IEBlog

Awesome. Can't wait for the beta!

Looks nice! Keep the good work!

And of course you're going to allow me to run the beta standalone aren't you. I mean anything else would be crazy wouldn't it, given the number of changes you're making to turn your product into something that has suddenly started to adhere to standards...

Oh wait... you aren't?  I'm going to have no way of testing compatibility between versions 5 through 8 of what has been the buggiest, most consistently inconsistent browser of all time?

"Run them in virtual machines".  No.  I'm not going to have 4 different installations of Windows on my machine.  I can't be bothered any more.  I'm just going to develop in Flash, Flex and AIR (Silverlight can take a running jump given how I feel about how I've been treated as a developer by Microsoft over the last 7 incredibly painful years) and if I do write HTML I'm going to make it standards compliant and include a "gold bar" at the top of the window telling IE users to download Firefox.

Including proprietary headers like X-UA-compatible is not something I'll be doing just to save Microsoft's blushes.  As much as the current team is to be commended for their efforts, their employers did nothing for *years*.  Developers were required to work hundreds of hours for no recompense to get it to work properly and I, for one, have had enough.

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?

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.

> 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.

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?

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

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

dump the mozilla 4 part of it just make it MSIE X.X Win NT XX

(the x's are the version)

Just break everything one time then let the standards compliant engine rise out of the ashes.

these stuipid multile rendering engine thing is plain dumb.

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

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).

"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.

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.

@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.

PingBack from http://blogs.dotnethell.it/vincent/The-Internet-Explorer-8-User-Agent-String.__12954.aspx

Thank you.

@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.

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? :-)

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...)

@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.

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! ^.^

appMinorVersion.indexOf('b') you mean, not indexOf('beta')

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.

@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.

@AnomynousCoward

"Changing the user-agent string to "Internet Explorer/8.0" will break a few poorly designed sites"

It would break many apps (not just internet based) and as usual, a lot of people would scream bloody murder.

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.

@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.

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!

"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 ?

Mnnnnnn, many questions.

This blog entry doesn't state what user agent string IE8 will send out when in IE7 mode. Can this be clarified?

I can see in my stats in the future many UA strings with IE7, but how do I know that this is not IE8 in IE7 mode, where the majority of visits is actually IE8 masqurading as IE7.

Which one of these is IE8?

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

or

Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.2; MS-RTC LM 8)

It you could hit one of my pages one night with both in IE8 standard mode and IE8 in IE7 mode I will be able to tell the modes apart from each other.

That will make some IE team members wonder. :-)

So does IE8 in standard mode still use the IE7 targeting hack?

*+html div {min-height:1%} /* hasLayout trigger */

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)

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 ...

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

"if ((null != window.navigator.appMinorVersion) &&"

shouldn't that be " null !== ".

Hey Now!

The beta is going to be great!

Thx 4 the info,

Catto

@ 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?

it is really annoying to post a legit comment only to have it filtered out.

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".

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)

That's a good idea! Good luck with IE8.

Actually, the whole test case is overblown. navigator and alert are global objects, and never null...

@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.

@EricLaw [MSFT]

Thank you for your reply. Another question please if I may. Does IE8 use the following IE7 targeting hack?

*+html div {/* IE7 style */}

What would happen if IE8 in standard mode is targeted by *+html. IE7 style targeting the edge. Is Microsoft going to suggest to developers that we should use conditional comments? Add the very things that make these *  hacks possible.

Keep up the awesome work.

And please do your best to ignore the negative comments out there. ;-)

@Mark Wisecarver

What negative comments? Where? This has been a good start for a recent IE8 announcement.

@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.

@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.

@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.

@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?

@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?

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

Remove me from the ice blog list. Thanks

Remove me from the ice blog list. Thanks

If the ASP Atlas AJAX folks can put together a website that allows for tracking bugs/features and upcoming releases of their software - why can't the IE Team do it?

http://www.codeplex.com/AtlasControlToolkit/WorkItem/List.aspx

This seems so simple, and best of all you can get an accurate count of which bugs/features users care about most (which you CANT right now.  right now you only hear from the developers that are paying the most for their MSDN licenses)

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.

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.

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....)

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

Could I ask what will the IE 8 RSS user-agent be?

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.

http://blog.mozilla.com/gen/2007/02/27/the-cost-of-monoculture/

@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.

Why are you using Mozilla in the string? It shows both a pirate approach and a low self esteem.

I thought Microsoft had stopped stealing from people.

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

@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.

@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.

Here's to another 500mil downloads of Firefox plus as many downloads as possible of Opera and Safari.

Firefox 3.0 is shaping up to be a very incredible release, very pleased with the nightlies, write code and it just works. If it didn't work, fix site code and it works, if bugged then just report to Bugzilla and it gets sorted out by browser devs. Amazing, unlike here where this is just a marketing too for spreading misinformation about some browser passing Acid2, when it won't out of box.

@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?

@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.

@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.

In der ersten Jahresh�lfte 2008 soll die erste Beta des Internet Explorer 8 erscheinen. Eric Lawrence, einer der Program Manager f�r den Internet Explorer, hat jetzt im IEBlog einige Informationen zum User Agent String des Browsers ver�ffentlicht. Der

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.

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...

Heh... Let me get this straight.

So, basically, the main excuse for IE to make the super standards mode opt-in (not default) was to avoid "breaking the web". Insisting on bad sites that couldn't be modified (like, on CDs), and bad intranets.

Therefore, the big idea was to make the IE7 standards mode the default, in order to ensure that not a single site would be broken, at the cost of giving the middle finger to web standards...

Yet... with the new UA string, these examples may still not work. Not to mention badly coded sites are even more likely to do browser sniffing than your average site... In other words, the super standards mode was disabled by default for no good reason, in the end.

None of this makes sense... Or actually, it does, sadly. Microsoft acts like they suddenly want to play nice and respect standards, but still do everything they can to hinder web development and unfairly abuse their monopoly to kill off competitors.

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.

i  have just about had it with Microsoft i have been a web developer for over 8 years and at the point of just doing something else!!! coding now for 3 versions of ie is making everyone nuts do you all not read what people are saying, i think not cause now your going to put another one out there come on would you all over there just please resign and have them ship the os with firefox For The Love Of God please BILL just let them all go.

crying web developer  

@Andrew

"I imagine the performance hit of inspecting a page for a meta tag would be minimal."

What if the meta tag is at the end of the head section, or after the closing html tag? The team never specified WHERE in a page the tag has to be. Will the browser hang while the page is downloaded and then rendered?

My solution is to have true standards mode enabled by default, with IE7 mode as an opt-out.

"[T]ell me how a server can target IE8 standards mode (via the opt-in tag) without a change to the User Agent string within IE8."

Set your server to transmit the IE=edge HTTP header to ALL browsers forever. The userAgent has NOTHING to do with the opt-in.

And as I've said, I have nothing wrong with IE8 identifying itself as IE8 in true standards mode.

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.

Why the need to detect a user agent string again?

@Gyrobo: As stated in the original article, the UA-Compatible META tag must be placed in the HEAD of the document.  Placing at the top of the HEAD would be wise.

@EricLaw

But how would the page render if the meta tag was placed at the END of the head, with a script ahead of it? And what if it was nested inside a conditional comment, or added dynamically by the aforementioned script?

You still haven't explained how conditional comments will work in IE8. Will the IE7 rendering engine use comments coded for IE7? How will conditional comments be interpreted before the meta tag is discovered?

Also, you seem to have ducked my request for an example of a SINGLE case where removing the "mozilla" part of IE's userAgent would cause breakage. Were you lying when you said that removing "mozilla" from IE's userAgent would cause breakage, or are you still looking for such a case?

Would you rather I phrased that differently so you could save face?

@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?

@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. ;-)

Great job guys. Please don't disappoint us with IE8. I've got high hopes.

I'm quite sure Microsoft understands that if they don't make IE8 something really revolutionary of some kind, then there's little or no hope of making people like me switch to alternative browsers that probably do better.

So please, add a ton of new features that are non-existent in other browsers, or improve your features so they do the best.

Don't disappoint us!

@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?

@Gyrobo you wrote:

"You still haven't explained how conditional comments will work in IE8. Will the IE7 rendering engine use comments coded for IE7? How will conditional comments be interpreted before the meta tag is discovered?"

Conditional comments in IE8 standard mode will work the same way as the currently do in IE7, and I mean "exactly" the same way.

It think really is important where you place the meta tag. If the meta tag was placed after the conditional comments for IE7, then it's possible that any linked script or CSS inside these comments would be loaded first before IE8 had a chance to realized that it should be in standard mode.

As EricLaw wrote, "Placing at the top of the HEAD would be wise."

@EricLaw MSFT

Please confirm (yah, or nah) if IE8 will be available on Windows XP.

Thanks.

@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.

@Alan Gresley

Thank you for your engaging and provocative response. It fills me with queasiness and unease, but that's hardly your fault.

gyrobo {roboshrub inc??} why do you suppose it that they're not answering your questions?  you are clearly not rational, and some people have better things to do than try to reason with a nutcase.  the fact that opera works is completely irrelevant... your question was what breaks if {mozilla} is removed from the IE ua string... two examples were provided, and rather than admit your lazy error, you start hurling insults and trying to change the subject... you can hide your name behind your anonymity, but you can't hide your ignorance.

"This is a very exclusive program, by invitation only."

http://www.activewin.com/awin/comments.asp?HeadlineIndex=42767&Group=1

You other web developers need to quit whining. You don't HAVE to do all sorts of extra work for IE8. Say you already have a standards compliant page (and don't have special mods for different browsers), then you don't need to do a damned thing.

If you already do a check for something like "IE6, IE7, Opera, Firefox, and All Others", then you can probably let IE8 fall into the all others and let it render the page compliantly, or you can make the check go like this "IE6, (IE7 or IE8), Opera, Firefox, and All Others".

The only problem I can see arising is for sites that just do a basic check for IF IE then render the page this way. In which case its still an easy fix. I've realised that most other web developers create a lot of extra work for themselves thats just unnecessary, and instead of stepping back, taking a look, and figuring out if they should change it, most of them just keep charging ahead and do everything the hard way.

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.

@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.

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

Stifu: I'm glad that your sites will be fine, as will mine. But it shouldn't matter for the web in general. The IE Team is making perfect sense. Yes, the whole situation sucks. But the "don't break the web" mantra makes sense from the user point of view. The real "super standards" mode will function and view the page just as FF or Opera would WITHOUT modifying anyone's webpage, except for maybe a single line of code, at most. The "IE7 mode" will just allow the user to view pages "more correctly" if some web developer has that catch-all 'IF this browser is IE then display this way'. That way, for the user, the webpage is never broken. It either works because it worked in IE7, or it works because it is a standards-compliant site, that doesn't do something different for IE versions.

Now I'm no IE evangelist, as I do use Opera on about a 60/40 ratio to IE. But between the two I usually enjoy my IE experiences better for a couple of reasons. 1) pages tend to load better for me. 2) Opera has this weird tendancy to display images on sites in inappropriate places forcing me to refresh once or twice. 3) In Opera sometimes when opening a new tab, if the site hangs a bit on the load time, in hangs the whole browser.

So, why do I use Opera more? The native mouse gestures. That's about it.

As for FF, I just don't like it. I don't know why, but when I sit there and use it, I just get angry for no reason. It's not like I need to get used to it or anything, I just don't like it, and I probably never will.

*And yes, I'm aware that there is an IE add-on that allows the use of mouse gestures. The only 'add-ons' I run are things like Flash, trusted AtiveX's, and the sort.

@Al Billings

Al, I've said all I need to say about the reality of your posts here.  

I have no inside information, I just recall it being said that some form of bug tracking will be available during bet