IE8 May Not Pass the Acid2 Test After All

dotne writes "CNET has published an article called Acid2, Acid3 and the power of default. The article predicts that IE8 will not pass the Acid2 test after all: '[Another] scenario could be that Microsoft requires Web pages to change the default settings by flagging that they really, really want to be rendered correctly. Web pages already have a way to say this (called doctype switching, which is supported by all browsers), but Microsoft has all but announced that IE8 will support yet another scheme. If the company decides to implement the new scheme, the Acid2 test — and all the other pages that use doctype switching — will not be rendered correctly.' Microsoft's IE8 render modes have been discussed here previously, and they've caused an uproar in the web development community. According to the scheme, authors must put Microsoft-specific <meta> tags into their pages in order for them to be rendered correctly. I doubt Acid2, nor Acid3 will have Microsoft extensions in them."

Page specific tuning

It's possible that IE8 will contain code that detects the presence of an ACID test and switches to the proper renderer to pass the test.

Re:Page specific tuning

The author of Acid2/3 is not amused [hixie.ch] by this meta tag. From the tone of that blog post, to me it sounds like he wouldn't shy away from actively try to break a mechanism like that by changing the URI to make sure that the browser that passes the Acid test actually does so for real.

Note, though, that he doesn't say that explicitly, and you shouldn't assume that he will. It's my own conclusion, and you should draw your own, etc...

Re:Page specific tuning

Forgive me if I'm wrong (as I'm not an HTML guru in the least) but isn't that the point of DOCTYPE? Meaning, if a broken page wants to use the buggy renderer they shouldn't be setting a strict DOCTYPE.

Microsoft is so committed to their long-standing policy of coddling the incompetent that they want a way to be lax on pages that specifically request a strict interpretation.

Re:Page specific tuning

The issue is that IE6 allowed people to use the strict rendering with out truly strictly rendering stuff. When IE7 was released that was more strict, it broke a lot of pages that assumed that strict worked because IE6 worked.

They already corupted the doctype tag. The logic behind the new tag is to indicate the renderer you want so future releases do not break the current page.

They dont want to break the web again (like IE7 did). So the web will work by default (as defined by IE 6) and new stuff that targets the new browser should not break when IE 9 is released.

They are attempting to end a cycle of new browsers breaking older pages.

What they need to do is do it right the first time and fix rendering bugs quickly. force people to fallow the standard and it will work. Microsoft should be the most accurate implementers of the specs because they have such a huge install base.

Re:Page specific tuning

Here, let me fix that for you:

They are attempting to end a cycle of new Microsoft browsers breaking older pages.

Opera, Mozilla/Firefox don't have this problem so why should Microsoft. By adding a tag that specifies the browser it makes the page browser specific and since we are talking IE it is now OS specific .

Sure a web developer could code other browsers in the tag but it will be ignored by other browsers because other browsers already follow standards and don't require the hint. The metatag might as well be called "IEVersion(tm)" because that's what it is. Once again, Microsoft is trying to make a standard Microsoft specific because they are too stupid, lazy or ignorant to implement what everyone else on the planet has agreed to.

Microsoft has no intention of ever being the "most accurate implementors of the specs" because they have no incentive. Implementing a standard does not make Microsoft more money, tying people to Microsoft products does.

The tag gives Microsoft a get out of jail free card so they don't have to follow standards ever again. Using the tag only reinforces their behaviour.

And 100000 slashdotters...

shout "SURPRISE!" in unison.

Steamrollin' On

> I doubt Acid2, nor Acid3 will have Microsoft extensions in them.

But lots of web pages will.

Re:Steamrollin' On

My pages WILL have them, if it's easier to make render them correctly than custom hacking. I may not like IE8 but my clients will use it and they will require that my pages work ok on ie(6|7|8), firefox and opera.

Amazing

Another Microsoft "We'll do it our way, and you'll do it our way too if you know what's good for you."

I wish Microsoft would at least learn to fake sincerity in actually following common standards. This isn't even lip service. This is "We follow standards (for certain Microsoft-centric values of 'standards')."

Of course, the market has rewarded them, so why should they change? All they need is smoke, some mirrors, and some moderately-skilled PR, et Voilà! "standards-compliant!"

Re:Amazing

I think you're missing the point somewhat here, as are a lot of people. The core issue is that Microsoft have painted themselves into a corner by not following the specifications in the past. This makes it difficult to follow the specifications today, as all the pages that expect Internet Explorer to deviate from the specifications will break if Internet Explorer starts following the specifications.

They solved this in the past by inventing doctype switching, which is a pretty poorly-thought out hack, and with conditional comments, which uses non-standard syntax. This approach, on the other hand, uses standard HTML syntax and is designed for this purpose.

Far from being "A Microsoft extension to HTML" as some people are describing it, this is an attempt to retain backwards compatibility while fixing Internet Explorer to more closely follow the specifications. It does this using the standard HTML way of incorporating additional metadata. Let me repeat that: what you are seeing is Microsoft improving conformance with the specifications. The mechanism they are adding is so that older code doesn't break, which is an entirely reasonable thing to want.

If you have a better idea of how they can satisfy the constraints of backwards compatibility and closer conformance to the specifications, please, describe it. As far as I can see, this is the best way of doing it. Sure, it's their own fault they are in this mess, but bitching about their past behaviour doesn't mean that this attempt to fix things is the wrong approach. Microsoft are doing the right thing here.

Make Acid2 the Default

Why not make Acid2 the default? I'm sure the browsers interals could look for IE6/7 "hacks" and provide a icon on the bottom to have it viewed in compatibility mode? If the broken mode is going to be the default, I think standardization will be slow, unless common developer tools like MS Visual Studio and Dreamweaver put in the MS Specfic renderer tags in by default.

I think we're at the moment when developers want standards, where in the IE4/NS4 war, everyone and their brother was trying to hack-together web pages, and IE did some nice exposition of the DOM via the ID attribute in tags, which accomodated less-skilled programmers. Now that the baseline-developer's skills are improved, and the IDEs out there are actually pretty decent (e.g. Not FrontPage, Not MS Word) I'd say the time is right.

While the Acid2 test is niceity, what I'd really love to see is a standard plugin model shared by FF and IE. It has been a while, but I always thought the "EMBED" inside of an "OBJECT" tag was lame. I don't like ActiveX but I get in intranet environments where it can be useful, where the code should be "trusted" and "signed", where you're essentially using a browser to "publish" applications that should probably be desktop applets, or use a native HTML (AJAX?) interface rather than "VB applet on a webpage." That being said, we need an out in the wild, "safe" plugin/viewer model.

Re:Make Acid2 the Default

Why not make Acid2 the default?

Because lots and lots and LOTS of pages would break, among other things. Earlier today there was another article in ./ with a link to the full rationale behind this [alistapart.com], and to me is makes a lot of sense. Basically, with this tag you can specify a version for each browser on which the site was tested and is known to work well, then all browsers might keep internally working versions of their legacy rendering engines (or a compatibility mode built in their newest engine, whatever works best in each case), and forever in future you'd have old sites being 100% readable in new browsers, no matter how much actually existing "de facto" and "official" standards change or get deprecated/replaced over time. An example from the above link, specifying the page renders correctly in IE 8's engine, Firefox 3's engine and (say) Opera 4's engine:

<meta http-equiv="X-UA-Compatible" content="IE=8;FF=3;OtherUA=4" />

What is there to not like in this? It's a simple, elegant and practical solution to this very real problem. Sure, it could have arrived earlier, but better later than never.

Re:Make Acid2 the Default

It's a simple, elegant and practical solution to this very real problem.

It's a problem, though, that only Microsoft has. Everyone else is just expected to conform to the standards.

Read here [webkit.org] for the WebKit team's response to this and why they're not going to define or obey any such tags themselves.

Re:Make Acid2 the Default

<meta http-equiv="X-UA-Compatible" content="IE=8;FF=3;OtherUA=4" />

What is there to not like in this? It's a simple, elegant and practical solution to this very real problem. Sure, it could have arrived earlier, but better later than never.

I do know what all the non-IE users here on slashdot will say, and there's quite a few of them (myself included). First off, it looks like only Microsoft will use this, other browsers seem to have broken compatibility early and often and have a fairly rigorous standards compliance. I'm sure evil tounges will say it's because only a minority use these browsers, but I've not heard any noise over Firefox or Opera breaking pages by fixing things. On the whole they seem to be way ahead of web developers, so the benefit of extending it to other engines is pretty much neglible I'd say.

The possible downside is that precisely because these buggy pages work in IE, noone will bother to fix them. Imagine you're at IE12 and it's now a super anal 99.9% standards compatible browser in IE12 mode. What good will it do if pages still call for IE7 compatibility mode and are still broken on every other browser? That said, between "We won't fix the browser because it'd break sites" and "We won't fix the pages because they work" I think the latter is easier to fix and new code would presumably be written to spec and work in all browsers. So it's something of a percieved downside which I don't think is real.

Now, if I take on my hat as an IE user it's a slam dunk. More pages will work. If I take on the hat as a web developer, I know my pages won't suddenly break (because you were forced to use an IE hack to make it work right in the first place). Also a slam dunk. If I take on the hat of Microsoft, they don't get screaming customers about how they broke the intranet. Also a slam dunk. If Microsoft was really really nice, they'd put in their status line "This web page is designed for a higher version of IE. Please upgrade to ensure optimum performance" too, so we'd lose these buggy versions as quickly as possible.

It's easier to cut the developers loose and let them fix IE than the pains of trying to slowly break IE while not getting killed by angry customers. The sooner they get a standards compliant browser out there, the sooner we can get to *really* fixing the problem.

Re:Make Acid2 the Default

What is there to not like in this? It's a simple, elegant and practical solution to this very real problem.

First of all, I've been working in software long enough to know that I can't trust it for a minute. Oh sure, they may say they're going to maintain all these special compatibility modes, but in reality one or more of the major browsers won't do it, or they'll try but they'll screw it up, or they'll do it for a while and then suddenly drop all the backwards compatibility stuff because they don't feel like maintaining it any more. Regardless, eventually it will come to pass that I can't trust the version specifying mechanism to do me any good, and I'll have to update my pages anyway.

Secondly, it encourages the web to come to a halt, technology wise. As an expensive consultant to big companies, my experience has been that they all want to try to tell you "our web site has to look perfect on every version of every browser ever invented, period." Of course that's impossible, so once I drill it through their head that it's impossible, they have to settle for some major subset, at which time they always want me to use ancient UI technology for maximum compatibility. (Seriously, I deal with people who freak out if I want to use CSS positioning or an iframe, and god help me if I mention AJAX.) If they got the idea that browsers have a magical compatibility mode so that all future browsers will support pages written for today's browsers, they'll instantly write up a corporate policy that basically says that nothing will ever change again and for the rest of time their web site will be maintained as if it is forever 2005, and then they won't change it until someone practically holds a gun to their head to force them to.

Now, you're asking yourself, why should you care? Because it's more than a few idiots, it's a substantial portion of the web. Sure, there will always be little guys who will come along and innovate, but do you really want to deal with a web full of sites that forever use 2005 technology and just a few sites that have caught on to the latest stuff?

IDGI

I don't get it. Why is there such an issue. As a dominant software company, Microsoft should be allowed to do what it wants. After all, what is best for Microsoft is also the best for America.

Therefore, if you are against Microsoft, you must be a terrorist.

Please report yourself to the nearest detention center for correction.

History repeats itself

1995 called, they want their <blink> and <marquee> tags back

The real problem is Dreamweaver

The real problem isn't on the browser side. It's in Dreamweaver, the most popular web page design tool. Dreamweaver does not create valid HTML or XHTML. Not even close. Create a page in Dreamweaver for anything later than HTML 3.2 and run it through the W3C validator. It will fail.

The basic problem is that the Dreamweaver people never really figured out what to do about CSS. In theory, CSS is supposed to have some abstract model of the format of some block of text, like TeX does. In practice, there's usually a big block of CSS with machine-generated names at the beginning of the web page. There's a fundamental disconnect between the CSS model and the Dreamweaver "Properties" box. So Dreamweaver is still inserting I, B and FONT tags.

In layout, Dreamweaver does table-based layout quite well, but DIV/FLOAT/CLEAR layout isn't handled well. Much of this is due to the limitations of the DIV/FLOAT/CLEAR approach. Tables are a 2D grid system, and map well to the drag-the-lines editor in Dreamweaver. DIV/FLOAT/CLEAR doesn't map well to a visual layout editor.

The end result is a mess. And HTML 5 doesn't help.

Worst... Proposal... Ever!

Microsoft claims that the X-UA-Compatible flag is necessary on standards-compatible content to avoid breaking IE-specific content. I call BS.

For years, Microsoft has been telling everyone to put version-specific IE hacks in conditional comments [microsoft.com], in case IE's behavior improves in future versions. Now that they are finally fixing IE, they spring this X-UA-Compatible "solution" on us, punishing those who have been producing standards-compliant content and rewarding the zombies who have been writing IE-specific code. If your site is standards-compliant, you have to do the extra work to tag it as such, and keep that crufty tag around for the foreseeable future!

If you sat down today and wrote a new standards-compliant browser, it would work just fine with almost all the content and web applications out there. Apple did this recently with Safari. Microsoft claims to have done this with IE 8. Safari didn't need any X-UA-Compatible flag. Why should IE 8 need one?

The only reason IE 8 would need the X-UA-Compatible flag is simply because it is IE 8. If their new browser identified itself as, say, "Microsoft Trident VI [wikipedia.org]" instead, things would just work. Microsoft could still call it "Internet Explorer 8" for marketing purposes, but web developers would know that "MS Trident VI" means IE 8, just as "WebKit 4xx" means Safari 2 (or similar browsers) and Gecko means Firefox (or similar browsers).

Dear Microsoft, here's a sane solution for you:

1. Ditch the X-UA-Compatible flag; it's a stupid idea.
2. Continue supporting HTML conditional comments as you have been doing.
3. Fix the layout engine and the CSS parser at the same time, so that any existing IE-specific CSS hacks become irrelevant.
4. Add support for CSS conditional comments, to give web designers an escape route. Let's face it, CSS hacks are a reality, so we might as well have a tool to do it cleanly.
5. Send this as the User-Agent string: Mozilla/5.0 (compatible; Microsoft Trident VI; Windows NT x.y; ...; Microsoft Internet Explorer 8.0; ...). Any server-side code doing browser sniffing, not seeing the "MSIE" string, should send a standards-compliant response. User-Agent strings have never really been logical anyway -- IE started this mess years ago by sending the "Mozilla" string, and Opera continued the trend by optionally sending the "MSIE" string -- so additional games in this area wouldn't do any more harm to the Web.
6. In JScript, navigator.appName should return 'Microsoft Trident', and navigator.userAgent should return the string above. Client-side scripts doing explicit browser sniffing, not seeing the "MSIE" string, would suppress their legacy IE hacks.
7. In JScript, document.all should evaluate to false (although expressions involving document.all can still behave as in older versions of IE). This approach worked for Mozilla [mozillazine.org], and it will work for Microsoft too.

As you see, it is possible to fix IE in a backward compatible way without introducing a X-UA-Compatible flag. The chances of Microsoft taking these steps is almost nil, since it places IE 8 on an even playing field with other standards-compliant browsers. That's why they are proposing X-UA-Compatible -- they can claim to support web standards while knowing that web developers will find it easier to muddle along than to use their stupid flag.

Re:It's the most logical decision

IF MS were to change the way pages rendered with existing doctypes, millions of pages could/would render differently requiring businesses and individuals across the world to either re-vamp their websites or at least change the existing doctype to a new name that referred to the old rendering style.

Sounds like just the medicine they need for creating browser specific web pages.

In other news, many Americans against government bailout of mortgage companies that made bad loans...

Re:It's the most logical decision

On a purely philosophical point... what is the use of having an international standard if the said standard changes based on the whims of a single corporation?

I get your point, and you may have a more real-world-ready opinion than some of us, but I am not ready to concede anything to MS in this regard.

There are other ways MS could address this issue rather than continuing their embrace extend destroy strategies. There is no reason IE specific tags should be required to make a page display according to an international standard.

If anything, broken pages should require tags to inform the browser that they do not conform to standards, and thus require special attention - not the other way around.

Re:It's the most logical decision

That ACID(2,3) tests are designed to test browsers, browsers are not designed to test ACID. As such, ACID should be updated to include the new doctype option for IE.

So, if I take a test and don't pass it, the test should update itself to include my wrong answers?
ACID is designed to test a browser's adherence to a set standard. Its not designed to just 'test' a browser to see if it works. It's designed to see if it works the way a browser should. I say break the millions of web-pages and force them to get updated.

Re:It's the most logical decision

IF MS were to change the way pages rendered with existing doctypes, millions of pages could/would render differently requiring businesses and individuals across the world to either re-vamp their websites or at least change the existing doctype to a new name that referred to the old rendering style.

I don't buy it (or maybe I just don't get it--if so, please explain).

IE (just like Firefox, etc.) has a "quirks" mode which renders things in a non-standards compliant way, but is designed to "more or less work" with all the pages out there that are not strictly coded. This new tag is supposed to apply to web-pages where the web author has already explicitly said he wants strict rendering, because he said so in the DOCTYPE. But instead of just fixing IE so that it renders that standards-compliant code better and better, they propose to freeze that rendering sub-engine, and force web-developers to add a new tag that basically says "yes I really meant I wanted you to render strictly!"

It seems to me that the majority of pages that rely on rendering quirks will be okay, since they will be rendered in quirks mode. But pages that were intended to be standards-compliant should be treated as such.

Microsoft's plan isn't sustainable or elegant: they basically want the entire web-community to add another tag each time MS releases a new version of IE. (If they want a custom tag for the IE7->IE8 transition, they probably will want a new one for the next transition...) The entire point of these standards was to get away from browser-specific tags and hacks. A web developer shouldn't have to think about what browsers are on the market today (or 3 years from now): he should just code to the standard.

Put otherwise: Instead of asking everyone who has written a standards-compliant page to add-in a non-standard tag to make it work in IE... wouldn't it be easier to tell everyone "hey, if you've coded a page that is ~almost~ standards-compliant, but relies in some way on IE7-specific behavior, then add in this <NotQuiteStandard> tag, and IE8 will render it like IE7."

Re:It's the most logical decision

This new tag is supposed to apply to web-pages where the web author has already explicitly said he wants strict rendering, because he said so in the DOCTYPE.

No, that's not at all the case.

Way back when the web was first taking off in a big way, browser vendors were adding lots of proprietary element types and attributes. Most of them weren't very well thought out, but there was a pressing need to get them into a specification to ensure interoperability. The result was HTML 3.2, with a lot of stuff that really shouldn't be in there.

The next version of HTML took all these out again because better ways (e.g. CSS) had been developed to handle their functionality. Of course, there needed to be an upgrade path from HTML 3.2 to HTML 4, so the W3C published HTML 4 in two major forms — a backwards-compatible document type, called "Transitional", and a modern document type, called "Strict".

That's what the "strict" refers to in the doctype. It's got nothing to do with rendering modes.

Now, when Internet Explorer 5 for the Mac was being developed, Microsoft realised that they couldn't improve its conformance to the specifications without breaking backwards compatibility. They came up with a hack that guessed whether to be backwards-compatible or specifications-compatible by looking at the doctype to see if it was there, if it was out of date, etc.

This "doctype switching" is a proprietary, non-standard hack, that happened to catch on amongst browsers. The fact that a document refers to the strict doctype does not imply that the developer is choosing a particular rendering mode, it just means that they are using a modern document type.

But instead of just fixing IE so that it renders that standards-compliant code better and better

No, this is exactly what they are doing, and this is the very reason why such a switch is necessary.

Microsoft's plan isn't sustainable or elegant: they basically want the entire web-community to add another tag each time MS releases a new version of IE.

No, this isn't the case. Any developer who wants Internet Explorer to use its most recent rendering engine at all times can select "edge" and be done with it.

The entire point of these standards was to get away from browser-specific tags and hacks.

Yes, and this is an attempt to do just that. Doctype switching is a proprietary hack. <meta> is the standard HTML method of including metadata. Furthermore, providing a switch like this allows them to slowly deprecate the older rendering engines over time.

Put otherwise: Instead of asking everyone who has written a standards-compliant page to add-in a non-standard tag to make it work in IE... wouldn't it be easier to tell everyone "hey, if you've coded a page that is ~almost~ standards-compliant, but relies in some way on IE7-specific behavior, then add in this <NotQuiteStandard> tag, and IE8 will render it like IE7."

No, this would cause mass regressions because there are a hell of a lot of unmaintained websites out there.

Re:so?

Firefox 3 should be ale to pass the acid 2 test. They have been working on it, and they don't introduce custom extensions to doctypes to render correctly, rather they are fixing their browser itself.

MSFT dug itself into a hole. instead of getting out and filling the hole back in they are digging another hole to bring people down to their level.