The joy of being me


So, we’re having a fun thread on the HTML WG.  I’m explaining what Microsoft must do, given the half-billion or so users that rely on us to not break their web experience.  A certain contingent disagrees with me (okay) and seems to want to beat it out of me (not okay). 

I’m reading through email.  My wife leans over my shoulder at a random one – sent privately to me, even, and honestly, it didn’t even wiggle my annoyance meter.  (We have a new out-of-office system on the server, I’d accidentally mis-configured it and it was sending external messages.  This person was letting me know, in a moderately polite way.)

My wife says, “wow, people are really rude.  You don’t need that.  Why are you still doing this?”

 Hmm.
 


Comments (63)

  1. Ok, I’m replying in entirely the wrong forum, because this is a reply to the linked message from the mailing list, but one of the joys of having one’s primary email be webmail is that mailto: links in web pages don’t work any more. (Ironic, huh?)

    But your post said "If they’d just sent us the content they sent to firefox, we’d have done a better job" and then later on "Show me another way to Not Break The Web". Well, how about sending a firefox-ish UA string then?

    Implement the standards at least as well as Firefox does (still got a long way to go there, but I hope you’re planning to be a lot closer by IE8) and then send a Firefox-like UA string (take out "MSIE", stick in "like Gecko" like Safari or Opera or whoever it is does). Presto, you’re getting served standards-compliant content.

    It’s not like this is a new idea. You were ON the team that decided IE’s UA string should start with "Mozilla/"…

  2. Sandi says:

    re:"We have a new out-of-office system on the server, I’d accidentally mis-configured it and it was sending external messages"

    You’re not the first person to have done that :o)  Back in January another softie committed the same faux pax, leading to his cell phone and office number being sent to anybody who sent an email to an MS competition email alias.

  3. If Microsoft ends the current Microsoft Internet Explorer line of browsers and call the next (more exciting!) browser Microsoft Internet Adventurer then you have a fresh start. IA would have a new navigator.userAgent string (something straightforward like Opera’s) and no document.all, conditional comments, etc. IA would have all the hackable stuff like the "child selector hack" fixed so that those hacks don’t apply in IA just like they don’t apply in other browser. If IA does not respond to any of the current browser sniffs for IE then IA is a clean path to the future. IE is in a big popularity decline and the trend will probably continue without something big changing.

  4. My wife asks me the same thing.

  5. I haven’t yet been accepted to join the discussion, so I’ll post my comments here for the time being.

    I don’t buy into this notion that every new version will have to break the Web again if standards support is improved. Sure, that will be true with IE for the next few versions, but it isn’t an eternal thing and it isn’t a problem significantly experienced by anyone but IE.

    The reason is simple: IE is in a hole right now. It was left alone for half a decade while the other browsers fixed up their engines and developed a solid standards-compliant foundation. IE doesn’t have a solid foundation yet. Maybe after five more years of unrestricted improvement. But for now, with every improvement, IE currently has to change some pretty fundamental stuff like what the "width" and "height" properties even do. That’s major. Other browsers don’t have bugs anywhere near that level of impact.

    Let’s say you build a site using all of the modern CSS features, and it works in current versions of Firefox and Opera. Then new versions of Firefox and Opera come out with majorly improved standards support. Guess what? Your site still works in the new versions. Magic? No. It’s just that the browsers didn’t *have* to significantly change its handling of the features you used. They had it right in the first place.

    IE’s problem is that it *isn’t* doing what it’s supposed to be doing in the first place. If you guys could wave a magic wand and bring the IE engine to where Firefox and Opera are today, then you wouldn’t have anything in the way for future additional standards support. And, assuming you implement things correctly the first go in future additions, you won’t run into this "paradox" again. Even with 80% market share, adding standards support wouldn’t break anything worth mentioning.

    IE is very much in a transitioning stage. Yes, there will be some major growing pains for the next few versions, but this is not the result of just being the #1 browser. It’s the result of trying to incrementally fix an engine which gets a huge amount of things wrong in widely-used areas.

    As a web developer, I see no reason for Firefox, Opera, or Safari to implement a new kind of quirks mode. Newer versions just don’t break things like newer versions of IE do. If you want to implement a new quirks mode in IE, I guess that’s your call. I for one hope that I don’t have to be stuck with IE 7’s engine until 2010 or whenever HTML 5 is finalized.

    If you must implement a new quirks mode system, here is how I would do it: HTML 4.01 gets a snapshot of the current layout engine. This would be quirks mode 2 or whatever you want to call it. XHTML 1.x served as text/html would also get quirks mode 2. HTML 5 and anything run through the XML parser (including application/xhtml+xml once it’s supported) would get the new standards mode. Preferably, there should also be a way to get standards mode in HTML before HTML 5 is ready. I suppose some kind of IE-specific comment thing could do the job. Maybe a doctype comment?

    <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd&quot; –[ie 8]–>

    Something like that? Assuming that non-IE browsers don’t bother supporting a new quirks mode, this should be sufficient.

  6. cwilso says:

    @Stuart:  We couldn’t send a firefox-ISH string, we would have to send exactly the firefox string from some particular version.  And we’d be expected to have that version’s bugs, etc., an not support ActiveX, and…

    The short version is that production web coders are pathologically precise today.  🙂

    @Peter: that move would be majorly disruptive.  At least as disruptive to users as just fixing all standards bugs in quirks mode.

    @David: I agree with you that the bulk of the problem is because we’re in that hole today, and internally we’ve already made it a credo to "do it right the first time" – cheesy, I know, but it’s exactly what you’re talking about, that if we were very careful to properly implement according to spec, it wouldn’t be such a problem.  (There are still some potential problems, some of which affect us today – because the specs aren’t clear or complete, for example).

    We certainly will provide some way to get into "really standards" mode prior to HTML5.  There’ll be a post asking for community feedback on that soon.

  7. Chris, what exactly is the big technical disruption of starting fresh with a newly labeled browser? It would only require a few technical changes to make the browser not respond to the IE sniffs. Leak the newly labeled browser slowly to the public with beta release numbers below 1.0 so you can fix up the important bugs that have cause people problems with IE. Then when you hit 1.0 you don’t support the pre 1.0 versions. People using IE can continue without any pain until all the computers using IE are dead and gone.

    Developers should not have to do something extra to a validating HTML 4.01 document that has a DOCTYPE clearly saying that to get into standards mode on any browser. The DOCTYPE already says they want standards mode.

  8. Your assertion that you’d have to send EXACTLY Firefox’s UA for some particular version is contradicted by one simple fact: Firefox is able to release new versions without "breaking the web"…

    Point is, web developers *expect* Firefox to be releasing new versions on a regular basis, because they always have. They (understandably) got out of the habit of expecting IE to do so.

    Send something that looks vaguely like a plausible future version of Firefox, implement standards as well or better than current Firefox versions, and you’ll render the web as well as Firefox does.

  9. Stuart, I think that changing the UA string would be a good move but that is only one of the several common sniffs that people use in JavaScript program specifically for IE. The other sniffs (document.all, conditional comments etc) would need to be removed also which would also be good moves. I don’t know any developers that lament these missing features in other browsers. Perhaps some intranet developers but they can continue using IE6 if they want to.

  10. Chris,

    I too caught the email that Chris Messina linked to last week. As a developer and a user, I’m kinda struck by some of the things that you were mentioning in the email thread. The main thing though that caught my attention was when you started speaking on how the I.E. dev team has always had this mantra of "Don’t Break The Web!". For the most part, that’s ok, but it also struck me as odd: Because I.E. doesn’t support current CSS, HTML/XHTML (and Javascript) standards and because I.E. is the browser with the largest market share, THE WEB IS ALREADY BROKEN.

    For the life of me, I can’t understand why you fail to see that. I understand that you fully want to help the company that you work for remain a market leader, but I think that you hold a bigger responsibility to the millions upon millions of users around the world whose lives revolve around the Internet and what it provides.

    It is for this reason and this reason alone, that you, along with the other devs in the team, should be wholeheartedly pressing for building the I.E. rendering engine from scratch so that it supports established web standards from day one. I know, I know…you’re going to say that you do hold users as more important which is why the team has the mantra "Don’t Break The Web!". However, I would assert that that is an incorrect statement.

    According to the email, the main reason why the dev team still rely on quirks mode is because of all of the negative attention that I.E. received from web developers angry that all of their hacks no longer worked. This is the problem in a nutshell: Microsoft (which I understand you can’t wholeheartedly speak for) doesn’t foster an atmosphere for educating developers that they should be coding using todays established web standards. The first thing that I asked myself after reading that passage in the email was: Did Microsoft and the I.E. team before hand really go out and educate that hundred of thousands of developers about web standards and urging them to fix and re-code pages and stylesheets so that when I.E. 7 was finally released you could avoid the exact reaction that you received? If I had to answer, I don’t think "you" did.

    And I have to say that I’m a bit taken aback by the assertion that the I.E. team shouldn’t have to do away with Quirks mode or work extra hard at producing a stable, standards complient web browser simply because I.E. holds the largest market share. You may have only been speaking for yourself in that thought, but regardless, it makes Microsoft sound wholly arrogant. That IS NOT a good thing.

    The way that I see it, you have a two fold problem:

    1) You need to code and test a fully standards complient capable web browser.

    2) You need to place more work and effort into educating web developers as to WHY they need to create their sites using standards complient code and HOW to produce those sites without resorting to CSS- and browser-sniffing hacks.

    The answers to each problem, IMHO, are easy ones.

    1) Reach out to the Opera, WebKit and Mozilla devs so that you can all come to aggreement on how to render the basic aspects of modern HTML/XHTML/CSS the same. Don’t worry about the advanced features like Canvas just yet. Just concentrate on the basics.

       Reaching out to them for help doesn’t make your team weak or put them at a disadvantaged position (after all Microsoft is still a business). It just shows that the I.E. dev , and by virtue Microsoft, is concerned with rendering the web like the other browsers do so that EVERYONE receives the same experiences while utilizing the web.

    2) Do more to educate the masses of developers about Web Standards and why they should be used so that you can avoid things like shipping a new browser and not having your Moms bank website break. If you show that using Web Standards is important to Microsoft, then Web Standards will become important to the developers who only code for I.E. and Windows (and trust me….there are still millions of developers that do this. I come across them daily while surfing in Safari.)

    I know that it feels that everyone is dis-regarding what you are saying and as the head of the I.E. dev team that has to suck. But you have to understand that we want a Web that works…correctly, the way that it was meant to, not the way that I.E. makes it work. The only way that this is going to come about is for you to realize that "THE WEB IS ALREADY BROKEN!" Chris will you help us fix it?

    P.S. – I sincerly apologize for the extrmemly long comment. I’ll try to NOT let it happen again.

  11. mrpinto says:

    I’m in agreement with the theory that anonymity combined with audience leads to rude behavior online.  It seems logical to assume that there will always be critics of such an enterprise as IE.

    That said, isn’t there a risk that adopting a mantle of martyrdom might lead to conflating the rude ones with those whose gripes are both legitimate and politely (or at least acceptably) voiced?  

    What I see is a situation of frustration without legitimate outlet.  Web developers are frustrated that IE was stagnant for 5 years, that IE7 is still behind, that IE7 is taking a long time to replace IE5-6, that public bug reporting was of limited utility to begin with and then shut down, etc.  These all strike me as legitimate concerns.  

    On the other side, I see the IE team circling the wagons, claiming the status of righteous martyrs, ignoring (or at least failing to publicly address) legitimate concerns, and failing to offer a satisfactory forum for productive dialog.  You are (correctly) concluding that there’s just a group of assholes out there who won’t be happy no matter what.  Then, it appears to me that you’re (incorrectly) concluding that this means that you should close ranks and shoot back at anyone who offers constructive criticism..

    How many times have we seen the IE team belittle the efforts of those who have publicly collected bugs?  Engage in sideshow debates over things like what exact percentage of this or that CSS spec remains to be implemented in IE distracting from the plain fact the IE isn’t even in the same ballpark as its competition?

    I’m not going to call you names or demand that you give up and switch to Gecko, or anything like that.  I understand that there’s a huge userbase and a complicated codebase and backward compatibility.  I’m sure it’s hard.  

    Despite all that, the folks from Gecko, Webkit and Opera are demonstrating that even if it’s hard, it’s possible.  MS should have more resources in this arena, not less.  Nonetheless, IE is now and barring unexpected change in course will remain a pain in the butt for serious web devs.  To make matters worse, I really feel that the IE team tends to treat web developers disrespectfully, even those of us who don’t belong to the "rude" camp.

    My humble advice would be to acknowledge and empower the reasonable ones and ignore the unreasonable ones.  

  12. Why IE v.next? Why not IE2 v.1

  13. Dave Sussman says:

    I think that whatever you do you’re going to get hosed. If you don’t fix the bugs and adhere to standards (and IE7 is a vast improvement) then you’ll be accused (even more) of monopoly and not caring. If you fix the bugs and break existing sites, then you’ll be accused of not caring about customers (and yes, I agree, the web is already broken, but stopping existing sites working isn’t a good solution). If you create a new browser (my choice), then you’ll be accused of confusing the market place.

    It’s a matter of which you consider the lesser evil. Personally I’m in favour of a new browser; yes it’ll be a lot of work, but the long term benefits will be worth it – how far ahead do you want to look? Do you still want to be maintaining the IE code base in 5 years, modifying it for HTML 5 and CSS 3? At what stage do you consider the cost effectiveness of bolting on features when you know that it would be better to start again.

    I believe that we – Microsoft, web developers, educators, authors – have to educate the entire web world about the problems and why standards are good. We know that many developers don’t produce compliant code, but that’s partly a matter of education, partly toolset. Users don’t care, but if they understood the reasonings behind a new browser, they’d be more forgiving.

  14. Cecil Ward says:

    Chris,

    I think you’re on the right lines with your ideas on HTML versioning.

    It seems to me that when the IE team examine a particular web site’s behaviour and then try to make generalisations as to how best to do the right thing with maintaining backwards compatibility one of the hard things to ascertain is just what the website designers were expecting/intending way back whenever it was built. You pointed out that seeing a website marked with all the signs of "standards mode" doesn’t tell us enough about the author’s understanding or intentions. Indeed a website may carry a STRICT doctype simply because someone copied it from somewhere thinking it was the "right thing to do", or because some tool wrote it in there.

    I would suggest that we really do need an opt-in mechanism, but such an opt-in needs to be rather more "intentional" and "explicit".

    Your point about the misuse of STRICT doctypes in my view is resolvable in this way. If a website author goes to the trouble of explicitly declaring that they have validated their page, then they are probably serious about standards compliance, and so this is to be treated as a case for _real_ standards mode when considering how to render the page.

    So I had a go. I just made the following syntax up, in 30 seconds as an idea of what kind of things we might want to know in order to do archaeology well.

    <!–

    $$[[target_browsers="microsoft.com/ie/6.0;mozilla.forg/firefox/1.5/;opera.com/9.0"

    validated="w3.org/validator/xhtml/1.0/strict"

    date="2007" ]] –>

    Which tells us

    i. Author intent: I tested against the given set of browsers.

    ii. Validation: I validated my web page. I validated it against the (say) XHTML1.0 Strict spec, using some well known tool such as the W3C validator.

    We would need to issue firm guidance – that these directives must not just be copied into pages robotically in boilerplate fashion, and that authoring tools may only insert a validated directive if they incorporate validation as part of the build process. Authors must be honest about the target_browsers directive, which specifies the set of browsers they have actually tested against.

    That is the kind of thing I’m thinking of, but the details need to be worked out properly by first making up use cases.

    I’m well aware that this could indeed be abused, but we could take further steps to discourage this.

    How about this. Add a validation button to IE8 and a coloured "quality flag" indicator. A "high quality web site" lights up the quality flag in some way. Sites that declare themselves to have been "validated" get a gold star (providing of course they are not full of obvious errors). On the other hand sites that are not marked as validated get no gold star and sites that are full of soupy badness as reported by IE’s parser get a golden onion indicator. These indicators should not be in the user’s face, but they could be the start of a culture change in site authoring, as there would be a drive to get your site to "gold starred" qa level. (A bit like the "reward" that you get from using an EV certificate.)

    You guys really have a great chance to reform the culture of web development just now, as well as a chance to make the process of working out how to do backwards compatibility with more certainty in future, so the web can advance without being held back by "BC-FUD".

    Go for it.

    Best.

  15. Dean Edwards says:

    Chris, can you explain why Peter’s idea of discontinuing the Internet Explorer *brand* would be "majorly disruptive"? Disruptive to whom?

  16. cwilso says:

    users.  Users who visit their bank sites and expect them to work.  

  17. Al says:

    Personally, I’m still waiting to hear why my baby was shut down after I left Microsoft (the bug database). It was a huge uphill slog both internally and externally getting that thing going and then it goes away.

    That might not be a public discussion though. 🙂

  18. Al says:

    After reading this, my innate reaction is:

    "Well, duh."

    Since I was one of the people (the only one for a while) buried in the triage of these bugs, I can certainly speak to the number that were basically "You SUX0R!!!".

    That being said, it seemed like it didn’t get as much support as I would have liked. This was probably a factor of it starting very late in the product cycle. I dunno. It was a shame to see the IE Blog go to running a lot less posts with bigger time periods in between after I left combined with that happening to the bug DB.

    It seemed like a lot of outward facing community efforts got effectively defunded to some degree.

  19. Dean Edwards says:

    "users.  Users who visit their bank sites and expect them to work. "

    If that was the answer to my question then I won’t bother to ask another one.

  20. "users.  Users who visit their bank sites and expect them to work."

    Chris, this is not a very fulfilling answer. It answers the whom but not the why. Such a short answer is difficult to interpret. The terseness appears to be intended to shut down conversation without a discussion or at least educating the developers that are disappointed with many previous IE design decisions but willing to actually spend time commenting here. Unfortunately my initial interpretation is "look here peons, we know best, we don’t want to do that and we aren’t going to tell you why so just accept it."

  21. Al Billings says:

    The point that Chris has made over and over again is that IE won’t break backwards compatibility, period. Neither users nor developers give Microsoft and the IE team any slack if they go to their site and it doesn’t work, regardless of why.

    Other browsers seem more inclined to just update things and figure that users will come along. Given IE’s browser share, that isn’t really an option.

  22. cwilso says:

    Yeah.  What Al said.

    Dean and Peter – I’m not trying to rudely shut down discussion of this topic.  I’m burned out discussing it right, now, actually – go take a look at the HTML WG thread and see how many posts I’ve made on it explaining why we believe in this.

    No, I can’t give you every piece of data we have.  Some of it, we don’t have rights to give to the public (e.g. contracted studies) and some of it paints a negative picture of web developers that I’d rather not explain in the news.

    In short, the "why" is that there are a tremendous number of ignorant web developers out there, for whom DOCTYPE is cargo-cult boilerplate.  (Ian Hickson’s term, btw.)  They copy-and-paste with abandon, then they tweak it until it works in IE.  They’re annoyed, today, because we have bugs they have to work around.  Then, we ship a new version that fixes those bugs, and their tweaks break – and they scream at us.  This is how the IE7 cycle went.  It doesn’t make our browser share or our market perception go up, despite being the "right" thing to do.

  23. cwilso says:

    Forgot to point – the HTML WG thread on versioning is at http://lists.w3.org/Archives/Public/public-html/2007Apr/thread.html, look for the "Versioning and html[5]" thread started by me.

  24. Craig Ewert says:

    Dean Edwards asked:

    <blockquote>Chris, can you explain why Peter’s idea of discontinuing the Internet Explorer *brand* would be "majorly disruptive"? Disruptive to whom?</blockquote>

    And cwilso replied:

    <blockquote>users.  Users who visit their bank sites and expect them to work.</blockquote>

    Let me try to elaborate.  I’ve been to many sites (My bank may be one) that flatly refuse to work on anything besides IE.  I’ve even seen one that refused to work if IE wasn’t your default browser.  If Microsoft put out IE8 that claimed to not be IE, all of those sites would fail to work, and users who just updated their browser would be screwed.  If, on another hand, Microsoft put out a 2nd line of browsers (or 3rd, counting the butterfly line), then the bulk of everyone everywhere would be still using IE, not the new thing.

    I should also point out that I’ve tried once or thrice to contact the webmasters of sites that refuse to work with Firefox et al; those contacts have been uniformly disappointing.

  25. Vanessa says:

    my ie7.0 worked good yesterday.

    but today it always appears a error message and then it asked me if i wanted to send an error report. i used to have this problem, after i reinstall windows XP SP2, the problem seemed disappear, but now it comes out again.

    if you could help me, could you email me at "voicesay@hotmail.com"

  26. Sander says:

    But Craig, if MS put out IE8 that claimed not to be IE – and announced they’d soon be auto-updating everyone, banks would be falling over themselves to fix their broken sites. (And who knows, they might even start just allowing _all browsers_.)

    Despite understanding the whole "not wanting to break websites" idea, I can’t help but to see this as a very positive outcome; this is a case where the entire world would applaud Microsoft strong-arming some utterly broken websites into the modern age like that.

  27. It sounds like the crap developers are stopping Microsoft from fixing it’s browser because these developers use hacks. That is very unfortunate. Every developer should know that using a hack is fragile and will require maintenance. These developers would be irritated in the short term but within a few months developers would start thinking Microsoft’s browser is so much better and easier to satisfy when building pages. Developers might even stop recommending Firefox to their family and friends since IE comes installed. I think this would be a great time to just rip off the band-aid and fix the long standing bugs in IE. Then the evolution of the web can move forward much faster.

  28. Mitch 74 says:

    Interestingly, I contacted my bank a few years ago because their website wasn’t working with Phoenix 0.6 (‘Firefox 0.6’, if you want). They answered me that their website required either Internet Explorer 5+, Netscape 4.7 or the Mozilla suite.

    When I told them that Phoenix was basically a stripped down Mozilla, I got the answer:

    ‘we’re evaluating it’.

    When Firebird 0.7.1 came out (Firefox 0.7.1), the website worked. I scrapped IE that same day.

    If you tell me that your bank won’t do that, then maybe your bank isn’t very customer-oriented – and you might want to consider a change for a better bank, as this may be an indication of bad things to come.

    About bad web designer: it is unfortunately true that many web designers don’t know what a DOCTYPE is, get scared when mentioning ‘validation’ – and those same developers make their clients pay a bundle for their services, and don’t plan to ever come back on the website they’ve created.

    Interestingly, several designers I’ve met don’t work freelance but for an internal service, and tend to use standards for their designs – so that they don’t HAVE to come back to the page later on.

  29. Tino Zijdel says:

    This isn’t about users but about forcing your way into the specs in such way that it will be more difficult for other vendors to compete.

    Proclaiming IE.current’s behavior as ‘the standard’ and requiring authors to specify something in addition (specced or not) to get ‘the other’ (W3C) standard’s behavior is preposterous and will only lead to users continuing to cater to IE’s bugs and shortcomings and thus even more broken websites for non-IE users.

    I thought the browser-wars were over, but it seems that only the battleground has shifted…

  30. cwilso says:

    Tino, I don’t believe you’ve actually considered it from the other side of the fence for a minute.  I’m not trying to "force my way into specs" – I’m trying to not break the web for a half a billion users, the bulk of whom don’t understand the difference between IE’s current behavior and the W3C standard, in my opinion shouldn’t be forced to pay the price for that ignorance.

  31. Tino Zijdel says:

    Chris: has MS even considered it from the other side of the fence? Have you not considered the alternatives that won’t break the web at all but don’t require authors to go out of their way to use HTML.next in IE or fall into the trap where they are unknowingly still using IE’s substandard mode (because they don’t understand the difference) resulting in broken sites in other browsers?

    Here are my thoughts:

    – use the HTML5 DTD as an explicit opt-in switch (and treat documents w/o DTD but with XHTML-mimetype the same)

    – make sure that IE.next’s implementation of HTML.next is as close to perfect as possible

    – shorten the release-cycle for standards-compliance fixes so authors won’t start relying on bugs in your implementation

    – re-instate the public bug database so people can follow bugs, know when they are expected to be fixed and/or find out whether work-arounds exist

    I don’t think it is fair to have authors that want to use standards, and other browservendors to pay the price of your ‘problems’ either.

    Oh, and when I say ‘you’ I mean Microsoft ofcourse 😉

  32. Dean Edwards says:

    Quit Microsoft. Remain head of the HTML WG. Keep your integrity. Make a difference.

  33. Lionel says:

    >Quit Microsoft. Remain head of the HTML WG.

    >Keep your integrity. Make a difference.

    Ugh… and how rude is this, do you think?

    I do have the impression that many people are missing the point Chris is making.  This is not about HTML5 and IE.next.  IE will use the HTML5 doctype to trigger really-really-standard mode.  It will also use some kind of comment-based switch to trigger this mode in previous versions of HTML.  As for matching the standard as closely as possible, it’s another obvious decision.  (Does anyone believe they enjoy shipping bugs and the resulting support problems?)

    The whole point about versionning is about the far future, when HTML5 will be commonplace.  Maybe everyting will be perfect, HTML will always be backward- and forward-compatible, and will evolve in such a way that no breaking fix will ever be required from browser vendors.  In this case, any versionning may be safely ignored.  On the other hand, it might not go so smoothly forever.  Web developers might go on using browser sniffing to serve different pages, with different bugs and bug workarounds, to different browsers, and with no clue about how recent a document is, it would again be a choice between backward compatibility and standards, or back to ugly opt-ins.

    When Chris is asking for versionning, he is not wearing his "Microsoft hat", he is wearing an "experienced browser developer" hat.  He’s not asking for something Microsoft needs, he is asking for something that is already best practice for network protocols (HTTP, SSH, …) and does no harm in the (unlikely, IMHO) event everything goes as planned for HTML5.  No doubt he could give in and not worry about hypothetic post-HTML5 opt-ins in IE14, but I guess *this* is where integrity matters.

    I’m happy to see in the Internet Explorer team someone who is ready to point out future problems even if it makes him unpopular.

  34. @ Dean Edwards

    If Chris W really wants to make a difference then Microsoft is where he needs to be, as that’s the place that needs to start being different.

    For my day job, I keep sites in quirks mode, as there’s less chance a new version of IE is going to come out and break it. If IE fixed certain parts of its renderer to a current version of HTML, it would encourage larger companies to adopt modern standards.

  35. Anders Borum says:

    Interesting discussion (as always around here).

    I work with (some of) Microsofts server products. It’s quite strange that Microsoft wants to push standards and yet still are utterly unable to deliver a framework (such as MOSS 2007) that renders wellformed HTML / XHTML.

    The left hand says: "We need to push standards. Standards rule!"

    The right hand says: "Who needs standards. Let’s continue rendering our all-caps-broken HTML and give developers hell on earth."

    Quite strange that you have such talented developers, and yet we still see such low quality HTML / XHTML from your software (again relating to MOSS 2007).

    I am simply unable to state just how dissapointed it is keep working around software that doesn’t comply with standards.

    That be MOSS 2007 or IE7.

    Best suggestion so far (among the previous posters) is starting from scratch with a new browser. Make sure you keep the current quirks mode (as implemented in IE7) to ensure that the millions of users you keep referring to are able to browse broken sites.

    We seriously need a new render for HTML5 in IE8. As suggested elsewhere, please start using your own user agent string and make sure the browser is as close to the standards as possible.

    Chris, you’re in a position in Microsoft to make a difference (and your work is appreciated – so is the work by your team on IE). I truly hope you’re listening to all the (relevant) feedback provided in these blog posts.

    Creating a standards compliant browser is not for the benefit of external developers alone.

  36. thacker says:

    The Broken Web.

    Wilson–

    You made an excellent point regarding bank Web content. Many of which were built exclusively around IE6.

    It took some balls for Microsoft to introduce IE7 and piss off large corporate users because their code was written without standards or forward thinking, users who make significant contributions to Microsoft’s bottom line.

    You also hit the nail on the head about the crap

    produced by what I refer to, politely, as "Hobby Content Developers". However, Microsoft has to accept responsibility for that to a large extent. Frontpage was the most widely used "Web development tool" in existence. It helped to promote and foster the belief that anyone who purchases a canned application can develop Web content and call themselves a Web designer or developer. Microsoft promoting Word as an HTML authoring tool didn’t do anyone any favors, either.

    Not to mention, IE being the biggest guy on the block and its attraction to the script monkeys in regards to security.  I damn sure do not envy you.  

    The IE team and the re-direction of Microsoft in its Web development strategy, including Word 2007 abandoning the IE engine, has shown a lot of guts and courage.

    That, folks, has earned my respect and from an SOB who rarely passes out compliments to any man, woman or child for any reason.

    All things considered for having to maintain balance for past, present and future, your idea of including some trigger within IE8 that will initiate a true hard-core standards mode makes a lot of sense. Just keep it simple and stupid, please.

    By the way, thanks for support of the W3C P3P standard. I am probably the only existing fool who considers that standard as having merit and who consistently develops P3P policies.

    Hopefully, IE8 will get banged out the door sooner rather than later in 2008.

    Thank you very much.

  37. Steven Clark says:

    The funny thing about people being rude in emails, its very similar to when you phone the local department store with a complaint – very rude…

    when you trot down to said store standing in front of you is some little guy in a suit usually…

    its that feeling of anonymity (I’m safe in my little room so can say anything). Its really hard not to take it personally and not let it bring your day down, especially when you read that at 6am on the way between bed and work.

  38. Jalf says:

    Isn’t the problem of "breaking the web" a bit overstated?

    I mean yes, a lot of content has been written for and around IE6, but try using Firefox for a few weeks. There are hardly any sites that work *exclusively* on IE, regardless of version. Which means that somewhere, somehow, in some form, pretty much all sites are able to deliver reasonably compliant HTML. They might not do so when it’s IE asking, but they’re able to do so when other browsers do.

    The only sites I know of that I can’t view properly in Firefox are:

    – Netbanking (requires ActiveX)

    – Microsoft Update (requires ActiveX)

    – (and the occasional MS Update-like download manager)

    Within the last two years or so, I have seen precisely no websites that fail to render correctly (or acceptably) in alternate browsers. (Which might lead to another interesting point. Perhaps IE isn’t the market leader, despite its huge market share. Perhaps the market leader (as in, the one people go out of their way to support) is Firefox? Perhaps there is no longer a clear market leader (because people make sure to support both)

    True, if even the tiniest percentage of IE users experience problems, that’s a lot of people you’re pissing off, but another thing to consider might be this:

    You say that 1% of websites would break if you broke backwards compatibility. That’s a lot of sites, I agree. But how many of them already have a more standard-compliant rendering mode? I’d guess 99% (since most of them already work in alternate browsers.)

    That means only 0.01% of the web would break. Now consider that *if* Microsoft were to require some form of standards compliance, how long would it take for, say, 90% of the remaining websites to get cleaned up?

    I’d give it a month or two.

    Okay, so all our numbers (yours and mine) are completely made up, but my point is simply that it’s not *just* a matter of "how many websites would break if we dropped a bit of backwards compatibility.

    There’s also "how many websites might seem to break now (ie. how many break if I tell Opera or Firefox to identify itself as IE6), but could and would be fixed almost by flicking a switch so they use the Firefox/Opera/Standards-rendering path they already have, for IE as well?

    How many *fundamentally* broken (as in, works *only* on IE6-compatible browsers) are there really? Because they’re the ones that cause problems.

    The ones that already handle better browsers than IE6 would be able to handle a non-backwards compatible IE with ease as well.

    And of the fundamentally  broken websites, how many would get fixed very quickly if a new version of IE suddenly behaved roughly the same way as other modern browsers?

    Now I suspect we’re down to a very small number of websites that would actually be a problem.

    And finally, there is one other point to consider.

    True, some web developers might scream and shout if they can no longer use their IE6-specific bugs to control rendering, but on the other hand, how many people are inconvenienced by having to literally develop two versions of their websites?

    Where are the cost biggest? In causing so much extra work for 99% of the web, who have to support ancient and mysterious bugs, or by forcing the last 1% to upgrade?

    The point is that keeping IE6 compatibility is not the "no-cost" option you make it out to be. The costs are huge because of all the quirks web developers have to deal with.

    True, the alternative, of actually breaking compatibility isn’t exactly no-cost either, so the question is which solution is cheapest (in terms of development time for web developers, in terms of number of people who hate Microsoft, in terms of IE marketshare or whatever metric you prefer)

    We’re not talking about the expensive (upgrade and break compatibility) path vs. the free one (leave things as they are, only improve where we can do so without breaking anything), but about two different kinds of expensive.

    That said, I do think what you said in that "fun thread" is true. The only way to sort out the current mess is to use an "opt in" approach to standards compliance.

    (By default, websites are a mess. The developer has to specify explicitly that he wants his page treated according to the standard)

    It’s hardly a perfect solution, but I can’t see a better one.

    Oh, and finally, another way to speed up the process of cleaning up the web might be to have a little chat with the ASP team. Where is the sense in *by default* render noncompliant code? Why is XHTML *Transitional* used as the default doctype?

    If part of the idea behind ASP is to hide all the low-level details of actually messing with html, shouldn’t it at least do it right?

  39. harold says:

    Just read the linked msg on WG.

    Wow!, that cleared up so much stuff for me.  I’ve been one of those "kinda have it in for Microsoft" folks, but I started to see exactly what battles you are up against.

    I like (no, scratch that….) LOVE the idea of being able to set something (doctype) or otherwise, that indicates my desire for stricter rendering.

    harold

  40. v1lDaY says:

    view party who else http://glucotrol.8000web.com/ green world elephant sorry. glucotrol specific is very good http://glucotrol.8000web.com/order-glucotrol.html is great site super. order glucotrol Order green world everything http://glucotrol.8000web.com/glucotrol-dosage.html specific my love and Good work! glucotrol dosage very good If date http://glucotrol.8000web.com/medications-glucotrol.html specific birthday property. medications glucotrol anyway automatically past http://glucotrol.8000web.com/glucotrol-2.5.html while this time have a good day! glucotrol 2.5 elephant green world drink party http://glucotrol.8000web.com/side-effects-of-glucotrol.html Order detach Good work! side effects of glucotrol

  41. Kat says:

    Again, thank you for explaining this position, in Australia this year. It was teh awesomeness.

  42. bilgi yarışması says:

    That said, I do think what you said in that "fun thread" is true. The only way to sort out the current mess is to use an "opt in" approach to standards compliance.

  43. Webdesign says:

    The reason is simple: IE is in a hole right now. It was left alone for half a decade while the other browsers fixed up their engines and developed a solid standards-compliant foundation. IE doesn’t have a solid foundation yet. Maybe after five more years of unrestricted improvement. But for now, with every improvement, IE currently has to change some pretty fundamental stuff like what the "width" and "height" properties even do. That’s major. Other browsers don’t have bugs anywhere near that level of impact.

  44. Webhostin says:

    There’s also "how many websites might seem to break now (ie. how many break if I tell Opera or Firefox to identify itself as IE6), but could and would be fixed almost by flicking a switch so they use the Firefox/Opera/Standards-rendering path they already have, for IE as well?

  45. The reason is simple: IE is in a hole right now. It was left alone for half a decade while the other browsers fixed up their engines and developed a solid standards-compliant foundation. IE doesn’t have a solid foundation yet. Maybe after five more years of unrestricted improvement. But for now, with every improvement, IE currently has to change some pretty fundamental stuff like what the "width" and "height" properties even do. That’s major. Other browsers don’t have bugs anywhere near that level of impact.

  46. medyum says:

    If you must implement a new quirks mode system, here is how I would do it: HTML 4.01 gets a snapshot of the current layout engine. This would be quirks mode 2 or whatever you want to call it. XHTML 1.x served as text/html would also get quirks mode 2. HTML 5 and anything run through the XML parser (including application/xhtml+xml once it’s supported) would get the new standards mode. Preferably, there should also be a way to get standards mode in HTML before HTML 5 is ready. I suppose some kind of IE-specific comment thing could do the job. Maybe a doctype comment?

  47. Medyum says:

    The IE team and the re-direction of Microsoft in its Web development strategy, including Word 2007 abandoning the IE engine, has shown a lot of guts and courage.

  48. medyum says:

    Your assertion that you’d have to send EXACTLY Firefox’s UA for some particular version is contradicted by one simple fact: Firefox is able to release new versions without "breaking the web"…

    Point is, web developers *expect* Firefox to be releasing new versions on a regular basis, because they always have. They (understandably) got out of the habit of expecting IE to do so.

    Send something that looks vaguely like a plausible future version of Firefox, implement standards as well or better than current Firefox versions, and you’ll render the web as well as Firefox does.

  49. Personally, I’m still waiting to hear why my baby was shut down after I left Microsoft (the bug database). It was a huge uphill slog both internally and externally getting that thing going and then it goes away.

    That might not be a public discussion though. 🙂

  50. adana haber says:

    tank u very much for this site very nice tanks again

  51. hikaye says:

    Stuart, I think that changing the UA string would be a good move but that is only one of the several common sniffs that people use in JavaScript program specifically for IE. The other sniffs (document.all, conditional comments etc) would need to be removed also which would also be good moves. I don’t know any developers that lament these missing features in other browsers. Perhaps some intranet developers but they can continue using IE6 if they want to.

  52. Again, thank you for explaining this position, in Australia this year. It was teh awesomeness.

  53. The IE team and the re-direction of Microsoft in its Web development strategy, including Word 2007 abandoning the IE engine, has shown a lot of guts and courage.

  54. There’s also "how many websites might seem to break now (ie. how many break if I tell Opera or Firefox to identify itself as IE6), but could and would be fixed almost by flicking a switch so they use the Firefox/Opera/Standards-rendering path they already have, for IE as well?

  55. sikiş says:

    I am grateful to you for this great content.

  56. The IE team and the re-direction of Microsoft in its Web development strategy, including Word 2007 abandoning the IE engine, has shown a lot of guts and courage.

  57. Again, thank you for explaining this position, in Australia this year. It was teh awesomeness

  58. thank you for information admins..

  59. medyum says:

    If you must implement a new quirks mode system, here is how I would do it: HTML 4.01 gets a snapshot of the current layout engine. This would be quirks mode 2 or whatever you want to call it. XHTML 1.x served as text/html would also get quirks mode 2. HTML 5 and anything run through the XML parser (including application/xhtml+xml once it's supported) would get the new standards mode. Preferably, there should also be a way to get standards mode in HTML before HTML 5 is ready. I suppose some kind of IE-specific comment thing could do the job. Maybe a doctype comment? http://www.medyumoguz.net

  60. mehmet can says:

    güzel çalışma admin

                                                      http://www.cityotokiralama.com