of my site I still could not track down "phantom pixels".

    I am a person who prefers to look forward. I love cutting edge stuff provided of course someone can demonstrate why it’s useful. Yes we all know in a lot of ways IE is behind, yes even now IE8’s JavaScript is stuck at 1.3 (1.5 with proprietary implementations), and yes IE8 does not support XHTML. However I still think Microsoft could do a lot of good in the future. I just don’t remember any one in any way complaining about Apple when they went to the W3C with transitions and transforms. Obviously I think every company should have the freedom to bring forth forward thinking ideas and should Microsoft decide to do (let’s say with IE 11) people will jump them just because of past politics. I honestly think in any organization and most especially in the business world there are going to be sharp internal differences of opinion and policy. I think most people care about what software they build and want to see their work succeed.

    Looking at IE7/8/9 I can sort of relate to the development process with my own site. I’ve relied on third party modules (e.g. vBulletin, WordPress) that could be compared to things like VML. Building my own solutions (compared to MS implementing CSS 2.1, SVG, etc) is *not* an easy task. Building my own stuff (e.g. blog, CMS, chat room, forums, private messaging, etc) has taken almost two years however when it’s finally finished it’ll serve as a platform for things to come…and that can be compared to the initial iteration of OS X which didn’t have much to begin with for those who need a non-MS comparison.

    If there were twenty rendering engines really I probably wouldn’t mind the WebKit comments so much. There was a recent study that showed too many subscribers on a video channel would drown out the personalization of that channel. At the same time it’s important to retain the rendering engines we do have (I am actually very impressed with KHTML’s sub-system performance which is only currently matched by Gecko 1.6+ and would prefer to see them not cease development). We lost Yahoo’s search for ten years (thanks for forcing monopolizing on us Carl I.) reducing American’s (major) search engine choice from three to two (most search sites like Lycos simply rehash Bing, Google, and Yahoo results) and I would not like to see in ten years having only a choice between Gecko and WebKit. Consider consumer’s "choice" for PC CPU’s…can you run Windows XP/Vista/7 without an X86 CPU? If so could I purchase the parts from an online retailer and get similar performance?

    Diversity (not the political but the scientific definition) of rendering engines is very important. I use all the developer tools (Web Developer Toolbar by Chris Pederick, Opera’s DragonFly, IE’s Web Developer toolbar, Safari’s developer tools, etc) and find that each will tell me various things the others won’t. In example accessibility is important to me (I love the tab key!) and only Microsoft’s tools trigger errors about visibility and focus. WebKit will alert me to incorrect mime/media type on files (such as a CSS file being interpreted as text/html). So if Microsoft axed Trident and replaced it WebKit I would lose the benefit of ensuring my visitor’s accessibility isn’t broken (Safari took an alarming amount of time to enable accessibility). As more standards are implemented in the future I very much would appreciate having Microsoft included in having their own perspective considered to help balance the standards being implemented in to browsers.

    …and lastly consider how things will be in about ten years. When we’re trying to shake off IE9 with 10% market share someone may look back and consider a different browser with horrid implementation of a similar spec that was released at the same time and think…IE9 wasn’t as bad as X Y or Z. Because as much as I can’t stand IE6 I’ll take dealing with it’s CSS issues any day over trying to get Netscape 4 to work and considering the differences in how IE handles things like XML between say IE7 and IE8 I am pretty confident that IE9 will rock harder then most anticipate it will. This decade may start a little slow because yes, we all miss our desired release dates (e.g. myself and Mozilla with Firefox 3.6 for January 1st as just two examples) however but if you give people sufficient time you’ll might just see things in a different more objective yet also more positive light.

  39. Matt says:

    @Lenny: this is entirely off-topic. Why not comment on the proper post?

    >the question prompt is now a double negative.

    No, it’s not. The prompt is "Do you want to view only securely-delivered content?"  There are zero negatives in that sentence.

    >the user should be able to "dismiss" the warning for a site they trust

    You’ve just demonstrated that you don’t understand the mixed content threat. You should go re-read Eric’s post.

  40. AndyC says:

    @Lenny. Actually a far, far better approach would have been to remove the Yes/No aspect from the equation altogether. Something more like the Vista Task Dialogs that simply offer the choice between two options would have solved the understanding issue.

    Of course, that would have been harder to implement on XP, since it doesn’t have native Task Dialog support, but it surely wouldn’t have been that much harder than the current dialog box.

  41. Will Peavy says:

    I agree with John Bilicki, competition among browsers is good. I’m a developer and I think Trident is fine.

    The main thing I’m hoping for in IE9 is faster JScript. Carakan and V8 are awesome and totally blow JScript away.

  42. Sylvain Galineau [MSFT] says:

    Quick correction for @TheLudditeDeveloper: Microsoft did not invent CSS; that credit belongs to Hakon Wium Lie (from Opera) and Bert Bos (W3C). It is, however, true that IE3 was the first implementation of CSS1 in a major browser.

  43. @ Dean Hachamovitch [MSFT]

    > A good starting point is Microsoft’s interoperability principles

    The first starting point with any Microsoft-controlled webpages, including those talking, discussing about standards, conformance, interop repeated commitments, etc is

    This is true for the first entrance webpage to the last (and most recent) created webpage.

    has 76 errors and does not even use a strict DTD.

    Gérard Talbot

  44. @ Dean Hachamovitch [MSFT]

    > You can look at how different tests

    > run even today in modern browsers

    > (…)

    > The work in developing a public CSS

    > 2.1 test suite and contributing it > to the W3C

    [MS-CSS21]: Internet Explorer Cascading Stylesheets (CSS) 2.1 Conformance Document

    3 Appendix A: Test Suite Failures


    lists 18 tests which fail in IE8 but those were regarding the 2009 pre-alpha test suite. Some of the tests I created and then filed as bugs at connect IE beta feedback (even in some cases, 9 months before IE8 RTW eg bug 354316) which were valid bugs, confirmed bugs are not even mentioned in that document … and they are still unfixed in IE8.

    Today, we know IE8 fails at least 150 testcases from the CSS 2.1 test suite (build 20100127; alpha 1) and I approximate that about 100 (or more?) Microsoft submitted testcases need to be updated, modified, corrected and/or clarified.

    The whole [MS-CSS21].pdf document has not a single issue to report regarding IE8 or IE8’s most/best standards compliant rendering mode. So, was it really required/useful/helpful/relevant to release such (even if preliminary) document now in an IE blog post titled "Documenting Standards in IE"?

    I personally filed a bunch of valid, confirmed HTML 4 bugs at connect IE beta feedback (and a few of them are still unfixed and postponed) and none of them are mentioned in that [MS-HTML401].pdf document.

    Documenting Standards at starts with publishing error free webpages using valid markup code and valid CSS code, especially all those MSDN webpages supposedly explaining/lecturing/tutorializing/examplifying how to author webpages.

    A convincing starting point with regards to Microsoft’s interoperability principles starts with publishing error free webpages using valid markup code and valid CSS code at all times and as a Microsoft policy.

    Gérard Talbot

  45. ie + webkit = win says:

    IE should use Webkit. Even with hundreds of comments similar to this, it is not repeated enough times.

  46. @ John

    > The same applies to CSS2.1; read

    > the standard, then read the

    > hundreds or thousands of

    > ambiguities that exist

    > in the standard.

    > No standard is perfect.

    Your statement is over-excessively exaggerated. There is a list of ambiguities or issues requiring to be untangled, clarified and fixed. Most of those issues have been already fixed and closed. There is not thousands of ambiguities in CSS 2.1. There is not hundreds of ambiguities in CSS 2.1. And they are being discussed, addressed.

    See for yourself:

    regards, Gérard Talbot

  47. AntiTroll says:

    @Gérard: How many web browsers have you implemented exactly? For someone so keen to point out bugs, it’s a shame you don’t devote the same level of attention to standards as you do to implementations.

    <<The whole [MS-CSS21].pdf document has not a single issue to report regarding IE8 or IE8’s most/best standards compliant rendering mode.>>

    You clearly did not read the document. Look for the string "IE8 Mode" in the text. You’re welcome.

  48. Rob Parsons says:


    My deductions from the IE Web Developer forum is that Developers and enthusiasts do not fully appreciate what a web browser is. HTML, XHTML and CSS are not programming languages.. they are not compiled, they are not strictly typed. There is no guarantee that what comes down the wire to the client is free of syntax or logic errors. Scripting languages are interpreted at the client end.

    The expectation is that a browser should be able to handle anything that is thrown at it and that magically a browser should be able to read the developers mind and render pages as they intended. Somehow the browser has to know what is junk, what is a crafted attack, or what is a simple typo.

    This is folly…. we are left with a situation where we are continually chasing our tails tweaking browser rendering engines to cover ALL reported scenarios resulting in bloated client side rendering code. This is in-efficient (from a computing point of view) as each client browser is in effect using resources to detect and correct errors. Feeding junk to the browser has made it a fat client! the total energy consumption by both our servers and the 100’s of millions of clients is an utter waste.

    HTML and CSS are well entrenched as the cornerstones of web based publishing. Logic dictates that instead of relying upon a browser to interpret our markup and script, we should be using a different development environment that does some pre-compiling chores before the content is published to a server and ultimately to the client.

    The biggest innovations introduced by IE7 and 8 has been the inclusion of Developer Addons that has allowed developers to detect and correct errors before they are pushed out.

  49. Rob says:

    @Rob Parsons,

    IE did not innovate Developer AddOns as these were in Firefox and Opera long ago.

  50. Lenny says:

    @Matt – no I didn’t miss-read or miss-understand the post at all.

    You access a site… you are on a secure connection (HTTPS), thus you WANTED to go there.

    A dialog pops up, asking you to confirm that you are ok with this security miss-match.

    "Do you want to…"

    YES of course I do!

    Wait, NO, that’s not the answer you are actually looking for because the remainder of the question reverses what it is really asking.

    "Yes" is a positive response.  "Yes" I want ice-cream.  "Yes" I’d like to install this program.  "Yes" I’d like to receive a big pile of money.

    Thus "Yes" I’d like to NOT load the portions of this page that I intended to see… is not a valid combination.

    UI is all about Usability.  That new dialog is a usability fail. period.

    I’m not saying I have the perfect solution – what I’m saying is the current dialog is a WORSE solution than the previous one.

    I’m sure it was well engineered and well intended, but the end result failed.

    The original dialog had the correct context… "Do you…"  …"Yes"

    …then again even mentioning usability on the IE blog is a dead end.  The bad layout of this blog has been discussed numerous times and the only "fix" that has occurred is that a working search was added via a bing popup window.

    Even still the bing search is broken too.

    search this blog for "test" and you get just shy of 4 MILLION results!!!!! 4 MILLION! wow!

    search this blog on Google and you get the correct number: about 644 results.

    So, Bing 4 million results and decide?


    Google it – and find what you need!

    Grrr! the captcha on this site is STILL broken!

  51. EricLaw [MSFT] says:

    @Lenny: You are misunderstanding the security threat. In your example, the security threat isn’t that the secure CRM site is evil. The security threat isn’t that the insecure Google Map hosted in that CRM page is evil.

    Rather, the security threat is that because the Google Map content was delivered insecurely, an attacker could use that "hole" in the secure envelope to compromise the secure CRM site. That’s why secure pages should include only secure resources.

    I think there’s little sense in arguing over whether or not "Do you want to see only secure resources" is a "positive" question or not.

    Vis-a-vis the CAPTCHA on the IEBlog: While this isn’t under the control of the IE team at all, a common thing that causes problems for folks is that the CAPTCHA uses a session cookie across the entire MSDN Blogs page. So, if you do something like:

    1> Open this post. Start writing a comment.

    2> Decide you want to refer to some other post. Open that other post in another browser tab. Read it. Close it.

    3> Try to submit a comment on the post in #1.

    You’ll type the CAPTCHA on page #1. That will fail, because the blog system is expecting you to type the CAPTCHA that was shown on page #2, which you likely didn’t even bother looking at, since you were only reading that page and not trying to comment.

    Sorta lame, yes, but once you know how it works, it’s fairly easy to avoid.

  52. Gary says:

    Ignoring all the other commenters who see this as anothere excuse to hit on old, current or even future versions of IE, I for one welcome the decision to map interoperability between IE and other browsers.

    Will the information be a comparison between IE and other browsers, or IE and W3 standards (and those "standards" still going through the process)?

  53. Adrian Bateman [MSFT] says:

    @Gary: we are documenting where IE has variations from certain web standards including some from W3C. The reference point is the standard specification. We document final approved standards now and when relevant standards going through the process become final we will document those too.

  54. Rebecca says:

    @Eric and @Lenny

    There’s 3 problems with these scenarios.

    1.) using a url reference of about:blank, javascript: data: etc. is expected in an application.  Historically IE choked on some of these legit urls.

    2.) a secure application may well mash in an insecure resource (not ideal, but sometimes you do what you have to do)

    3.) the dialog.

    – The dialog is a "Security Warning"

    – Thus the buttons are [Yes] / [No]

    Ignore the exact warning and circumstances for a moment and you’ll see why the dialog is bad.

    example (bogus) dialog:

    Security Warning:

    Do you want to load this site that may contain rick-roll links?

    [Yes] / [No]

    The key points are:

     "Do you want…." and the Yes/No buttons

    In the English language when you ask a question that starts with: "Do you want…." and the possible answers are Yes and No, you can substitute the users action by re-reading the question substituting "Yes I" or "No I do not" for the "Do you" portion of the question to get a statement that signifies their response.

    Thus the 2 answers are:

    [Yes I] want to load this site that may contain rick-roll links


    [No I do not] want to load this site that may contain rick-roll links

    If you apply this logic to the Old and New questions you get:

    [Yes I] want to display the nonsecure items.


    [No I do not] want to display the nonsecure items.

    versus the new dialog:

    [Yes I] want to view only the webpage content that was delivered securely


    [No I do not] want to view only the webpage content that was delivered securely

    The last statement is the issue.

    "No I do not want to view only the webpage content that was delivered securely" – very confusing!

    This is the typical answer that users want to select yet it doesn’t read cleanly at all.

    Its typical because this is only a warning and for most users they would only worry about this if the page being loaded was where they enter their credit card details…. all else they don’t worry.

    By definition a "warning" indicates that there *may* be an issue *if* you proceed along the path you are already on.

    If Microsoft(1) really feels that this should be stronger than a warning, then the red page that BLOCKS the user from accessing nasty content should be presented instead.

    (1)Everyone else is fine with a warning.

    In changing the dialog I would also change the word "webpage" back to just "page" as the page might be loaded up inside a context or a corporate application that doesn’t resemble the web or even display telltale browser chrome.

    Can Microsoft confirm that IE8 currently only shows this warning for http: urls that exist in the page? and that any other "helper" urls don’t trigger this dialog? (e.g. is the problem with this dialog appearing inadvertently in IE now fixed?)

    I should also add (since the topic is standards) I’m quite glad to see that Microsoft is working on becoming standards compliant.  It will be good to see IE on par with other browsers for the first time.

  55. Rebecca says:

    I meant to sum up:

    The dialog should present a yes/no question that is either:

    load the scary stuff – yes/no


    load everything – yes/no

    attempting to rephrase it as:

    load only stuff that meets a condition – yes/no

    just confuses the user and forces them to pause.

    I’d recommend that everyone on the IE staff get a copy of Steve Krug’s book: "Don’t make me think"

    Its well worth the read and will point out why the new dialog is bad and how it fails to clarify a simple question for the user to answer.

  56. Mark says:

    Rebecca, your "logic" makes no sense to me. I think the problem here is that change is always confusing for some.

    But none of this has anything to do with the topic of this post. Let’s stay on topic, shall we?

  57. Hilman says:

    @Mark – yes we are all going off topic here but Dan brought up a good point.

    The dialog changed, the behavior changed, and the new dialog is confusing and illogical.

    We’ve all found the dialog to be annoying and awkward to read it is just that now that Dan has pointed it out we’re bringing up the reasons why the new dialog has made things worse than the original dialog that was simple and to the point.

    I think this post would have had more comments on topic if it had been posted 4 years ago.  We all care deeply about standards but we all gave up on IE supporting them when IE7 was released without including any JavaScript improvements.

  58. wai says:

    Hi IE team, I have an experience want to share. Some days ago I uninstall a svg plug-in and IE begins to crash when visiting acid3 test website(it crash when it reaches 12/100, at this point usually there is a top yellow bar ask for running MSXML, but it doesn’t before it crash)

    I tried no-addon/in-private mode, reset setting in Option page, but still problem exists. XML DOM Document in "Manage Add-ons" showed "Enabled".

    Finally I have a trial to do install msxml3.msi, this fix the crash on acid3 page.

    It would be nice if IE can report source of crash in next browser start-up, or a tool to do self-test on its dependencies?

  59. Wow. The IE team never ceases to amaze. Basically, this post says "keeping up with standards is tough".

    You have the market share… please develop a PROPER browser, or do what you do in Europe worldwide: give the people a CHOICE.

  60. helpless net user says:


    I for one welcome our new "Do you want to view only the webpage content that was delivered securely" overlord!


    I think we can all agree that the new warning dialog is a disaster. If you are asking the user an "important" question and the answer they need to provide isn’t clear there is a problem.

    With the old dialog, pressing the [X] or Escape would dismiss the dialog and Decline the prompt (e.g. no mixed content) which made sense.

    With the new dialog, pressing the [X] or Escape

    does the same thing EXCEPT it dismisses by ACCEPTING the default buttons’ action (e.g. Yes) – this goes TOTALLY AGAINST all UI behavior concepts in Windows.

    I hope this prompt gets fixed in IE9.  Users deserve better, readable, simple interfaces.

  61. infinte says:

    The only thing we need is:

     * addEventListener and event capture

     * standard Stylesheet

     * native application/xhtml+xml support

     * <canvas>

     * no VBScript and no COM, (maybe DLR and Managed assemblies, explicitly enable COM by user)

     * SVG

     * XML namespaces

  62. greg says:

    the only thing we need is (also):

    CSS vertical-align

    CSS rounded corners

    CSS opacity (that doesn’t use/rely on filters)

    CSS rgba()

    DOM NodeTypes declared e.g. TEXT_NODE=3

    DOM Mutation events

    JavaScript .cloneNode() (full support)

    JavaScript .getElementsByTagName() (full support)

    JavaScript ability to change "type" attribute

    JavaScript .getElementsByName() (full support)

    JavaScript .innerHTML support on ALL elements that can contain HTML!

    HTTP Referrer – needs to be sent on ALL requests regardless of origin

    HTTP requests for favicons should respect link tags that specify a desired favicon and not make a request for a root/favicon.ico (this would also include support for all valid web image types (GIF,PNG,JPG))

    Items inside res://ieframe.dll or res://C:{Winpath}system32IEFRAME.dll or res://wdsShell.dll should NOT be copied into the IE cache folder (no HTML, CSS or JS from the DLL)

    and a whole bunch more things so that IE is almost kind of sort of caught up to other browsers.

  63. Matt says:

    >I think we can all agree

    ROFL! You must be new here.

  64. infinte says:


     Managed, not Native

     DirectX, not GDI

     WPF, not WinForm

     Future, or DIE.

  65. Clint says:

    Looks like Windows Mobile 7 is dying before it even gets released!

    Skype has already ditched Windows Mobile devices:

    "Skype Pulls Its Windows Mobile Apps" – Mobile device owners running Windows Phone Classic (i.e. any version of Windows Mobile OS) will no longer be able to download Skype……With Windows Phones 7 not scheduled to come on the market until the end of 2010, though, this latest decision could mean that there will be an extended period of time during which Skype for Windows Phones will simply be unavailable.

    iPhone, BlackBerry or Android FTW!

  66. Todd says:

    Does it really matter what is in IE9? It’s not like anyone will use it.

    Heck, I don’t think I’ve even touched IE7 *or* IE8 since they were released. I don’t even develop for IE unless it’s I’m forced too – why bother.

    The whole thing is laughable. I love how you try and play it up as "embracing standards" when the only reason you started developing Internet Explorer again is because you were losing market share to better browsers.

  67. AntiTroll says:

    >It’s not like anyone will use it.

    Yeah, it’s not like IE is the browser used by far more users than any other browser.

    Oh, wait. It *is* exactly like that.

    Troll elsewhere.

  68. Marian says:

    To the Microsoft team, a friendly suggestion: don’t ignore the smell, and don’t lose the oportunity to revamp IE with a new engine.


    "General improvements:

    Now using a WebKit based rendering engine for the client and in-game overlay web browsing components (replacing Internet Explorer)"

    This cite is about a new Steam client released by Valve:

    The user base of this application is "25 million users, 1000+ games, 12 billion player minutes per month, and 75 billion Steam client minutes per month"

    Replace trident with webkit please !!!

  69. AntiAntiTroll says:

    Todd is right. A small chunk of the web population will use IE9 because MS insists on supporting its badly outdated brethren. IE may have a collective majority, but no single major version number has any kind of dominance. It’s saddening to see how few have upgraded to 8, or even 7. Version 9 will further diffuse the product line.

  70. MathIsHard? says:

    AAT: I know math is hard, but when IE8’s numbers are bigger than anyone elses (and they are, worldwide) that means that IE8 is the "dominant" browser version. When all IE browser versions combined have nearly twice the marketshare of *all* competitors combined, that’s means IE is the "dominant" browser.

    Marian: Video game web browsers necessarily have dramatically different use cases than desktop browsers. Steam’s move is basically meaningless.

  71. Peter says:

    While I believe Microsoft is making great efforts in getting back in the "league" with Internet Explorer 9, I also think they’re still being way too vague. There’s blog items on documenting standards now, there’s news about joining the SVGWG and even proposals to help editing the 2D Canvas documents.

    Next to this, however, there’s nothing confirmed. It’s unclear whether SVG will be available in IE9, how far CSS and HTML/DOM support will go, also in terms of new tags (including CANVAS, and, more importantly, VIDEO due to the Theora/H264 debate), or which version of the ECMAScript Standards will be implemented.

    Mitch 74’s statement basically said the same thing; all other browser vendors have open development processed, frequent snapshots, beta versions and sometimes even alpha- or nightly versions of their software. There’s public bug trackers, mailing lists and roadmaps sharing their visions. Their support of web standard is visible through active participation in the WhatWG and W3C mailinglists. Microsoft, however, still barely communicates.

    Open up already, invite a bunch of people to your internal mailing lists and get feedback on your plans and ideas. Be concrete instead of spreading hints and rumours like you’re doing now, with these posts. Listening to your users does not mean exclusively listening to people within Microsoft who happen to use your browser, it also includes listening to web developers, other browser vendors and basically everyone who might get in touch with the software at one point or another. Getting back in the league on the technical side of things really isn’t everything, you have to get back in the league in terms of participating in standards, openness and communication as well, and right now I’m truly worried about those aspects.

  72. Chris says:

    Maybe Microsoft’s trump card will be to release WMV and WMA as open specifications for web implementation, meaning no more worries about patents creeping up, as Microsoft will be giving up theirs for use within HTML5 <video> and <audio> elements.

    If Microsoft do this, I’ll forgive every bug IE has ever brought on us 🙂

  73. evan says:

    @Chris – Uhm… I don’t think I’ll ever be able to forgive every bug IE has brought on us… likely not even one.

    Just finished watching the Olymics where the world comes together to celebrate sport – it was awesome!

    I just wish the world could come together and drop IE6 and IE7. The world would truly be a better place and the Internet would thank everyone for a better web.

    Please everyone… make 2010 the year that IE6 dies and IE7 gets its eviction notice.

    The web wants to move on – only IE is holding it back.

  74. ivanetta says:

    @evan the feeling is mutual.  To see all the hate for IE6 in realtime just follow this link on Twitter:

    Only a few days until the funeral:

    Unbelievable that the most exciting moment in Web Development is not a new feature, a new technology, or a new browser…… but the death of an old one.

    Fare well IE6! You won’t be missed!

  75. JJJ says:

    Can anybody help me with this question:

    Internet Explorer remembers my username and passwords (e.g. for Only one username and password is associated with this website. Can IE fill this info automatically so I only have to click ‘sign in’ (like Firefox). Currenty I have to click the username field and select the stored username. Thanks.

  76. tammy says:

    The site appears in lots of search results near the top (or at the top) depending on the day and appears to be a copy of this blog.

    Is this a site that Microsoft is transitioning to/from?  The captacha thing there doesn’t work so I presume this is the site.

    If the other one doesn’t belong to Microsoft can you guys send the other site a notice to either re-design their site so that it doesn’t look the same or do a C&D.


  77. harold says:

    Ok – before I go nuts is there a setting in IE that has changed recently?

    Randomly but with a 70% chance or better when I launch IE – external urls (Google, Bing, Facebook, anything) work just fine – but attempting to load ANYTHING off my local environment (e.g. http://localhost:8080/foo/bar/baz.php) fails completely!

    The tab icon just spins endlessly, the status bar goes to 1 bar, and checking Fiddler2 IE never even sent the URL Get Request!

    Its not my local server as I can access the page(s) by localhost/ip address in Chrome, Firefox, Safari or Opera without any issue and all return a 200 or 304 (cached resource) status code.

    I’ve tried it (in IE8) in no addons mode and still no better. (seems to always work after a reboot – but thats a bit extreeme)

    Is there a registry setting I can look at? e.g. did something override the default "app" handler for local URL’s in IE?

    PS (its been happening for about a month+ now – finally got annoyed enough to post a comment)

    All ideas/suggestions on how to fix IE welcome.

  78. EricLaw [MSFT] says:

    @JJJ: Sorry, no, you need to select the username from the dropdown.

    @harold: Fiddler isn’t a reliable way to test this, because localhost traffic bypasses Fiddler ( by default. Do you have a 3rd party antivirus product installed? We’ve heard of "link scanner" type plugins that exhaust the per-site limit of HTTP connections and prevent further requests to target URLs.

  79. harold says:

    @EricLaw – this "link scanner" type plugin… might it be made by a company with 3 out of the following 26 letters?


    Since I might have said application installed but have said link scanner disabled could it still be causing interference?

    Note I have Fiddler installed to redirect all wininet.dll traffic for monitoring and when IE is working, I see all my localhost traffic just fine.

    I also do not have a certain S&D anti-malware app installed so it isn’t that.

  80. Philip Kahn says:

    How about taking actual steps to kill IE6? Like, the next service pack for each OS you have installs the most up-to-date IE version, with you needing to go out of your way and check a checkbox that says something like "Do not update Internet Explorer; keep version 6 over version 9"?

    There, you’ve suddenly halved IE6 levels. You need to commit to standards (which I’m not sure is actually going to happen, but I’m cautiously hopeful) AND help people move from the nine-year-old travesty that is IE6.

  81. hAl says:


    Yo might conisder uninstalling that anti-malware product and use Micrsoft Security Essentials in stead.

    That has also a lot less CPU footprint

  82. JJJ says:

    @EricLaw: Thanks, can you make this a feature for the next version of Internet Explorer?

  83. Mitch 74 says:

    @hAL: why not both? In S&D’s case, it’s not that difficult to have it installed but non-resident (you need to -gasp uncheck a couple checkboxes on install), so as to run it once in a while to ensure your system’s clean…

    As for installing newer IE versions with service packs… No. For a few reasons.

    – SPs are supposed to be a collection of system fixes, and at best update a few libraries that retain backward compatibility; as it stands, IE 8 is not backward compatible with IE 6. It is also worth noting that if you install your SP through Windows update, then you probably got IE 8 already anyway. Those that don’t, don’t use WU anyway.

    – that wouldn’t help in XP’s case: since it went to ‘extended support’ phase, it won’t get a SP4. And anyway other OSes will either stop being supported in a few months (Win2k) or already have a more modern browser (Vista, 7).

    However, if it were possible to slipstream IE 8 in a XP install CD, now THAT would be great! Or slipstream Vista SPs and IE 8 in Vista install DVDs, or with Seven… I must admit that while installing Vista is fast, updating it is doggone SLOW. But this is not the place.

  84. hAl says:


    Harold was not referring to Spybot S&D.

    But on S&D as you mention it. Even when not a resident scanner it was shown to interfere with IE8 workings when this browserversion was released by means of overpopulating the restrited sites list.

  85. Al says:

    I don’t think there was a conclusion stated about the broken mixed-content dialog.

    Since the Feedback site is now down I can’t track the bug there 🙁

    Can we get a status update on whether it will be fixed in IE9 or not?

    Additionally when it is fixed will it be back-ported to IE8 installations too?

    It seems quite strange that this bug made it past QA before shipping.  Is this because MSFT is still using the Waterfall programming method?

  86. hAl says:


    Trouble reading names ?

  87. Can read fine says:

    @Literacy – I think we all read the other article no problem.

    What we’re pointing out is that the new dialog is still a failure – and in most peoples opinion is now actually worse than the original since the dialog is essentially backwards.

    If Microsoft would like to address the comments and confirm/deny that they are going to address the usability/readability of the dialog in that would be great.  Being silent on the issue indicates 1 of 2 things.


    A.) they are naive to believe that there is no issue and haven’t stepped back to look at this with fresh eyes.


    B.) they do intend to fix this – however until they have a solution in place that they are ready to push to developers (eg. IE9 beta 1) they would rather not talk about it until there is something to show that is better.

    Unfortunately silence is perceived as (A) not (B)

  88. DT says:

    The old dialog was negative with the normally-correct response being no, while the new dialog is positive with the normally-correct response being yes.

    The text itself is debatably clunky, and I don’t really have an opinion on it one way or the other, but the current dialog is clearly not backwards.

  89. Do The Wrong Thing! says:

    I think it’s funny that everyone is so excited about the mixed content screen.

    You only see this screen when the website has a bug and doesn’t know how to use SSL properly.

    And yes, there are some sites that don’t use SSL right, and IE warns the user.

    It is funny to me that there’s so much whining at Microsoft, and not more people complaining that the other browsers are not as safe, and not more people complaining to the websites that aren’t using SSL correctly.

    But I guess nobody ever got fired for blaming Microsoft. And it’s easy for the lazy!

  90. Vidar says:

    @Eric – I agree. There’s no good reason why Microsoft should continue developing their own rendering engine, when they’re never going to catch up with the competition in terms of standards compliance.

    Adopt Gecko or Webkit (or Opera’s Presto for that sake) – I don’t care. If you’re worried about backward compatibility with legacy intranet applications – include Trident as well, and provide a compatibility mode, which will let the user set up a list of websites to be rendered with Trident.

  91. @Sylvain Galineau [MSFT] – thanks for the correction.

    @whydontmicrosoftusewebkitpeople  When IE was the only game in town it fully satisfied nearly all users.  Today we all have a choice.  The price we pay for this choice and daily innovation is a few inconsistancies. If eveyone still used IE then there would be no inconsistancies.

    @MicrosoftDevs Keep up the good work. Versions of IE continue to get better and IE9 is eagerly awaited by many.

  92. Vidar says:


    "When IE was the only game in town it fully satisfied nearly all users. "

    What planet were you living on? IEs implementation of standards were a major head-ache for web developers around the world. And I’m not just talking about how IE implemented standards differently from other browsers – there were (and still are) what I consider bugs, but since Microsoft documented them, has to be considered implementation specifics.

    Take this for an example: Try to register an onchange event handler for an input field. Works well in all browsers, including IE, until an IE user selects a value from IEs autocomplete feature instead of typing the value themselves. This does NOT trigger the onchange event, for reasons only Microsoft can understand. They were kind enough to document it [1], but that doesn’t mean it is good.

    That’s just one example from a developer’s view – then add all the security issues the everyday user were victims of.