The Internet Explorer 8 User-Agent String

As Dean announced, a beta version of Internet Explorer 8 will be released in the first half of 2008. 

The HTTP headers of every web request contain a simple string called the "User-Agent" that allows servers to identify the type and version of the visitor’s browser.  For a deeper understanding of the User-Agent string, please see Understanding User-Agent Strings.

When released, the IE8 beta will introduce an updated User-Agent string.  For IE8, we’ve simply replaced "MSIE 7.0" with "MSIE 8.0".  For example, on Windows Vista, IE8 sends:

Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0)

Best Practices
Websites that are sensitive to changes in the User-Agent string may need to be updated to support IE8. 

During internal testing, we have encountered a small number of sites which fail to recognize IE8 because they are performing exact string matches to look for specific IE version strings. Those checks will need to be removed or updated to accommodate IE8.  The article Best Practices for detecting the Internet Explorer version provides suggestions and code samples.

As noted in the Best Practices article, it is not recommended that you block access to content based on the user-agent string of the browser. If you do have to offer different content to different versions of the browser due to improved capabilities, you should ensure that future versions of the browser are not blocked. Serving content based solely on the user-agent string is often an unreliable way to detect the full capabilities of the browser, because the user might have adjusted some settings, such as disabling script or extensions.

Detecting Beta Versions
For IE7 beta, a "b" was added to the User-Agent string

Mozilla/4.0 (compatible; MSIE 7.0b; Windows NT 6.0

... to identify the browser as a beta release.  While this was consistent with prior IE releases, it caused many site compatibility problems.  Server code tried to cast the string between MSIE and the next semicolon into a floating point number, and failed when the "b" was encountered.

To prevent similar problems in the IE8 beta, we will not include the "b" in the IE8 beta User-Agent string.  Pages can detect if they are running inside a beta version using the following script:

if ((null != window.navigator.appMinorVersion) &&
(window.navigator.appMinorVersion.toLowerCase().indexOf("beta")>-1))
{
window.alert("This is a beta browser.")
}
else
{
window.alert("This is NOT a beta browser.");
}

User-Agent Spoofing

If you’d like to see how your sites react to the new IE8 User-Agent string today, you can use the existing registry-based User-Agent override mechanism to cause IE7 to masquerade as Internet Explorer 8. 

For the convenience of beta users, a new menu option is available in beta versions that will enable the IE8 beta to masquerade as IE7 without restarting.  When the “Report IE7 User-Agent string” option is checked, IE8 will send the IE7 User-Agent to allow beta users to interact with any sites that have not yet updated to support IE8.

Eric Lawrence
Program Manager

Comments

• Anonymous
  February 21, 2008
  Awesome. Can't wait for the beta!

• Anonymous
  February 21, 2008
  Looks nice! Keep the good work!

• Anonymous
  February 21, 2008
  The comment has been removed

• Anonymous
  February 21, 2008
  You've just ADMITTED that even WITH super standards mode as an opt-in, IE8 will STILL cause site breakage. The question remains: are you incompetent, or do you just hate developers?

• Anonymous
  February 21, 2008
  It's time to drop the "Mozilla/4.0" from your user agent string. The user agent string is relatively useless but keeping "Mozilla" in there is just another indicator that folks at Microsoft won't get with the times and are an anchor on the web. Of course, some is forgiven if addEventListener, event capture, and working get/setAttribute are included in IE8.

• Anonymous
  February 21, 2008
  > Serving content based solely on the user-agent > string is often an unreliable way to detect > the full capabilities of the browser Tell that to the Silverlight team! I know that Silverlight works in Firefox 3 and pretty much any other Gecko browser, but I guess Microsoft doesn't realise that.

• Anonymous
  February 21, 2008
  Considering the "don't break the web" mantra, I'm curious as to why you guys have gone with a UA string that will, apparently, break certain sites. Why is one level of breakage acceptable while the other is not? Do you guys perform any stats and then have a percentage cutoff point at which you decide that it's an acceptable loss? Is it possible to share these stats?

• Anonymous
  February 21, 2008
  Great news on the beta. Just please make sure that you can run it as a standalone and that it doesn't over install itself over IE7

• Anonymous
  February 21, 2008
  @Jonathan: If they didn't change the UA, how would the server know to send the UA-Compatible version of the page?

• Anonymous
  February 21, 2008
  The comment has been removed

• Anonymous
  February 21, 2008
  Feature detection is preferred over userAgent sniffing these days. What new features should we be detecting? What behaviours have changed? How do we identify these changes? Thanks, -dean

• Anonymous
  February 21, 2008
  Amen Tom.  Yeah it would be nice not to break the web, but IE will always be one step behind unless they get it out of the rut it is in.  No, super standards mode.  Have standards mode on from the get go and if developers want their sites to not break they can go back in and put in a special meta tag to bring the site back to non-super-standards mode.  As a developer I would rather go back into my site and limit it to a previous version of IE than tell all my new sites to use a new super duper cool standards mode.  Microsoft, listen to the developers.  You don't want to break the web, but the developers are telling you to break it ... we'll clean up the mess that your browser has left us with (once).

• Anonymous
  February 21, 2008
  "If you’d like to see how your sites react to the new IE8 User-Agent string today, you can use the existing registry-based User-Agent override mechanism to cause IE7 to masquerade as Internet Explorer 8." Another "brain fart", no surprises here, if we can already pretend to be IE8 why bother with IE8 at all, after all isn't IE8 just going to pretend to be IE7? The fumes in here are becoming more and more noxious.

• Anonymous
  February 21, 2008
  I continue to wonder why IE's UA string keeps the "Mozilla 4.0 (compatible" in it, when IE has grown far beyond those days. I think it's time IE created a more accurate and fresher UA string for itself. I don't think it would break the web either, since it would be for a new version anyway.

• Anonymous
  February 21, 2008
  @Devon: It's pretty surprising how much of the web breaks when the legacy "Mozilla" portion of the string is dropped.   If you're interested, you can see for yourself using Fiddler (www.fiddler2.com).  You can rewrite the User-Agent header to anything you'd like, then visit major sites (e.g. www.cnn.com) and see what breaks.

• Anonymous
  February 21, 2008
  PingBack from http://blogs.dotnethell.it/vincent/The-Internet-Explorer-8-User-Agent-String.__12954.aspx Thank you.

• Anonymous
  February 21, 2008
  @Fred Clown -- what about webpages burned to CD?  How do you put an opt-out tag in them? Even if Microsoft were to announce TODAY that sites would need to opt out of standards mode, as you say you'd prefer, I guarantee that when IE8 actually ships there would be millions of pages that would not have added this tag and would therefore be broken.  Remember when IE7 was shipped and how much outrage there was from web devs about their sites being broken, even though all the changes were made to improve standards compliance? And remember that users don't care WHY it's broken; they just want it to work.

• Anonymous
  February 21, 2008
  It is kind of history, given current the current history of Mozilla, the community, the foundation, and the company, that Microsoft's super-duper new browser will still identify itself as somehow compatible with Mozilla. Does that mean compatible with standards? :-)

• Anonymous
  February 21, 2008
  PatriotB, the new switch is awfully convenient for Microsoft's goals of keeping people using an IE6 compatible mode (rather than standards) as the web continues to try to move forward. Of course, with more than 1 in 4 browsers being something other than IE in most of the world, this may not matter eventually. (Soon enough, it will be 40% or more...)

• Anonymous
  February 21, 2008
  @EricLaw: Firstly, www.fiddler2.com is completely atypical of a common web page. Its code is an atavism and the fact that it's a Microsoft site makes it all the more dubious. Secondly, Opera completely removed the word "Mozilla" from its navigator.userAgent string two years ago and has unfettered access to cnn.com and all major websites.

• Anonymous
  February 21, 2008
  As a semi-related thing, since you mention it.. is there any chance of leaving the quick UA-change mechanism in the browser? Since you've added one here for testing, obviously at least part of the code exists already. Most of the other browsers have a method for doing so, even if it's accessed through an extension. Not a requirement (hardly) but it'd be an appreciated item. Yes, I know this is a poor spot to mention what amounts to a feature request, but since you mentioned that it was partly in there for the beta... If nothing else I'd like to know why not, if you have the time to answer. Thanks! ^.^

• Anonymous
  February 21, 2008
  appMinorVersion.indexOf('b') you mean, not indexOf('beta')

• Anonymous
  February 21, 2008
  Please, get us out of the rut we are in. Changing the user-agent string to "Internet Explorer/8.0" will break a few poorly designed sites, but on the whole it will make it quite a bit easier to identify user-agents.

• Anonymous
  February 21, 2008
  @Gyrobo "Firstly, www.fiddler2.com is completely atypical of a common web page. Its code is an atavism and the fact that it's a Microsoft site makes it all the more dubious." Fiddler is add-in to IE that logs HTTP traffic and alows you to edit it.  The content of the fiddler site has nothing to do with it. "Secondly, Opera completely removed the word "Mozilla" from its navigator.userAgent string two years ago and has unfettered access to cnn.com and all major websites." A changed agent string could still cause issues because pages are often modified depending on the browser.

• Anonymous
  February 21, 2008
  The comment has been removed

• Anonymous
  February 21, 2008
  That means that an IE8 using the IE7 rendering mode (engine?) will identify himself non the less IE8 which will break the web what exactly is the reason why you implemented the IE7 rendering mode? That is wired, because the people who are using user agent sniffing are exactly the ones who make websites work only in IE7. Please get it a) right or b) dismiss that whole “don’t break the web” mantra. Because the only reason the web is broken is your plain wrong implementation of international standards in IE6, which lasted too long. Changes in IE7 were more cosmetic. And IE8 continues to continue that tradition. Why don’t you allow people to get a meta element which says “behave like IE7” which is, what the browser does then. Including the UA string. That would make just more sense than an IE8 which uses the IE7 rendering engine for CSS and IE7 JS but says to a website that it is not an IE7. This is just not true and every website that breaks is right spot on. Not breaking the web? Thats sarcasm, obviously.

• Anonymous
  February 21, 2008
  @Al "PatriotB, the new switch is awfully convenient for Microsoft's goals of keeping people using an IE6 compatible mode" Now you're on the other side of the fence, you seem to be enjoy the kind of rubbish you used to put down.  Not to mention the conflict of interest here.

• Anonymous
  February 21, 2008
  Hi IE Team, Keep up the good work guys. A lot of developers appreciate the new openess and the strive for better standars support. IE is finaly becoming the browser it deserves to be and the new meta target tag helps the IE team to advance more rapidly. Looking forward to IE8 and I hope IE9 will soon follow!

• Anonymous
  February 21, 2008
  "Keep up the good work guys. A lot of developers appreciate the new openess and the strive for better standars (sic) support." You're being sarcastic, right ?

• Anonymous
  February 21, 2008
  The comment has been removed

• Anonymous
  February 21, 2008
  And what do I find here. Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0; MS-RTC LM 8) compared to Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.2; MS-RTC LM 8) and Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0)

• Anonymous
  February 21, 2008
  Given how much IE8 will improve standards support, wouldn't it be time to put Mozilla/5.0 in the front, to match other highly conforming browsers? Hmm, I wonder how many sites that would break ...

• Anonymous
  February 22, 2008
  I just don't get this. If you are expecting sites to need to modify code due to hardcoded matches on the user agent string, why won't you make them modify their code to opt out of IE8 standards mode? @Dean Edwards: re: "Feature Detection" - I would agree, if this were legitimately possible.  If I test for if(document.getElementById){...} this won't help me one bit in knowing if the browser is IE 7 or less thus informing me that the implementation is completely broken.  Ditto with (setAttribute), yes a browser may support it, but I wouldn't suggest for a second that IE supports it (without major hacking). I too look forward to the Beta.  I hope that I can run it STANDALONE on my XP machine for testing, but I have this strange feeling that I won't be able to do either (since no one from [MSFT] has talked about it yet). @MSFT - How about some news here?  is SVG in the Beta?, what DOM methods & properties are fixed/implemented in the Beta?, does PNG Gamma work in the Beta?, has the UI had the major overhaul it needed in the Beta?, can I move/merge toolbars in the Beta?, can I drag a link to a new Tab in the Beta?, Does > Fullscreen > Un-fullscreen still reveal toolbars that were hidden in the Beta?, are the Reload/Stop buttons moved back to the correct location in the Beta?  What version of ECMAScript will the Beta support?  Has VBScript been dropped yet in the Beta?  Has printing been fixed in the Beta? Looking forward to your responses MSFT, Clive

• Anonymous
  February 22, 2008
  "if ((null != window.navigator.appMinorVersion) &&" shouldn't that be " null !== ".

• Anonymous
  February 22, 2008
  Hey Now! The beta is going to be great! Thx 4 the info, Catto

• Anonymous
  February 22, 2008
  @ Eric Drop Mozilla from the useragent please. If something breaks then the blame falls on the web designer or web developer masquerading as as web designer. CNN? Seriously, I doubt they have abandoned work on the site. Can you give us a better example of a site that is still popular though not maintained that will break without Mozilla in the useragent?

• Anonymous
  February 22, 2008
  it is really annoying to post a legit comment only to have it filtered out.

• Anonymous
  February 22, 2008
  Ok, lets try reposting piece by piece... Correct me if I am wrong, but your JavaScript test is broken. navigator.appMinorVersion returns a number! If you release IE8 Beta, with this as a string, you've already messed up. Since 0.8 would indicate Alpha, even 0.9 would be better than the String "beta".

• Anonymous
  February 22, 2008
  you explain how to determine the beta version... But where is the example? this one? Mozilla/4.0 (compatible; MSIE 8.0beta; Windows NT 6.0)

• Anonymous
  February 22, 2008
  That's a good idea! Good luck with IE8.

• Anonymous
  February 22, 2008
  Actually, the whole test case is overblown. navigator and alert are global objects, and never null...

• Anonymous
  February 22, 2008
  @DOM: No, appMinorVersion, like most DOM attributes, is a string rather than a number. Furthermore, at least one of the browsers I tried returns null for this property, hence the additional test.  Code which fully specifies the scope of the objects it is using runs faster. @Autosuchprogramm: The post clearly states the IE8 user-agent string, which does not contain the "beta" indicator for the reason previously described. @John: Users don't really care whose "fault" it is when their sites break.  They blame the browser, don't upgrade, and then you have to put up with older browsers longer. @aaron: Either != or !== would work in the code given.  It's not clear to me why you'd want to use !== in the case of a null check.  http://www.ecma-international.org/publications/standards/Ecma-262.htm @alan: The User-Agent string isn't render mode-specific, as the page gets to determine its own render mode, but the UA string is sent to the server before IE has the page used to determine the desired render mode. @Robbert Broersma: Your test for "b" would also be effective in this case, although it could return false positives in the admittedly unlikely event that there was a different "b"-containing string inside appMinorVersion. @Eric: I've written a simple addon that changes the user-agent string to a string of your choice on the fly.  I will release it sometime around the beta, and I expect it will continue to work in the final release. @Gyrobo: If you watch your network traffic in Opera, you will see that on startup, they download a list of compatibility "shims" that contains the list of sites that they need to spoof their user-agent string in order to successfully render.  While it's a clever idea, it does have a performance impact, and frankly it doesn't scale to the entire internet-- there's sites they've missed.

• Anonymous
  February 22, 2008
  The comment has been removed

• Anonymous
  February 22, 2008
  Keep up the awesome work. And please do your best to ignore the negative comments out there. ;-)

• Anonymous
  February 22, 2008
  @Mark Wisecarver What negative comments? Where? This has been a good start for a recent IE8 announcement.

• Anonymous
  February 22, 2008
  @EricLaw, thanks for the info.  I wasn't aware that any browser returned null.  I'm a bit surprized that window.navigator.appMinorVersion is quicker than navigator.appMinorVersion, but I appreciate the knowledge tidbit! I see that you didn't comment on any of Clive's requests for more info on IE8. Can you verify if IE8 will be available for XP? Can you verify if IE8 has significant DOM fixes? Can you verify if IE8 has native support for SVG? Thanks.

• Anonymous
  February 22, 2008
  @Clive and Dom Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 5.1) Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 5.2) They have tested IE8 on XP if I read the above UA string correctly.

• Anonymous
  February 22, 2008
  @Andrew: "The content of the fiddler site has nothing to do with it." I wasn't talking about the site's content, I was referring to its antiquated code. Just from looking at it, I can see improperly nested elements and deprecated attributes. "A changed agent string could still cause issues because pages are often modified depending on the browser." Exactly. If the default rendering mode is IE7, then the userAgent should identify itself as IE7. That's why the whole opt-in scheme reeks to me of hypocrisy.

• Anonymous
  February 22, 2008
  @EricLaw: I'm fully aware of how browser.js works. It's something that would be become completely unnecessary if all browsers would just support the same standards. Do you have an example of a site Opera's browser.js "misses" that specifically uses the "Mozilla" aspect of the userAgent string?

• Anonymous
  February 22, 2008
  @EricLaw: I'm fully aware of how browser.js works. It's something that would be become completely unnecessary if all browsers would just support the same standards. Do you have an example of a site Opera's browser.js "misses" that specifically uses the "Mozilla" aspect of the userAgent string?

• Anonymous
  February 22, 2008
  Hello all, It is widely known, widely admitted and widely acknowledged that user-agent string detection is not reliable, not best, difficult to maintain, not forward-compatible and not recommendable. On the other hand, object/method support detection is much more reliable, much more manageable, forward-compatible and overall best (assuming browser manufacturers implements correctly objects, methods and attributes). Why Microsoft does not update several of its MSDN documents to explain that, to teach that, to document all that? Lots of MSDN articles and MSDN webpages teach the wrong way to do things in code examples. For once, Microsoft's left hand should say and should do the correct thing while the Microsoft's right hand (MSDN) is editing something. Checking browser version, browser name, browser build number, etc... has never been important, never been useful or helpful per se. Testing and verifying if the support of a method or the support of an object about to be used in a script (assuming that such method or object is being correctly implemented by browser vendors to begin with) is what truly matters and what is utterly relevant in a javascript-DHTML-driven webpage. Regards, Gérard

• Anonymous
  February 22, 2008
  Remove me from the ice blog list. Thanks

• Anonymous
  February 22, 2008
  Remove me from the ice blog list. Thanks

• Anonymous
  February 22, 2008
  The comment has been removed

• Anonymous
  February 22, 2008
  Pas de surprise, la cha�ne d'agent utilisateur de la version 8 ressemblera � celle des versions pr�c�dentes, MSIE 8.0 rempla�ant MSIE 7.0.

• Anonymous
  February 22, 2008
  La bêta d'Internet Explorer est prévue pour le premier semestre de cette année. Toutefois, il n'est pas nécessaire d'attendre pour connaître le comportement de la prochaine version. Pour cela, il suffit d'éditer la clé.

• Anonymous
  February 22, 2008
  Firefox reaches the 500 Million download milestone! Think about that for a second. That's 500 million times that someone thought... "Hmm, there must be a better browser out there than the one I have!" I wonder how many of them are Windows users? (yes, blatent trollin' but with no info on IE8 to read about....)

• Anonymous
  February 22, 2008
  @"In related news": 500 million "downloads" divided by at least 12 new versions with security fixes. ~42 million users, maximum.

• Anonymous
  February 22, 2008
  Could I ask what will the IE 8 RSS user-agent be?

• Anonymous
  February 22, 2008
  Andrew, it isn't rubbish if it is true. :-) There is no conflict of interest. I am still a Windows user as well, often enough, which means that I have to work with IE. I'm also someone with a bunch of web sites that I've written so I have to deal with it there. My attitude is transparently obvious and hangs out like a blazing sign on my brow written by the gods. If I had some sort of secret agenda, your comment might hold water but my agenda is open. I want IE to actually communicate openly in public AT LEAST as well as it did when I worked on the team on IE7. That's the low end of the goal. The desired goal is to have the bug database back (better and staffed with a commitment to use it) and actual open communication, at least to the extent that Opera and Apple can manage, as closed source corporate entities, if not to the level that Mozilla maintains. If the IE team had actually really had anything to say about IE8, long term goals, or standard support, during the last year, I really wouldn't have a place to complain from, would I? The fact that I work for Mozilla now is no secret since it is blazoned across my blog and well known here. I'm still friends with a few people on IE as well though many people that I knew have since left the team (and sometimes Microsoft as well). So, what is your complaint? That I say churlish things when we get relatively useless posts about IE on this blog instead of actual content? Well, I can see an obvious way for IE people to fix that problem but, really, I probably count as noise to them, as with so many other commentors here, at this point.

• Anonymous
  February 22, 2008
  http://blog.mozilla.com/gen/2007/02/27/the-cost-of-monoculture/

• Anonymous
  February 22, 2008
  @Gyrobo "I wasn't talking about the site's content, I was referring to its antiquated code." My mistake in terminology, the intent was the same.  I felt your statement was irrelevant to Eric's original point of modifiying the agent string. "Exactly. If the default rendering mode is IE7, then the userAgent should identify itself as IE7. That's why the whole opt-in scheme reeks to me of hypocrisy." I do see your point, however it would be nonsensical to release a new browser (backwards compatible or not) and not update the User Agent string.  I'm sure even non-standards mode will have many code changes, so regardless of the rendering mode you are still running IE8. I'm sure the IE team has canvassed most scenarios but I'll throw this out anyhow: Let's say IE8 uses the new rendering mode by default but can intelligently fall back to non-standards (backwards compatibility mode). This could be achieved by interrogating the page prior to rendering for (unmistakable) non-standard DOM-calls or non-standard HTML tags.  The browser could indicate to the user which mode IE8 was in (something like pure or compatibility - not that anyone but us would understand). Obvious flaws:

• A performance impact that would keep IE8 at a disadvantage to competitors and IE7/IE6.  Caching the test result might help with repeat visits. The Opt-in meta tag would help.

• It would get it wrong a lot, so sites would still break; hopefully not as many.

• Could not detect behavioural intent particularly with CSS.

• Anonymous
  February 22, 2008
  The comment has been removed

• Anonymous
  February 22, 2008
  Come on, just support standards from the 90s and upwardsa and people don't have to bother about user agent string, this is hilarious.

• Anonymous
  February 22, 2008
  @Al "Andrew, it isn't rubbish if it is true. :-)" I objected to the statement because it's particularly subjective given your employment.  Needless to say I disagree with it. “There is no conflict of interest” You work for Mozilla, so it is in your interest to encourage negative sentiment towards IE/MS to benefit FireFox. “The desired goal is to have the bug database back” It will be interesting to see if the bug tracking experiment (aka IE bashing log) will be effective in it's next incarnation with the release of IE8 beta. “If the IE team had actually really had anything to say about IE8, long term goals, or standard support, during the last year, I really wouldn't have a place to complain from, would I?” The IE team have made statements committing to standards support and have released information regarding IE8.  Whether or not they were made last year doesn't seem relevant to me. “So, what is your complaint?” You worded them well: “That I say churlish things when we get relatively useless posts about IE on this blog instead of actual content?” We obviously disagree on the merit of information posted. “Well, I can see an obvious way for IE people to fix that problem” My main complaint Al is that unless you post solutions we can debate, then your repeated statements about openness seem hypocritical, and are clouded by your position with Mozilla, regardless how much you claim to advertise this fact.

• Anonymous
  February 22, 2008
  @Andrew: My statement is FAR from irrelevant to the original post. It is the ESSENCE of the post. If the default rendering engine is IE7, the browser should identify itself as IE7. If the IE team wants to change the user agent in "super" standards mode, then by all means, they should go for it. But the IE team justifies the default nonstandard behavior of IE8 by saying that if the rendering engine is changed IN ANY WAY that pages will break. And by changing the default user agent, what we'll see is a browser rendering in the same way as IE7 but not identifying itself as such. It seems to me to be contradictory and illogical.

• Anonymous
  February 22, 2008
  The comment has been removed

• Anonymous
  February 22, 2008
  @Gyrobo "If the default rendering engine is IE7, the browser should identify itself as IE7." No. The agent string identifies the correct browser version. If it did not, a server would not know to serve standards compliant content.  We would be in a much worse position. "It seems to me to be contradictory and illogical." It's not an ideal situation. The alternative is to be standards compliant by default, which would cause a massive delay to the adoption of IE8 and have a negative effect on advancement of web standards. BTW, what did you think of my solution?

• Anonymous
  February 22, 2008
  @Andrew According to the IE team, IE8 already follows your solution. The meta opt-in intelligently "falls back" onto true standards mode. Nobody from the team has fully explained the implications of the opt-in on initial page rendering and its (deleterious?) effect on performance. Getting back to the userAgent: if a browser uses IE7's css rendering engine, IE7's JavaScript engine, and has all of IE7's flaws in said engines, is not that browser, for all intents and purposes, IE7? You can't put a sweater on a pig and call it a poodle.

• Anonymous
  February 22, 2008
  @Gyrobo "According to the IE team, IE8 already follows your solution. The meta opt-in intelligently "falls back" onto true standards mode. Nobody from the team has fully explained the implications of the opt-in on initial page rendering and its (deleterious?) effect on performance." Mine is similar however given the discontent about an opt-in tag I suggested one wasn't required.  I imagine the performance hit of inspecting a page for a meta tag would be minimal. You seem to disagree with the proposals, so what is your solution? "Getting back to the userAgent: if a browser uses IE7's css rendering engine, IE7's JavaScript engine, and has all of IE7's flaws in said engines, is not that browser, for all intents and purposes, IE7?" Do you accept that some websites will want to support a full range of browsers including  earlier versions of IE plus IE8 standards mode? If the answer is No, we can stop here.  If the answer is Yes, tell me how a server can target IE8 standards mode (via the opt-in tag) without a change to the User Agent string within IE8.

• Anonymous
  February 22, 2008
  The comment has been removed

• Anonymous
  February 22, 2008
  Andrew, you speak of the next incarnation of the bug database? What makes you think that there is ever going to be a next incarnation? Do you have some insider knowledge? It has been "temporarily" offline for more than a year. It isn't coming back. You state: "My main complaint Al is that unless you post solutions we can debate, then your repeated statements about openness seem hypocritical, and are clouded by your position with Mozilla, regardless how much you claim to advertise this fact." I have a solution, openness. The IE team can start actually communicating with people, especially developers, about the content of IE8, the roadmap of IE's future, both with IE8 but also beyond that, and how they are going to address the issues that web developers still have with them. They can recreate (or create) a bug database in which developers can log IE issues and they can actually do something with the issues logged so that people can see progress. How about starting developer chats again? How about having interviews with people outside of Microsoft? How about sitting down with the makers of the other browsers (like Opera, Apple, Mozilla, not to mention others) to discuss how they can work together for the common good of the web? They can make milestone builds available, call them "alphas", to the development community and others so people can actually see what the IE team is doing. They can use this blog to actually communicate substantive things on a regular basis, not two or three times a year, to show that they actually care about their own community. Or they can not do these things and appear to be arrogant and to only care about the development community (or their users) when it is in their own personal interest to do so. That's what the past year and a half has been like. You can label my comments all you want, Andrew, but I worked on a few versions of IE over a decade and I've been working on both Firefox 3 and security releases of Firefox 2 for a while now. I think that I have a pretty good picture of the strengths and weaknesses, not to mention attitudes, of Microsoft, and IE in particlar, in comparison to other people.

• Anonymous
  February 22, 2008
  Acid2 をうけて 2008 内のマイルストーン展開を見込んでいる IE 8。 個人的にも 2008 年の最大関心事と位置づけて期待していますがプロダクトマネ...

• Anonymous
  February 22, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  Let me sum it up this way, and I have said the same to Microsoft earlier as well. The only people Microsoft even considers polling are the people as someone way above in this thread said "pay for their MSDN subscriptions". I would expand that list to include its titular reps known as "MVP"s. Microsoft has long since stopped caring about the developer and user communities. It cares only for what its Program Managers think about the product.  No one else. Take a look at the Windows Live Messenger feedback page on MS Connect --- the MS team there were being BOYCOTTED because they were simply and blatantly marking EVERY logged item as "Closed" and "Wont Fix". It was only after a mass-outrage that items were reopened and considered. This is CALLOUSNESS. This is ARROGANCE. Let me put it this way. I have mostly spoken in favor of Microsoft for the past 15 years, but I am making this post from Firefox 2. My websites today are first tested with Firefox more than it is tested with IE. And that is a strong message from an ex-MVP.

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  Why the need to detect a user agent string again?

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  @Gyrobo: Since you're clearly too busy to follow my suggestion to Devon of simply trying the "Mozilla"-less user-agent, www.yahoo.com and www.wellsfargo.com are two cases where removing the "Mozilla" token causes the site to break. Would you rather I did not reply, so you could save face?

• Anonymous
  February 23, 2008
  @EricLaw Your answer here has me wondering? "@alan: The User-Agent string isn't render mode-specific, as the page gets to determine its own render mode, but the UA string is sent to the server before IE has the page used to determine the desired render mode." If this is correct then can you explain what I seeing below? "GET css-class.com/cssscript/class-base2.css HTTP/1.1" 200 526 "http://css-class.com/articles/explorer/guillotine/index.htm" "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0; MS-RTC LM 8)" and at the end of this session for the same page we have this. "GET css-class.com/favicon.ico HTTP/1.1" 200 1406 "-" "Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0; MS-RTC LM 8)" I believe that these above strings would indicate that IE8 is being tested in IE7 mode and the UA string will show IE7 for all files accept when the favicon.ico is requested. I have previously asked in this thread if the following UA string was IE8 masquerading as IE7. Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.2; MS-RTC LM 8) and I believe I was correct in the first place. This will be corrected with the beta? Here a riddle: I know that you know but I know that you know, and if you know that I know then we all know that I know, but if you don't know that I know then you don't know what I know. ;-)

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  @EricLaw Opera seems to render both pages fine without having the "mozilla" token. And before you start harping on Opera's browser.js, there's no apparent patch for www.wellsfargo.com and the fix for yahoo.com seems to be based entirely on document.all sniffing. Your response is insulting. I asked for a concrete example using only the "mozilla" token of navigator.userAgent, and you brush me off until I start talking about losing face. And now you refuse to answer my questions about conditional comments, and placement of the meta tag. Would you like help finding your face?

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  @EricLaw MSFT Please confirm (yah, or nah) if IE8 will be available on Windows XP. Thanks.

• Anonymous
  February 23, 2008
  @Rex Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 5.1) Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 5.2) Indeed I have many of these in my log files.

• Anonymous
  February 23, 2008
  @Alan Gresley Thank you for your engaging and provocative response. It fills me with queasiness and unease, but that's hardly your fault.

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  "This is a very exclusive program, by invitation only." http://www.activewin.com/awin/comments.asp?HeadlineIndex=42767&Group=1

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  Dave: although my sites will be fine, this isn't about me, but about the web in general. It also makes it clearer that the infamous "Don't break the web" song played by Microsoft was just a bad excuse to hinder progress.

• Anonymous
  February 23, 2008
  @John Hawks I didn't make a "lazy error," I checked those two sites and couldn't find anything that didn't work with a browser that didn't have "mozilla" in its userAgent. And every time I try to point out the inconsistencies in what the IE team is doing, a team member tries to misdirect my comment. Not to "change the subject" on you, but I take exception to your name calling. I was not hurling insults at anyone, I was expressing consternation that the IE team makes claims but cannot prove them to me.

• Anonymous
  February 23, 2008
  Greetings from the Internet Explorer Team! We are nearing the launch of Windows Internet Explorer 8 Beta 1 and we will be making it available for the general public to download and test. IE8 Beta 1 is focused on the developer community, with the goal of gaining valuable feedback to improve Internet Explorer 8 during the development process. We have identified you as a qualified beta tester and we would like to offer an opportunity to join our limited technical beta program for Windows Internet Explorer 8 Beta 1. Participation in the Technical Beta will enable you to evaluate a common release of Windows Internet Explorer 8, the ability to submit feedback, post bug reports, download software answer surveys on product quality as well as vote on top bugs filed by others from the technical beta program. This is a very exclusive program, by invitation only. The only way to submit feedback is to enroll in the Windows Internet Explorer 8 Technical Beta program. As such, we would be happy to have your participation. To accept this invitation and to apply to become a member of this program, follow this link: removed for obvious reasons (If this link does not work for you, copy the full link and paste it into the Web browser address bar.) Follow the steps shown to you by that program to apply to become an active participant. You may be asked to take a survey, or submit other preliminary information. To report a problem or to ask a question, visit the Contact Us page (found at the bottom of every page). We hope to see you in the technical beta! Best regards. The Internet Explorer Team If you do not want to be invited to participate in other Microsoft Connect programs, change the contact preference you have set on your Microsoft Connect Profile (http://connect.microsoft.com/myparticipation.aspx). This is an unmonitored e-mail address, so do not reply to this message. http://channel9.msdn.com/ShowPost.aspx?PostID=385426

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  Maybe have popup when it detects the site using "legacy" codes: "This site is using outdated codes. Please help us berate the webmaster for not keeping up with the times." If Vista can break old apps, why not let IE8 break old webs?

• Anonymous
  February 23, 2008
  @Andrew "My solution is to have true standards mode enabled by default, with IE7 mode as an opt-out." The typical user will not understand an IE7 opt-out mode.  Almost every site they browse to either does not work, or displays strangely.  What do they do?  Complain bitterly and uninstall IE8.  Public relations disaster.  Then you have a very small proportion of users who understand the IE7 opt-out switch.  They switch to IE7 immediately because it's the only way they can get anything done.  What have you now?  The current situation with IE8, without the opt-in Meta tag. Now, if you want IE to die, this may well suit your position.  If not, how can you justify it? "Set your server to transmit the IE=edge HTTP header to ALL browsers forever. The userAgent has NOTHING to do with the opt-in." I'm beginning to think you ignore parts of my posts. Yes you can use client-side scripting and such to target different browsers.  Many solutions however are server based, so they will inspect the User Agent string to construct pages targeting different browsers.  In this case, once a site has upgraded their IE8 content to standards mode, they commence inserting the opt-in tag.  This is all based on the User Agent string as this is the only version information available to the server within the request. "But all the scripts that work around IE7's flaws will suddenly break if the browser uses IE7's rendering engine but identifies itself differently. This isn't backwards compatibility, and it contradicts the IE team's "don't break the web" mantra." Let's repeat the initial post for you: "Best Practices Websites that are sensitive to changes in the User-Agent string may need to be updated to support IE8.   During internal testing, we have encountered a small number of sites which fail to recognize IE8 because they are performing exact string matches to look for specific IE version strings" ... This may be a suprise but it's in MS's interest for things not to break, so this is likely to be the path of least pain based on their research.  Given your solution of standards mode by default, your path is one of most pain.

• Anonymous
  February 23, 2008
  @Gyrobo That last post was for you.  Not me ;)

• Anonymous
  February 23, 2008
  @Andrew I don't see anything wrong with being vocal.  The IE team regularly ignores people who comment on this blog and only respond when they are called out on something, several people complain about something, or a user posts mis-information that they can prove incorrect. If I were to post that IE doesn't follow RFC-2345 properly with their HTTP Requests (2345 made up btw), then I would expect a response. But if I ask: Will IE8 be available on XP? - I don't expect a response until after Mix08 (and to be honest I don't expect a yes) Will super-standards mode fix the DOM? - I don't expect an answer until after Mix08 (and I don't expect a yes) Will super-standards mode fix more CSS and rendering issues beyond ACID2 stuff? - I don't expect an answer until after Mix08 (and I don't expect a yes) Experience on this blog has taught me that when the IE Team does not respond, 3 things are known. 1.) The answer is NO. 2.) They will say nothing, because they believe that telling people will upset them, and they don't want angry developers. 3.) Although they haven't figured it out yet, not answering us is what makes us Angry. One final question for the IE Team. Is Al right? Is "Temporary" the new "Permanent"?  When will public bug tracking resume for IE, and can we expect it to be a stable part of the 2 way transparent developer communication with the IE Team in the future? Anwar

• Anonymous
  February 23, 2008
  Anwar, the problem is Microsoft, not specifically the 'IE Team'. Microsoft never talks openly about anything, whether it's IE or other software. The 'IE Team' isn't allowed to say anything unless the powers-that-be allow it.

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  A fine example: "Microsoft never talks openly about anything"

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  The comment has been removed

• Anonymous
  February 23, 2008
  > the intent with IE7 was to increase support for standards while minimising the impact. The intent of IE7's release in relation to HTML and CSS was to fix the most popular, annoying bugs. The level of new standards support was relatively low.

• Anonymous
  February 23, 2008
  @Andrew "Isn't this one of the main points of our discussion?  The only thing opt-in dissenters have achieved is a thorough job of re-iterating the same slogan, which is along these lines. Force every site to upgrade to support IE8 standards mode." No Andrew, us dissenters are re-iterating our displeasure with hearing the same old mantra again and again from Microsoft which is "don't break the web." No`one is forced to write standard complaint code, it's just that with the evolution of the web, some sites may break. It is called progress. I have page layouts or demos that will remain broken in IE8 because by some grand wisdom they are deemed to be non standard. "Which of course will not happen, because IE8 will be boycotted by users.  So yes, I'm asking you to justify; you seem to have a bit of time (as I do) so include the forceful arguments from said blogs." And your forced argument is to say that anyone with a different opinion to yours is wrong. Microsoft is forcing developers to add non standard code, so they can safe face. They are also forcing IE8 onto the world as a critical update. Why can't they let the Intranet use there IE5, IE6 or IE7. "I think you might have worded that incorrectly; the point of opt-in standards mode is to support standards, not backwards compatibility." No Andrew the first motivation of Microsoft is to support backwards compatibility (part of their income base) and secondly support standards. If it was otherwise, then it would be an opt-in for backwards compatibility. I would think that some IE team members would feel liberated to not have big brother watching over their shoulders. "Anyway it's called compromise, not hypocracy.  Must you take everything so literally?  Can you not see the intent?" Comprise indeed, read this Andrew. http://en.wikipedia.org/wiki/Compromise "compromise is a concept of finding agreement through communication" Then we have extremism. http://en.wikipedia.org/wiki/Extremism "describe the actions or ideologies of individuals or groups outside the perceived political center of a society; or otherwise claimed to violate common moral standards." So Andrew, consider deeply, is IE8 as standard compliant as would be indicated by Microsoft's' claim in wanting to support standards. By CSS standards IE8 is NOT standard compliant and in the future this may all come back to haunt us. In hindsight we may say, if only if we did it different in 2008.

• Anonymous
  February 24, 2008
  The comment has been removed

• Anonymous
  February 24, 2008
  The comment has been removed

• Anonymous
  February 24, 2008
  Estoy escribiendo este post desde el Aeropuerto de Guadalajara. Los sábados si que esta bueno el tráfico

• Anonymous
  February 24, 2008
  Andrew, If the IE team uses Connect for a bug database during the Beta (whenever that is), that will be an improvement but doesn't change the overall issue. There needs to be a way for developers to log real issues, which the IE team then triages and potentially responds to, at all times, not just during a beta. If they have a temporary bug database and then turn it off, I will understand why they do so but that doesn't really solve the problem of "How do you report a serious bug or issue to the IE team?" and "If you do report an issue, how do you know if it will ever be addressed?" These are real issues to people. There are developers that hold off on implementing work on their sites because they either think or hope that an issue currently blocking them with IE use will be fixed in the future. If they knew that an issue was fixed and that it would be in a future patch, they would have a lot of relief, wouldn't they? Since you seem to have come out of nowhere recently, how about providing us with either your last name or a link to a site when you post? Otherwise, you kind of look like a troll yourself...

• Anonymous
  February 24, 2008
  Also, as to what other people have been saying here, no one outside of Microsoft really knows what IE8 will have changed other than the Acid2 mention and the new switch for different modes. None of the other feature work has been discussed, nor has any additional standards compliance. Andrew can say that it is company policy but people want to know what is coming or to have a roadmap before continuing to focus development on IE. I also, wonder, as well if IE8 will ship on XP. Somehow, I doubt if Microsoft has a choice with the lack of Vista uptake though.

• Anonymous
  February 24, 2008
  The comment has been removed

• Anonymous
  February 24, 2008
  Last time I read, Google estimated they we indexing 4+ billion pages, and that they were indexing 10% of the web. Even if you only break a small %age of them, that's a large number of pages that need to be fixed. Not breaking the web does have some merits worth considering.

• Anonymous
  February 24, 2008
  @Fred I like your rationale there, very nice.

• Anonymous
  February 24, 2008
  @Al Billings "Also, as to what other people have been saying here, no one outside of Microsoft really knows what IE8 will have changed other than the Acid2 mention and the new switch for different modes." I don't know javascript or the DOM but CCS is quite useful for testing browsers for CSS standards support. That is why I know that IE8 in standard mode (one build) will use conditional comments meant for any version greater than IE7. At least IE8 doing some things correctly. You know my name and my site so if you have any queries in what I know, please contact me. What is of great importance are the questions I keep asking IE team members again and again that are not answered. In saying this Eric Lawrence may not himself be able to answer my questions.

• Anonymous
  February 24, 2008
  @EricLaw You are using the real IE7, so if it was you who visited my site that day here is one log: [22/Feb/2008:11:07:30 -0700] "GET css-class.com/favicon.ico HTTP/1.1" 200 1406 "-" "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0; SLCC1; .NET CLR 2.0.50727; .NET CLR 3.0.04506; .NET CLR 1.1.4322; InfoPath.2; MS-RTC LM 8)" [22/Feb/2008:11:07:30 -0700] "GET css-class.com/ HTTP/1.1" 200 4092 "http://blogs.msdn.com/ie/archive/2008/02/21/the-internet-explorer-8-user-agent-string.aspx" "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0; SLCC1; .NET CLR 2.0.50727; .NET CLR 3.0.04506; .NET CLR 1.1.4322; InfoPath.2; MS-RTC LM 8)" The real IE7 will always request the favicon.ico at the start of as session Another IE team member or yourself: [23/Feb/2008:16:18:46 -0700] "GET css-class.com/ HTTP/1.1" 200 4092 "http://blogs.msdn.com/ie/archive/2008/02/21/the-internet-explorer-8-user-agent-string.aspx" "Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0; SLCC1; .NET CLR 2.0.50727; .NET CLR 3.0.04506; Tablet PC 2.0; .NET CLR 1.1.4322; InfoPath.2; MS-RTC LM 8)" String of request in between [23/Feb/2008:16:18:47 -0700] "GET css-class.com/favicon.ico HTTP/1.1" 200 1406 "-" "Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0; SLCC1; .NET CLR 2.0.50727; .NET CLR 3.0.04506; Tablet PC 2.0; .NET CLR 1.1.4322; InfoPath.2; MS-RTC LM 8)" The real IE8 will always request the favicon.ico at the end of as session so this is why I have asked this question. http://blogs.msdn.com/ie/archive/2008/02/21/the-internet-explorer-8-user-agent-string.aspx#7860099 Will this be corrected with the beta release or don’t you know the answer? Let’s say if IE8 arrived at a page on my site by following a link from another site. What confuses me is what is going to happen next depending of if the page is meant to default to IE7 mode or use this new meta to be in IE8 deluxe mode. Some severs will respond to a request by what is contained in the user string. Your initial answer was: “@alan: The User-Agent string isn't render mode-specific, as the page gets to determine its own render mode, but the UA string is sent to the server before IE has the page used to determine the desired render mode.” So IE8 gobbles up whatever is on my server and either determines either from my server or from an inserted meta tag in the source which mode to use. But what is going to happen with these conditional comments like: if IE if gte IE 7 Included within these conditional comments are scripts and css or linked files for these with rules meant for a typical IE. Two years ago there was the infamous “Call to Action”: http://blogs.msdn.com/ie/archive/2005/10/12/480242.aspx This means that it may be very simple to include this new meta in the page header by a script on the server, but then for any developer who has used such conditional comments, they will now have to change the conditional comments on each and every page they appear on. None of this would matter if IE8 was by default in standard mode and didn’t use conditional comments. All the solutions that the IE team has dreamed up over the years just to keep on supporting backward compatibility has now created this current mess and indeed the IE team members do not know all the answers for all the questions that we ask. Please Microsoft, break the web, you are expecting too much from the IE team members. Supporting backward compatibility just causes a continuous spiral into the abyss which no IE team member can possibly climb out of.

• Anonymous
  February 24, 2008
  The comment has been removed

• Anonymous
  February 24, 2008
  How is IE8 going to handle XSLT PI on XML where the XSLT now uses features of XSLT that require permissions to be set to true such as scripting, document fucntion, include, import? My understanding is these are set to false by default in IE8 and there does not seem to be any mechanism for turning them on inside XSLT or as part of the PI.

• Anonymous
  February 24, 2008
  How is IE8 going to handle XSLT PI on XML where the XSLT now uses features of XSLT that require permissions to be set to true such as scripting, document fucntion, include, import? My understanding is these are set to false by default in IE8 and there does not seem to be any mechanism for turning them on inside XSLT or as part of the PI.

• Anonymous
  February 24, 2008
  @bleh "Furthermore, you are LYING about all the arguments that have brought up against opting out of standards mode. That, or you are ignorant." I don't remember saying anything about an opt-out mode.  I was talking about all the arguments (of which there seem to be few) against an opt-in mode.  Forcing every site to opt-out is unrealistic, nobody likes being told what to do by MS or any company for that matter.  It costs money.  People just want things to keep working.

• Anonymous
  February 24, 2008
  The comment has been removed

• Anonymous
  February 24, 2008
  @Al Billings "If they knew that an issue was fixed and that it would be in a future patch, they would have a lot of relief, wouldn't they?" I agree with you.  I was arguing about the effectiveness of the system due to those wishing to disrupt it.  If we lived in a world of flowers and roses, the system would no doubt be in-place. "Since you seem to have come out of nowhere recently, how about providing us with either your last name or a link to a site when you post? Otherwise, you kind of look like a troll yourself..." My name is Andrew Hilton.  I don't have a URL.  I have no affiliation with Microsoft. It's interesting to note that you no longer appear to have an opinion on standards opt-in.  Perhaps Mozilla will adopt the same approach. "Andrew can say that it is company policy but people want to know what is coming or to have a roadmap before continuing to focus development on IE." How can a roadmap be relied upon if it cannot be guaranteed?  People now know that standards support is being worked on.  if it's good, we get closer to writing a purer set of code to multiple browsers.

• Anonymous
  February 25, 2008
  @Alan Gresley "What is of great importance are the questions I keep asking IE team members again and again that are not answered. In saying this Eric Lawrence may not himself be able to answer my questions." Vague questions about pre-alpha versions of IE8 accessing your site may be exciting to you, but no doubt Eric has more important work to do. "if gte IE 7" So you have to upgrade your site in order to support IE enhanced standards mode.  Isn't that the point of opt-in?  To give you time to upgrade? "None of this would matter if IE8 was by default in standard mode and didn’t use conditional comments." How would you support older versions of IE?  

• Anonymous
  February 25, 2008
  @Alan Gresley "Comprise indeed, read this Andrew. http://en.wikipedia.org/wiki/Compromise" I wasn't referring to compromise within the context of an argument.   "something intermediate between different things" (from dictionary.com) As in - "the information found in wikipedia is a compromise between convenience and accuracy".

• Anonymous
  February 25, 2008
  The comment has been removed

• Anonymous
  February 25, 2008
  "So you have to upgrade your site in order to support IE enhanced standards mode.  Isn't that the point of opt-in?  To give you time to upgrade?" Visit this page in IE6 or IE7 http://www.stopdesign.com/portfolio/promotional/hotbot_mousepad.html Are you going to suggest to Douglas Bowman that he needs to now include a new meta in the header and use conditional comments for IE6 and IE7 so IE8 doesn’t show the Guillotine bug on this page and other pages in this section of his site. If the default for IE8 was standard mode then the bug would not be present and all Douglas has to do is use one of the three fixes on my site for IE6 and IE7. All achieved with one CSS declaration.

• Anonymous
  February 25, 2008
  The comment has been removed

• Anonymous
  February 25, 2008
  The comment has been removed

• Anonymous
  February 25, 2008
  I hate IE7. Thought you might like to know that

• Anonymous
  February 25, 2008
  @Andrew As much as I'd like to continue our discussion, it's becoming untenable. I agree with Alan Gresley; every time I bring something up that is worthy of discussion, you act like having more than one conversation at a time is impossible and that we should only proceed on the track that benefits your argument the most. On that note, I'd like to briefly expound on my earlier comment on including IE7 in XP SP3, and why its exclusion proves why the opt-in tag is fallacious: If a web site currently WORKS in IE7, then it's almost certain that it is being maintained by someone who can insert an opt-out tag to maintain backwards compatibility in IE8. If a web site BREAKS in IE7, and STILL has not been fixed over a year after IE7's release, then it is PERMANENTLY BROKEN in Internet Explorer now and for all time. No opt-in can change that. By excluding IE7 from the service pack, Microsoft is implicitly admitting that even IF true standards mode is opt-in, all the pages that would have broken by such a move WILL be broken anyway.

• Anonymous
  February 25, 2008

• When I said "by such a move," I was referring to enabling true standards mode by default.

• Anonymous
  February 25, 2008
  The comment has been removed

• Anonymous
  February 25, 2008
  @Alan Gresley: There are many possible explanations for what you're seeing in the server log.  Using Fiddler or another tool, a prankster could send anything he wants as the User-Agent string. IE8 sends the IE8 UA string for ALL requests (including favicons), unless the aforementioned "Report IE7 User-Agent String" menu option is checked. While I did not discuss the Version Vector behavior in this post, rest assured that both "if IE" and "if gte IE 7" will continue to work.

• Anonymous
  February 25, 2008
  Awesome stuff! Keep up the great work guys! IE 7 is awesome.. IE 8 is gonna be super awesome.. thanks ;)

• Anonymous
  February 25, 2008
  The comment has been removed

• Anonymous
  February 25, 2008
  Any plans to use the correct English spelling of "favourites" in IE8?  The American spelling is very annoying!

• Anonymous
  February 25, 2008
  http://blogs.msdn.com/ie/archive/2008/02/21/the-internet-explorer-8-user-agent-string.aspx Yes yes I

• Anonymous
  February 25, 2008
  The comment has been removed

• Anonymous
  February 25, 2008
  The comment has been removed

• Anonymous
  February 25, 2008
  Wake me up when you support SVG.

• Anonymous
  February 25, 2008
  @Andrew "So your argument is that you don't believe the web will break." Oh, they will. Merely changing the user-agent string will break sites. So by your logic Microsoft can't even change that string to say "IE8". "Forcing every site to opt-out is unrealistic" They could opt-in to IE7 mode. By the way, if they don't opt-in to IE8 mode they are basically getting the IE7 engine. So why does it identify itself as IE8 when it's basically IE7 (mode)?

• Anonymous
  February 25, 2008
  Microsoft has a chance to redeem themselves. do what is right this time guys and stop crome plating turd! its the age old saying, beauty is skin deep!

• Anonymous
  February 25, 2008
  Com'era stato promesso lo scorso dicembre, entro la fine di questo trimestre Microsoft prevede di rilasciare al pubblico la prima versione beta di Internet Explorer 8. A confermare questa tabella di marcia è una email, inviata da Microsoft ad u

• Anonymous
  February 25, 2008
  The comment has been removed

• Anonymous
  February 25, 2008
  @EricLaw Could you kindly explain why IE7 won't be included with XP SP3? If IE7 causes so many problems that it can't be pushed on users, how will taking IE7 and changing its userAgent make IE8 more compatible? The pages that break now under IE7 would remain broken in IE8 whether or not true standards were the default. Isn't the REAL compatibility solution under an opt-in system to make IE6 the default rendering engine?

• Anonymous
  February 25, 2008
  IE7 will run just fine on XP SP3 when that service pack becomes available. Generally, it is our policy not to require new browser versions within a service pack.   For myriad reasons, some customers prefer to deploy new browsers on their own schedule, and sometimes, not alongside OS service packs.

• Anonymous
  February 25, 2008
  @Steve: Stay tuned to the IE blog for the latest news about the upcoming beta.  At the moment, there's no "link" to send.

• Anonymous
  February 25, 2008
  The comment has been removed

• Anonymous
  February 25, 2008
  Gyrobo, Opera will say literally anything to get press coverage.  (Reference their executive's claim that he was going to swim the Atlantic a year or two ago).  Remember, they have FAR less share than even Firefox, despite giving away their product for free. There's nothing at all in anti-trust law (or any other law) to suggest that Microsoft must require that users upgrade their browsers.  If users want new browsers, they can simply install them. As for the "meta" thing, you are making Microsoft's case for them!! Websites WILL break if they default to the newer engine.  Hence, the default is the OLD engine, and sites can "catch up at their own pace" by opt-ing in with the new META tag.

• Anonymous
  February 25, 2008
  EricLaw: Can't MS just include an option to install IE7 upon the SP3 setup? For example, they'll be a screen saying "Your Internet Explorer is out of date!" and will ask you if you want to upgrade IE during the installation with the default option set to no.

• Anonymous
  February 25, 2008
  @@sonicetc-- if someone is going to bother installing xpsp3 (big), they probably already have ie7 (small)... Since service packs go out on WindowsUpdate, the user would already see the IE7 update on Windows update.  And since Microsoft has ie on the autoupdates, someone on xpsp3 with ie6 will get asked to install ie7 anyways.

• Anonymous
  February 25, 2008
  When released, the IE8 beta will introduce an updated User-Agent string. For IE8, we’ve simply replaced ’MSIE 7.0’ with ’MSIE 8.0’. For example, on Windows Vista, IE8 sends: Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0) IEBlog : The Internet Explor

• Anonymous
  February 25, 2008
  @Stanley "As for the "meta" thing, you are making Microsoft's case for them!!" I am NOT making Microsoft's case for them, I was trying to follow to its conclusion what I saw as a contradiction in Eric's logic. He says that Microsoft can't afford to enable true standards mode by default, because that would break web pages. Then he says that Microsoft won't push an updated web browser because it might break web pages, a web browser that renders pages allegedly the same as the older version by default. The argument is strange to me: if a web browser is going to render the same (by default) as an older version, why not push the new version? This is especially pertinent on Vista. Will Vista SP2+ include IE8? If Microsoft goes through with this meta scheme, then the default rendering mode will be the same as IE7. "Websites WILL break if they default to the newer engine." Haven't I already pointed out at that by changing the userAgent, IE8 WILL break sites anyway?

• Anonymous
  February 25, 2008
  @Gyrobo: On the contrary, I have never said that Microsoft is not requiring customers deploy IE7 "because it might break web pages."   Customers elect to deploy browsers on the schedule of their choosing.

• Anonymous
  February 25, 2008
  @whut Your questions were answered previously. @Alan Gresley "It applies to a particular build of IE8 in standard mode. This build still has the some of the same CSS bugs as IE7." They haven't finished it yet. @Gyrobo "As much as I'd like to continue our discussion, it's becoming untenable." I'm sorry you feel that way. However you don't attempt to read and understand IE blog posts or replies from EL, so it's unlikely you'll understand me (I'm not nearly as eloquent). I want to see MS support web standards to the best of their ability.  I find it exciting that the op-in method frees them from the shackles of backwards compatibility.  It offers a chance to fully support the latest HTML/CSS/DOM standards.  The result of that freedom would be a rendering mode vastly different to IE7 and below, and possibly more compliant than competing browsers (who have their own bc issues). Forcing MS into standards mode by default would likely mean a watering down of IE8's standardisation effort, from fear of breaking existing sites. I'm sure you guys want to see IE more standards compliant, a complete break from the past is the only way forward. My time is limited so I apologise if you don't get replies in future.

• Anonymous
  February 25, 2008
  While the mozilla team are releasing betas of a really impressive browser to the public, you are revealing your new User Agent String. Really exciting...

• Anonymous
  February 25, 2008
  I beg to differ: there were differences between 'original' Internet Explorer 6 and IE6 for Windows XP SP2 (I want for proof that I once had to add fixes to handle IE6sp1 for Windows 2000, fixes that weren't needed for IE6sp2). So, not including a new IE version in XP sp3 because MS doesn't include new browser versions in their SPs doesn't hold much water. Still, I'd personally appreciate something else a lot more: porting IE8's engine to Windows 2000, so that everybody could get rid of IE6 - and not wait for 07/2010. Other than that, I'd really enjoy support for CSS2/3 advanced selectors. One question: what will happen when a page, using selectors like :before and :after are opened without the compatibility shim - sorry, the 'UI version indicator' in IE 8? Personally, I dynamically add empty elements under IE and give them the same style I give the correct elements, but will IE 8 pop me both out? Should I use a browser detect to tell IE 7 (and masquerading IE 8) not to display unsupported IE 7 extensions?

• Anonymous
  February 26, 2008
  @EricLaw You wrote: "@Alan Gresley: There are many possible explanations for what you're seeing in the server log.  Using Fiddler or another tool, a prankster could send anything he wants as the User-Agent string. IE8 sends the IE8 UA string for ALL requests (including favicons), unless the aforementioned "Report IE7 User-Agent String" menu option is checked. While I did not discuss the Version Vector behavior in this post, rest assured that both "if IE" and "if gte IE 7" will continue to work." My reply: So when the "Report IE7 User-Agent String" menu option is checked in IE8, this allows for testing IE8 in IE7 mode and all files requested will use the IE7 UA string accept the favicon.ico. This is a fact. Concerning the requesting of favicon.ico., this is the first true authenticated IE8 bugs. http://www.gtalbot.org/BrowserBugsSection/MSIE8Bugs/ "Rest assured that both "if IE" and "if gte IE 7" will continue to work." So any script or style or linked script or style within them which is meant for a buggy IE or a version equaling or greater than IE7 (remembering that we could only assume that IE would always remain buggy) will be used also by IE8 in standard mode. Is this correct? Isn't that IE bug behavior targeting the edge? I really don’t understand the true implications of this from a scripting perspective, but I do understand in simple terms that if there was some improvement in the DOM in IE8, then these scripts within these conditional comment would have some impact. Yes indeed a prankster is trying to fool me. I added a new test page on my site on the 13th of February and this prankster with amazing insight visits this page on the first day of it being live. This prankster is very clever because they knew well enough to check out the source of this page and discover all my reverse testing methods for IE8 which I have hidden in the code. So very quickly they rebuild Fiddler or another tool to fool me. [13/Feb/2008:02:00:04 -0700] "GET css-class.com/images/examples/t10.png HTTP/1.1" 200 180 "" "Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0; MS-RTC LM 8)" This prankster is really good too; they managed to build a UA that also has the ability to understand conditional comments which are only meant to target any version greater than IE7. The image (t10.png) is requested this conditional comment. if gt IE 7  img src="t10.png"  endif And for the real dinki-di IE7 [14/Feb/2008:11:08:49 -0700] "GET css-class.com/images/examples/t9.png HTTP/1.1" 200 180 "" "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0)" The image (t9.png) is requested by a conditional comment. if IE 7  img src="t9.png"  endif

• Anonymous
  February 26, 2008
  The comment has been removed

• Anonymous
  February 26, 2008
  @Andrew "It applies to a particular build of IE8 in standard mode. This build still has the some of the same CSS bugs as IE7." "They haven't finished it yet." Andrew I understand that you are for this opt-in to standard approach and I respect you for having a difference in opinion, but I don’t think you have considered fully the consequences of such an approach. The approach should be to put all the dramas of IE whatever behind us forever and move forward. IE7 showed a big improvement in standards support but some of these new features have there own bugs. One is with the dynamic pseudo class (:hover) on any element. Another one is with sibling selectors. The former bug with E:hover I am presently helping a IE team member work with. Hopefully IE8 in standard mode doesn’t have this bug. http://css-class.com/test/bugs/ie/recalculatedoffsetbug.htm http://css-class.com/articles/explorer/sticky/index.htm And doing a search with IE7, sticky and hover will show that this bug is epidemic. http://www.google.com/search?hl=en&q=ie7+sticky+hover&btnG=Search Microsoft is releasing IE8 to the world in a few weeks. A bug with sibling selectors is still present in IE8 in standard mode. H1 + * ~ p {} /* style for the 2nd paragraph onwards*/ This sibling selector string will select different elements in the source in IE7 and IE8 if ordinary html comments appear between these elements. This is non-standard complaint behavior because ordinary html comments are not elements. If I had a comment after the h1 header then the 1st paragraph onwards will all have the same style. To get around this you have to give IE7 or IE8 a whole new sibling selector. H1 + * + * ~ p {} /* style for the 2nd paragraph onwards in IE7 and IE8 / Then you have to overrule the sibling selector that you gave the good browsers in the first place. H1 + * + p {} / style for the 1st paragraph in IE7 and IE8 */ Please Andrew, explain to me the workarounds and where do all these go? Please don’t suggest to me that my universal selector could be replaced with a "p" for a paragraph because the universal selector could be selecting a list or blockquote instead. So can we all (the web development community) separate fact from fiction and consider the consequences across the whole spectrum of what this meta, UA string, and IE bug behavior will bring. Is this a true break from the past or is this just history repeating itself?

• Anonymous
  February 26, 2008
  @EricLaw As Mitch074 pointed out, XP SP2 included a version of IE6 that contained new functionality not present in the original IE6. Is this not tantemount to a new browser? "Customers elect to deploy browsers on the schedule of their choosing." This seems like a good reason to enable true standards by default. If a site "breaks" in IE8 then customers can schedule to install IE8 after the page is fixed. Side note: I don't know if you know this, but in XP the Active Desktop Recovery file seems to be broken with IE7. Will a fix for this be in SP3?

• Anonymous
  February 26, 2008
  Alan, I'm not sure where you get your information, seeing as how IE8 hasn't even shipped in a BETA form yet. <<"all files requested will use the IE7 UA string accept the favicon.ico. This is a fact">> I assume you mean "except" rather than "accept"?   I can tell you that your "fact" is incorrect, as I watch my network traffic constantly, and IE8 builds do not behave as you describe.  My original statement correctly describes the "Report as IE7" feature. The "bug" from Talbot's site is a by-design behavior, and has nothing to do with the User-Agent string.  As IE invented that FavIcon feature, it's ironic that the behavior is now being called a "bug" after everyone else adopted the feature.  Talbot fails to note that a web developer can avoid the query to the root by specifying the FavIcon via META tag, a feature we documented when we invented FavIcon support. There's no inherent linking between the UserAgent string and conditional comments; they are controlled from different sources.  The Conditional Comments version comes from the Version Vector registry key, while the UA string is in the code but can be overriden by other registry keys. Registry scripts from my site (http://www.enhanceie.com/useragent.aspx) have offered the ability to override both the UserAgent and the Version Vector for several years now. I'm not sure what you mean when you say "rebuild Fiddler"?  Fiddler enables user-agent string spoofing via either a menu option or checkbox; you don't need to "rebuild" anything.

• Anonymous
  February 26, 2008
  @Gyrobo: It's hard to imagine what your definition of "tantamount to a new browser" is.  There's no question that IE6 as shipped on XPSP2 included new features, pretty much exclusively focused on security.  It did not include significant UI changes (many business customers wait to deploy UI updates for support reasons).  It did not include changes to the rendering engine, and no changes to HTML support or DOM support.  Rather than including a new User-Agent Version or Version Vector Version, only a token was added to the UA. <<This seems like a good reason to enable true standards by default. If a site "breaks" in IE8 then customers can schedule to install IE8 after the page is fixed.>> That's a very risky game that has a chicken and egg problem.   Businesses (and lots of consumers) wouldn't install IE8 until all sites are updated, and sites wouldn't bother to update because no one is using IE8. Hence, web developers would be stuck supporting IE7 for YEARS longer than they will be under a compat-friendly strategy.

• Anonymous
  February 26, 2008
  . . Wow, Stanley, what a pathetic thing to say: "Opera will say literally anything to get press coverage.  (Reference their executive's claim that he was going to swim the Atlantic a year or two ago)." It's called a PR stunt. Of which Microsoft has had many. Remember the video when Bill Gates stepped down? Did you REALLY think that Opera's CEO was going to swim? You are more gullible than one could ever imagine, but maybe that's the norm for Microsoft employees? You are shooting yourself in the foot by criticizing Opera of PR stunts while your own employer uses them! . .

• Anonymous
  February 26, 2008
  @EricLaw <<"all files requested will use the IE7 UA string accept the favicon.ico. This is a fact">> "I assume you mean "except" rather than "accept?"" Thank you for correcting my spelling :-) "?Alan, I'm not sure where you get your information, seeing as how IE8 hasn't even shipped in a BETA form yet." "I can tell you that your "fact" is incorrect, as I watch my network traffic constantly, and IE8 builds do not behave as you describe.  My original statement correctly describes the "Report as IE7" feature." Can some in the IE team set Eric straight on this? "The "bug" from Talbot's site is a by-design behavior" "Talbot fails to note that a web developer can avoid the query to the root by specifying the FavIcon via META tag" A meta is needed to opt-out of automatically requesting favicon.ico when they may not be present in the root directory. How about that, we need another meta tag. "There's no inherent linking between the UserAgent string and conditional comments; they are controlled from different sources.  The Conditional Comments version comes from the Version Vector registry key, while the UA string is in the code but can be overriden by other registry keys." Can someone from the IE team please explain to Eric what I am talking about here? My question is will IE8 in standard mode still use IE buggy rules inside conditional comments like [if IE]? Eric, I do understand that they are not linked, they operate differently as do the IE7 UserAgent strings and conditional comments for IE7. "I'm not sure what you mean when you say "rebuild Fiddler"?  Fiddler enables user-agent string spoofing via either a menu option or checkbox; you don't need to "rebuild" anything." The latest hit. 131.107.0.102 - - [25/Feb/2008:22:28:57 -0700] "GET css-class.com/articles/explorer/sticky/index.htm HTTP/1.1" 200 18171 "-" "Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0; WOW64; SLCC1; .NET CLR 2.0.50727; .NET CLR 3.0.04506; .NET CLR 1.1.4322; .NET CLR 3.5.21022)" I have approximate 500 hits from IP addresses beginning with 131.107 which all trace back to Redmond and the other approximate 400 hits from IP addresses where the user is using IE8. I make a conclusion from my stats that there are bugs in IE8 in "standard" mode. Why are some pages visited again and again, over and over? Eric, test this page in IE7, it will explain to you the cause of the escaping floats and pee-ka-boo bugs in IE. http://css-class.com/test/bugs/ie/renderingbands.htm Eric, here is a file on my site to make you reconsider what you have been arguing about. http://css-class.com/cssscript/images.css I do feel empathy for any IE team member who has to report something about a new version of IE on this blog. Is it decided by the person who chooses the shortest straw?

• Anonymous
  February 26, 2008
  @pathetic: Heh... I don't work for Microsoft, and I often prefer to use Firefox.  Opera is a nice browser too, but it's a darn shame that Opera cannot compete on their merits instead of pulling PR stunts.  Poor Opera engineering team!  They write fine code, but their management makes them look absurd. PR stunts that use the legal system cost us all...

• Anonymous
  February 26, 2008
  Alan, posting on the IE blog is a privilege I enjoy; I just wish I had more time to do it.   Vis-a-vis the FavIcon issue: As noted, the IE team invented this feature ~long~ ago, and we support specification of the favicon via a meta-tag. All of the other browsers adopted this feature and the meta-tag.  The fact that IE built the feature originally to support a fallback icon (/favicon.ico) merely demonstrates that we want the webplatform to be easy to use, and we didn't want to force web developers to update every single one of their pages with a META tag specifying the desired icon.   Vis-a-vis the "Report IE7 User-Agent" option, I somehow doubt that anyone else on the IE team is going to try to tell me how my feature works.  :-)  If you examine your logs from this morning, you'll find that your site was hit by "MSIE 12.0" running on "Windows NT 9.1" ... You can either believe your web logs, or believe me that the User-Agent string can be spoofed. <<I make a conclusion from my stats that there are bugs in IE8 in "standard" mode.>> That's one possible conclusion.  The other possible conclusion is that the hits are not coming from IE at all, as noted previously.

• Anonymous
  February 26, 2008
  @EricLaw Thank you for your explanation. It's given me a great idea about how you could mollify me on the whole issue of changing the userAgent if you decide to stick to the opt-in: keep the IE7 string, but add a special heretofore unknown little flag token to designate it as IE8 for people looking for that. I'd still disagree about the concept of an opt-in, but at least you wouldn't seem so hypocritical. And I'm afraid I still don't understand why, if IE8 won't be pushed on Vista users in a later service pack. In comparison to IE7, it (under your opt-in scheme) has no "changes to the rendering engine, and no changes to HTML support or DOM support." "Businesses (and lots of consumers) wouldn't install IE8 until all sites are updated, and sites wouldn't bother to update because no one is using IE8." When Firefox was first released in 2004 it had 0% of the market, back when virtually all sites were tested only in IE6. The secret to driving adoption is having a good product. The yearlong (and continuing) communication blackout on IE8's feature-set has left me with sadly little to say about IE8 as a product. "[W]eb developers would be stuck supporting IE7 for YEARS longer than they will be under a compat-friendly strategy." People aren't buying new computers as rapidly as they used to. If you don't insist on encouraging your customers to migrate to newer browsers, web developers may have to code for a small but significant minority of IE6 users for the next QUARTER CENTURY.

• Anonymous
  February 26, 2008
  . Stanley, you are either a Microsoft employee or a troll. "it's a darn shame that Opera cannot compete on their merits instead of pulling PR stunts" What does "compete on their merits" mean? Are you so ignorant that you think marketing is irrelevant? Firefox is where it is today BECAUSE of marketing. BECAUSE of stunts like these. And you completely failed to address the fact that YOUR EPMPLOYER DOES THESE KINDS OF STUNTS TOO. "Poor Opera engineering team!  They write fine code, but their management makes them look absurd." How? Absurd like Bill Gates' "funny" video when he stepped down? Funny how you ignore the fact that all other companies pull these kinds of stunts as well. "PR stunts that use the legal system cost us all..." You mean like Microsoft's recent antitrust complaint against the Google/Doubleclick deal? Or Google's antitrust complaint against Microsoft before that? Get a clue. . .

• Anonymous
  February 26, 2008
  You better fix the idiotic problem of not being able to run multiple versions of IE on the same PC before releasing IE8.  Webdevelopers shouldn't have to hop around to different PC's just so they can test their sites on IE6, IE7, IE8 - fit it to where you can run them all together.  No more integregating IE into the operating system. Just make the stupid thing a standalone program that saves all its settings in files and not the registry, like Firefox.  Haven't you learned anything from Firefox yet?

• Anonymous
  February 26, 2008
  @Gyrobo: We spent a surprising amount of time agonizing over the user-agent string plan.   The idea to "add a special heretofore unknown little flag token to designate it as IE8" was considered, but ultimately put aside.  Such a strategy has a number of downsides, particularly in that it would require rewriting all of the web frameworks that exist for parsing user-agent strings.  Once the web frameworks are rewritten, we're right back where we started.

• Anonymous
  February 26, 2008
  @EricLaw And if Microsoft had had those types of discussions publicly and asked for feedback, I wouldn't have wasted your time reiterating it. But correct me if I'm wrong: by introducing a new UA string, all UA parsers need to be rewritten anyway, won't they? Forgive me, but I don't see a difference compatibility-wise, between adding an extraneous token and changing the version number, for UA parsers. Except that adding an extra token would avoid sites not correctly receiving data meant for the IE7 engine. If possible, could you make public a transcript of the talks that resulted in the UA version bump? If I could understand the viable options you chose from and why you deemed this the best approach I would be a happy camper.

• Anonymous
  February 26, 2008
  @Gyrobo: The user-agent string's !format! did not change, and thus parsers need not be rewritten.   The popular frameworks already correctly identify IE8, as does any code that follows our recommendation for UA parsing (http://msdn2.microsoft.com/en-us/library/ms537509.aspx#ParsingUA).

• Anonymous
  February 26, 2008
  @EricLaw The CURRENT frameworks may recognize IE8's UA (hard to believe considering you just formally announced the UA changes five days ago), but sites using older versions of those frameworks will inexorably break. Imagine the following scenario: if (navigator.userAgent.indexOf("MSIE 7") > -1) {  // Deliver crucial data that only IE7 needs. } Correctly identifying IE8 won't be of any use if there's no opt-in and that crucial data doesn't get through under IE8 legacy mode. The format of the UA may not have changed, but that single digit could mean the difference between a functioning site and catastrophic failure.

• Anonymous
  February 26, 2008
  @Gyrobo: A properly written framework immediately recognizes IE8 because the format hasn't changed.  For instance, see http://www.enhanceie.com/useragent.aspx, and you'll see that the version of ASP.NET shipped back in 2003 correctly identifies Name = IE Version = 8.0 Major Version = 8 Minor Version = 0 You're correct to note that a site hardcoded to target a ~specific~ IE version could be broken, which is only one of several reasons why we do not recommend such targeting. As noted in the Best Practices article, it is not recommended that you block access to content based on the user-agent string of the browser. If you do have to offer different content to different versions of the browser due to improved capabilities, you should ensure that future versions of the browser are not blocked. Serving content based solely on the user-agent string is often an unreliable way to detect the full capabilities of the browser, because the user might have adjusted some settings, such as disabling script or extensions.

• Anonymous
  February 27, 2008
  @EricLaw [MSFT] "As noted in the Best Practices article, it is not recommended that you block access to content based on the user-agent string of the browser." The Hotmail team doesn't follow your recommendation, might of fixed the "bug" already. http://www.freesoftwaremagazine.com/columns/hotmail_doesnt_work_with_firefox_2

• Anonymous
  February 27, 2008
  Don't forget: 1 - Standalone!!! No more poor/freak/.../ridiculous workarounds, please! 2 - If you want to keep conditional comments, be sure to implement a CSS version too...

• Anonymous
  February 27, 2008
  Is there any way I can sign up to be a beta tester?

• Anonymous
  February 27, 2008
  The comment has been removed

• Anonymous
  February 27, 2008
  @Will: This is an exciting time for the IE team and we can't wait to share IE8 with our customers.  Despite what you may have heard, the IE8 Beta is not public at this time.   Stay tuned!

• Anonymous
  February 27, 2008
  The comment has been removed

• Anonymous
  February 27, 2008
  "A properly written framework immediately recognizes IE8" I find this comment by EricLaw to be hilarious. If history has taught us something, in large part thanks to Microsoft's browsers, things are often NOT properly written! If they were, then IE8 would have been able to default to standards mode.

• Anonymous
  February 27, 2008
  The comment has been removed

• Anonymous
  February 27, 2008
  @EricLaw "A properly written framework immediately recognizes IE8 because the format hasn't changed." Eric, my issue is not that IE8 won't be identified correctly. It's that the IE7 legacy mode won't be identified correctly. "[T]he user might have adjusted some settings, such as disabling script or extensions." I'm talking about the things users CAN'T change, like CSS and JavaScript bugs that apply to all instances of a specific browser version. "Serving content based solely on the user-agent string is often an unreliable way to detect the full capabilities of the browser" Yes, but the developers your opt-in mode is designed to protect historically HAVE done things like this. When -- not if, but when -- a version of IE8 that (by default) renders as IE7 but identifies itself as IE8, when that browser hits the market, such web sites will break. It will LITERALLY be another IE7. P.S. Are you at least considering having a special beta version of IE8 where true standards mode is the default, for comparative purposes?

• Anonymous
  February 27, 2008
  @Eric Lawrence, I just sent you an email regarding your IE8 user agent string posted message. Once IE 8 fixes its CSS bugs - all of the CSS bugs -, then it will be time to recommend to avoid user-agent string (exact or not) detection and to propose, to suggest to go with object/method support detection instead. There is nothing better (overall), more reliable, easier to manage than object/method support detection. Regards, Gérard

• Anonymous
  February 27, 2008
  @Eric > The "bug" from Talbot's site is a by-design behavior, and has nothing to do with the User-Agent string.  As IE invented that FavIcon feature, it's ironic that the behavior is now being called a "bug" after everyone else adopted the feature.  Talbot fails to note that a web developer can avoid the query to the root by specifying the FavIcon via META tag, a feature we documented when we invented FavIcon support. Eric, IE 8 requests the favicon.ico regardless if the webpage author has declaratively linked it or not: that's one bug. I didn't write W3C Quality Assurance tip, How to Add a Favicon to your Site www.w3.org/2005/10/howto-favicon I didn't write www.w3.org/TR/webarch/#uri-opacity Other browsers support more than just the .ico file but .pgn or .gif too. What Jeff Davis {The first is to put a file in the root of your domain called favicon.ico.  The second is to use a <link> tag with the rel="shortcut icon" value 2007/03/01/why-doesn-t-the-favicon-for-my-site-appear-in-ie7.aspx} recommend is not web standards compliant. rel="shortcut icon" is not conformant (although valid) to HTML 4.01 spec. Therefore, Microsoft and IE dev. team has to proceed from this and assume the issues accordingly, from a web standards perspective. Regards, Gérard

• Anonymous
  February 27, 2008
  The standards bodies came along years AFTER the Internet Explorer team created the FavIcon feature, AFTER it was broadly used by lots of sites, and THEN wrote a new specification for how THEY thought it OUGHT to work.  Of course, this was different than the actual implementations in the REAL world. It's not clear WHY they felt that doing this was a good use of their time, but hey, who is anyone to question the STANDARDS??!?  Not you or I, for sure.  :-P

• Anonymous
  February 27, 2008
  This seemed a fitting place to make this wonderful announcement. See the list w/ babyfaced Mr. Gates on it here: antiaging4geeks.com Just scroll down the page a bit. Peter J. Lupo Esq. antiaging4geeks.com

• Anonymous
  February 27, 2008
  The comment has been removed

• Anonymous
  February 27, 2008
  On the IE blog , Eric Lawrence presented the User-Agent string for the beta version of Internet Explorer

• Anonymous
  February 28, 2008
  The comment has been removed

• Anonymous
  February 28, 2008
  The excuse for a meta tag enabling the super standard mode is not breaking the web. Yet the new user agent string will still break some sites. So you have the worst of two worlds: angry developpers as shown by numerous comments on this blog AND some broken sites The obvious solution to overcome these two problems simultaneouly is to link the user agent string send by IE 8 to its rendering mode:

• if UA is "MSIE 7.0" then rendering mode is classic IE 7 => no broken site
• if UA is "MSIE 8.0" then rendering mode is super standard => no meta tag There still is the debate on which rendering mode is the defaut but with less heat as:
• current mode is controlled from the client (and not from the site)
• the option to switch between the two modes should be immediately accessible Personnally, as a web surfer, I would use IE 8.0 mode and whenever I reach a broken site I would temporarily switch to IE 7.0 mode. Thierry

• Anonymous
  February 28, 2008
  You've just ADMITTED that even WITH super standards mode as an opt-in, IE8 will STILL cause site breakage. The question remains: are you incompetent, or do you just hate developers?

• Anonymous
  February 28, 2008
  @EricLaw "Alan, posting on the IE blog is a privilege I enjoy; I just wish I had more time to do it." Yes, my responsibility with family members also makes it hard to spend more time at my favorite passion CSS. "Vis-a-vis the FavIcon issue" Fair enough but if I had one site in the root directory with a favicon and another small site in a higher directory. This site is also going to show the favicon but this is of small matter to me so no more said. "Vis-a-vis the "Report IE7 User-Agent" option, I somehow doubt that anyone else on the IE team is going to try to tell me how my feature works.  :-)  If you examine your logs from this morning, you'll find that your site was hit by "MSIE 12.0" running on "Windows NT 9.1"" Wow..,  are they in alpha already, the UA an OS. :-) From today I have stopped my reverse testing for IE8. This file is not linked to from anywhere on my site. http://css-class.com/cssscript/images.css BTW, the world paid a visit yesterday with that nested IE6 UA string. I though it had been very quiet day until I search for the nested string. The only mistake (or was that on purpose) was to leave the nested string attached for a similar UA string but with a different IP address which allowed me to linked to many different IP addresses over a 2 hour period. Now that you have had the guided tour of my site (if you were present) you may now know what my site is all about and know other browser are far more standard compliant than IE7 and that IE7 is far superior to IE6. The removal of hasLayout from IE8 is welcoming. I was hoping for the best for IE8 but now I have concerns. "... You can either believe your web logs, or believe me that the User-Agent string can be spoofed." Of course they can be spoofed, I never said that they couldn’t, but only IE8 will download a file that is within conditional comments targeting greater than IE7. Can this be spoofed? Anyhow, I know by other means when IE team members have visited with IE8 which I will not declare publicly. <<I make a conclusion from my stats that there are bugs in IE8 in "standard" mode.>> "That's one possible conclusion.  The other possible conclusion is that the hits are not coming from IE at all, as noted previously." Since we can only dispels fact from rumor with the real McCoy, I will be welcoming IE8 beta testing soon.

• Anonymous
  February 29, 2008
  @Webdesign It's considered customary to put that kind of thing in quotes and post a response. Unless you felt I captured your reaction to this post perfectly and just wanted it said again.

• Anonymous
  March 01, 2008
  "If you do have to offer different content to different versions of the browser due to improved capabilities, you should ensure that future versions of the browser are not blocked." Gotta love that one after the whole v.targeting fiasco... IE team seems to be schizophrenic...

• Anonymous
  March 03, 2008
  Hey guys, keep up the good work on IE. Hopefully you can manage to have a good conversation with reasonable developers while ignoring those folks who have their whole self image tied up in hating Microsoft and all things that come from it.

• Anonymous
  March 04, 2008
  With Neelie Kroes looking over their shoulders and the entreaties of Web developers ringing in their ears Microsoft is promising that the default mode for Internet Explorer 8 will follow the latest Web standards: Consistent with its efforts to promot..

• Anonymous
  March 08, 2008
  You tout a beta browser for web developers, but it seems you've completely missed the mark completely on what it means to release a new browser for developers. IE should NOT be tied to the OS and should easily run as a stand-alone against IE7 and others, it would help it stand apart. Isn't that what you're going for? Why make this so difficult? You're Microsoft, ruler of the closed-source world. It shouldn't be hard to do this! Release IE8 as a standalone, or stand back and stay out of my way! "It's just another case of history repeating."

• Anonymous
  March 08, 2008
  http://www.update.microsoft.com/windowsupdate/v6/thanks.aspx?ln=en&&thankspage=5 You must be running IE5 or later :)

• Anonymous
  March 09, 2008
  @John: Yes, unfortunately you need to use the "Emulate IE7" button for now.  The WindowsUpdate team is working on fixing that.

• Anonymous
  January 09, 2009
  As announced in February 2008 , Internet Explorer 8 sends an updated user-agent string when interacting

• Anonymous
  January 18, 2009
  Обновленная cтрока User-Agent в Internet Explorer 8 Как уже сообщалось в феврале прошлого года , при

• Anonymous
  April 16, 2009
  2008 년 2 월에 말했던 (영어) 것처럼, Internet Explorer 8 은 웹 서버와 상호작용할 때, 업데이트된 User-Agent 문자열을 발송합니다. User-Agent