IE8 Will Be Standards-Compliant By Default

A number of readers wrote in to make sure we know about Microsoft's change of heart regarding IE8. The new version of the dominant browser will render in full standards mode by default. Developers wishing to use quirks mode for IE6- and IE7-compatible rendering will have to opt in explicitly. We've previously discussed IE8's render mode a few times. Perhaps Opera's complaint to the EU or the EU's record antitrust fine had something to do with Redmond's about-face.

Comments

Huge assumption in the title

[Tumbleweed]

Let's make one thing clear - IE8 may be in standards-compliant MODE by default, but whether it's *standards-compliant* has yet to be proven. What Microsoft HAS proven (repeatedly) is that it considers compliance with standards to be a relative term. Only time will tell. I sure hope that they actually accomplish it this time; I'm tired.

Re:Huge assumption in the title

[Your.Master]

Compliance to standards is a relative term. No browser exists today that is completely compliant. What we can say is that this one appears to be more compliant than before -- it renders ACID2 at the very least (and probably does right everything IE7 did right).

Re:Huge assumption in the title

[SEE]

They've said it already passes Acid2.

Re:Huge assumption in the title

[Phroggy]

What Microsoft HAS proven (repeatedly) is that it considers compliance with standards to be a relative term. So do all other browser makers. The various standards involved are non-trivial to implement, and as another poster commented, nobody has implemented all of them.

Just because a browser passes Acid2 doesn't mean it's "standards-compliant". It means it complies with the specific parts of the standards that Acid2 tests for, which is only a few things that most browsers (at the time Acid2 was created) got wrong.

Re:Huge assumption in the title

[Anonymous Coward]

> Compliance to standards is a relative term.

No, this statement is incorrect.

> No browser exists today that is completely compliant.

That is true. But it has no connection with the last statement.

I understand your point, and it's well taken, but you are introducing a tautology. Standards compliance is absolute, by _definition_.
Some attempts to comply with written standards may fail, and as such are not compliant. It may well be true that no browsers exist that are standards compliant, as the standards are written. However, please don't go waving around poisonous ideas like "standards compliance is a relative term".
Americans seem to have adopted a very lax relativism of late, a kind of fuzzy belief that everything is subjective. Some things are not. Some things are just facts that must be heeded. The definition is not up for negotiation, that's what _makes_ it a standard.

Re:Huge assumption in the title

[Craig Davison]

This is good news. The important thing is that MS is now saying they're willing to sacrifice backwards compatibility in IE. They have no reason not to follow the standards now (barring bugs or technical limitations).

Re:Huge assumption in the title

[Merusdraconis]

That was true five years ago, but no longer. With the amount of competitive alternate browsers out there, and the new rise of the iMac, the standard is the W3C. While most web developers will put in extra effort to work around IE's bugs, they're starting from W3C-standard webpages and kludging in IE support, not (as it worked years ago) building pages that worked in IE first then trying to make them work on Netscape later if at all.

Re:Huge assumption in the title

[mysticgoat]

Would you agree that a physical theory being "correct" is equivalent to a physical theory being "standards compliant" where the standard is reality?

No, I don't agree with that stipulation.

Reality is in its essence unknowable. Theories are models of reality that are simpler, and are based on a multitude of assumptions. And many of those assumptions go unstated. For instance, I am aware of no theory of gravity that takes into account the color of the objects being described, yet there is no scientific basis on which we can exclude color (or smell, or taste) from gravitational considerations. We do so because at this moment in history it seems silly to include it, but that is a literary arts judgment, not a scientific judgment. If you want to get your pet theory on Electric Pulse Gravity published, you'd do well to heed the literary aspects, but don't mistake them for the science.

A standard, however, is the formal statement of a group's conceptualization about a process, such as how a distance shall be measured, or how a web page shall be rendered. A standard has nothing to do with reality. It is all in your head (and the heads of everyone else who familiarizes themself with the standard). Because a standard is a human production that has no physical reality, it is possible to fully comply with its every detail (assuming that it is a well-written standard). Perhaps more to the point, it is possible for someone to completely learn a standard, including any of its weaknesses like internal contradictions or ambiguities. However it is impossible for anyone to completely learn reality, or learn all there is to know about any theory of reality.

In this sense, Euclidean geometry is a standard. You can do a lot of neat things with it, and you can spend lots of time exploring places where it is still ambiguous (things not yet proven). But you can't violate its established rules and still claim it is Euclidean geometry. You can replace those rules with other rules, but then you have a non-Euclidean geometry, like spherical geometry as one instance.

It is possible for a web browser to be standards compliant in the absolute sense. It is also reasonable to describe the relative compliance of non-compliant browsers. And since in nearly every case the context will make it clear as to whether the meaning is absolute or relative, there is no rarely any need to specify that. Unless, of course, one is pushing a hidden agenda, where the intent of talking about the subject is to create as much heat and smoke as possible while putting out no light.

There's probably a really succinct way of saying all the above, but I left my Zen Pocket Companion at work.

Re:Huge assumption in the title

[statusbar]

With respect to the http 1.1 standard, it _is_ relative...

From the standard:

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [34].

When the people writing the standards write standards with the words "SHOULD" or "SHOULD NOT" or "RECOMMENDED" or "MAY" or "OPTIONAL" you now have a standard which can have many different faces, or compliance levels. IMHO, this is poor standards writing. They MUST make the specs using the terms "MUST" and "MUST NOT" and bump the version number. Then you can easily have automated unit tests which show absolute compliance. But we don't, and must rely on what developers "THINK" or "MAY NOT THINK" is correct about the spec.

--jeffk++

Re:Huge assumption in the title

[cp.tar]

More standard than IE7 isn't really a high bar to aim for though.

It is much higher than "more standard than IE6".

Re:Huge assumption in the title

[edwdig]

Have you ever tried implementing something described in an RFC ?

When you get to the should / should not stuff, it comes down to in most cases you really want to listen to it, but there tend to be specific cases (say, embedded devices) where it really doesn't make sense to follow the normal behavior. Generally, if you run into one of those cases, it tends to be obvious that deviating from the spec is the right thing to do.

The optional and recommended stuff tends to be things that really depend on the specific product and shouldn't be forced.

Making things more strict would be a bad thing and make people break the standards more. The current setup acknowledges that different implementations have different needs and does a good job of accommodating.

Re:Huge assumption in the title

[Your.Master]

[quote]yet there is no scientific basis on which we can exclude color (or smell, or taste) from gravitational considerations.[/quote]

Sure there are. For instance, parsimony. Repeated experimentation holds that these properties are pretty much unrelated to gravity*. Besides, other models that appear consistent seem to adequately explain colour and smell and taste in ways that are incompatible by virtue scale with consistent gravitational theory.

Anyway, I recognize and respect your distinction here, although I do think it's hypothetically possible to come up with a model for reality accurate in every respect, we can just never truly comprehensively know that we have found the answer. That's neither provable nor disprovable, and thus, neither here nor there.

But I agree that context makes things clear pretty much always. If you look at the original context of my statement, he first used compliance as an absolute term, then declared that Microsoft viewed it as relative. I argue that the relative interpretation is quite valid. As for the hidden agenda, I don't think any of us (you, me, the guy I originally responded to, the Anonymous Coward in between) was pushing any hidden agenda, so I don't know where that came from.

I apologize if I looked like I was pedantically claiming that the absolute interpretation was invalid and retract any implication thereof. From my perspective, the person I responded to was saying that the relative interpretation was invalid, and the guy who responded to me agreed with absolutism-only.

* pigments for colours are slightly different composition, smells are different aromatic molecules, taste from that and other factors; all of which can reflect subtle molecular differences that lead to different mass per molecule which can in turn lead to different gravitational forces. Let's not analyse that one too deeply :).

Re:Huge assumption in the title

[Phroggy]

I understand your point, and it's well taken, but you are introducing a tautology. Standards compliance is absolute, by _definition_. I take your point as well, but I want to interject that this entirely depends on what the standards are. Can you build a word processor that is 100% compliant with Microsoft's OOXML standard? Not really, in any meaningful sense, because the standard is incomplete and refers to behavior that isn't described as part of the standard (that's a large part of what all the OOXML vs ODF fuss was about). Can you build an IRC client that is completely standards-compliant? No, because the RFCs that describe how IRC works are incomplete, inconsistent, contain errors, and aren't strictly adhered to by any popular implementation.

In the case of HTML/XHTML and CSS, there's been quite a bit more effort invested into making sure the standards are properly documented and are internally consistent, but these standards are constantly evolving. Is it enough to support HTML 4.01 and CSS 2, or must you support HTML 5 and CSS 3? Do de-facto standards count? Remember that XMLHttpRequest (the basis of AJAX) is mostly a de-facto standard; the W3C has published a working draft of a specification for it.

Standards compliance isn't always as cut-and-dry as you make it sound.

Re:Huge assumption in the title

[DanielJosphXhan]

So what you're saying is that by IE15 we should see a fully standard-compliant browser?

Microsoft has the be the only organisation on earth *slower* that the W3C. I mean, it's not exactly a moving target.

Re:Huge assumption in the title

[strikethree]

There is a world of difference between not fully implementing a standard and doing something against the standard.

strike

Re:Huge assumption in the title

[Random BedHead Ed]

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119

For those too busy to consult RFC 2119 in detail, it basically states the following:

• Should should mean "should."
• Must must mean "must."
• Must not must not mean anything other than "must not."
• Required is required if you want to express the idea of a requirement.
• Shall shall mean "shall."
• Shall not shall not be construed as indicating that something "shall." (In fact it shall be the opposite.)
• Should should usually mean "should," but not always.
• Should not should not mean anything other than the opposite of "should," but also not always.
• Recommend is recommended for use in RFCs as well, but may be optional.

Correct link

[RockMFR]

http://blogs.msdn.com/ie/archive/2008/03/03/microsoft-s-interoperability-principles-and-ie8.aspx

Re:I don't care about IE at all

[bunratty]

Firefox 2 is one of the most standards compliant browsers around. What other browser does significantly better overall at standards compliance than Firefox? Check out the link I provided to webdevout's information on browser standards support before you reply...

Put it all on Silverlight!?!

[99BottlesOfBeerInMyF]

I wonder if they're serious. Will they really be standards compliant enough so that I don't have to hack around IE8's deficiencies? Will this still be true for IE9? It's possible. Will this include SVG and XHTML and CSS3? What about XUL and HTML 5?

If all of the above work in the next couple of version of IE, do you know what that would indicate to me? That would indicate that Microsoft is betting on Silverlight to lock in users in the next 5 years... because they've pretty much convinced me they will never compete based upon features and the merits of their software, rather than trying to make it as hard as possible for users to switch to anything else.

Re:Hmmmm

[garett_spencley]

Competition is good. If Microsoft actually goes and creates a superior product then IE users get a better browser which forces Firefox to either "up it's game" (giving FF users an even better browser) or remain the same while everyone switches back to IE because it's superior.

Either way everyone gets a better browser. Win-win.

Re:I don't care about IE at all

[99BottlesOfBeerInMyF]

What other browser does significantly better overall at standards compliance than Firefox?

Well, since the link you provide is largely question marks for the Webkit based browsers, that's hard to say. Also, the comparison you link to is missing a lot of standards where Firefox is a bit behind. These include:

• Javascript - Safari, Opera, and Konquerer all have at least some support for Javascript DOM 3, which Firefox lacks in the released versions so far.
• image formats - Konquerer supports MNG, Tiff, and PDF. Safari supports JPEG 2000, Tiff, and PDF. I know of no standard image formats Firefox supports not supported by both of those (yet).
• XHTML 1.1 lists Firefox at 63% and question marks for Safari and Konquerer, but wikipedia currently lists both of those as having "full support" and Firefox as "partial."
• Web Forms 2.0 - Opera supports, Firefox doesn't
• Voice XML - Opera supports, Firefox doesn't
• WML - Opera supports, Firefox doesn't

That is not to say Firefox is necessarily behind other browser for standards compliance in general. No one with a clue would cite the Acid tests as proof of anything in that regard, but it does indicate that the link you provide is not particularly strong evidence one way or another. The whole question is probably too vague to be answered. There are a lot of Web standards and what really matters is which ones are most universally supported and what functionality cannot be used because of lacking support in one browser or another.

In summary, I reject your assertion, not because I'm convinced you're wrong, but because you haven't provided enough evidence to support it and there is significant contradictory evidence (cited above).

Re:Hmmmm

[WK2]

Firefox is an absolute memory whore.

OMG WHAT ARE YOU DOING TO YOUR BROWSER!?!? And how? Whatever it is, I don't think Firefox actually wants your memory that badly.

On the other hand, perhaps you meant, "memory hog."

It's a trap!

[tobiasly]

It's a trap! First Microsoft lures us all into using interoperable web standards, and then... then.... shit, I can't figure out how they can use this for evil. Gimme a sec...

Re:I don't care about IE at all

[99BottlesOfBeerInMyF]

Which, in actual terms, means that people code to Firefox just as they code to IE. It just so happens that coding your page to look right in Firefox is a helluva lot closer to the standard (if not it exactly) than when you do the same in IE.

I disagree. At my last employer I used OmniWeb for a while (a very niche browser). Most of the Web UI developers used Firefox, but a couple used Konquerer. A few used Safari. A few used Camino. A few used Opera. Regardless of what you used, when you found a bug, you tested it with a couple of other browsers and if the remote Windows box was available (or you had an emulator running), you tested it on multiple browsers and multiple platforms.

The upshot of all of this was, when a bug was listed, it was pretty easy to see which bugs were specific to a given browser. Bugs that appeared in some version of IE, but in no other browser at all, were by far the most common occurrence. Realistically our approach boiled down to, "write to standards; then hack for IE. " Make no mistake, we did not code for some other browser then try to make it work on every one, because that was not needed for the most part. We were programmatically generating Web pages and interfaces from XML data and a couple of databases. For the vast majority of the time, all browsers but IE were close enough to the standards we used (HTML3, CSS2, XHTML) so that there were no discrepancies when tested.

Developers & the half-life of accumulated cont

[weston]

Developers, developers, developers, right?

I think Microsoft has finally genuinely started to realize a very simple fact:

Client-side web developers hate them.

And it's probably the one thing MS has thoroughly earned with all the IE bullsh*t over the last 10 years.

This is a really great gesture, it's a good start if they want to allay any of that and gain back trust. But honestly, nobody gets over 10 years of being treated like crap overnight, and the half-life of contempt isn't short.

Personally, I'd like to offer my congratulations to the IE Product management team, and let them know that in time, I'll probably only wish debilitating terminal illness on them, rather than painful and extended death by torture.

Re:This Will Cost MS Dearly

[tobiasly]

I can't for the life of me figure out why MS would want to put their neck on the line like this

You must not have read the press release!

"While we do not believe there are currently any legal requirements that would dictate which rendering mode must be chosen as the default for a given browser, this step clearly removes this question as a potential legal and regulatory issue"

They aren't putting their neck on the line... it's already there. :)

Improved standards isn't the story here

[trepan]

The real story here is that "Developers wishing to use quirks mode for IE6- and IE7-compatible rendering will have to opt in explicitly."

If you've been following any of the design / developer blogs and community response about this, you'll know that in a previous plan, all web pages would render in IE7 standards mode unless the developer inserted a specific meta tag

<meta http-equiv="X-UA-Compatible" content="IE=8" />

into each web page of a site. (For the truly avant garde, one could set the content to "edge", which would tell IE to render in the most current standards compliant version available). The outcry was that while it was clear that IE was making progress in standards, in order to take advantage of those improvements, developers were being asked to touch each page of their sites and tell IE to use its more standards compliant mode. That discussion is what was at play here.

Re:Hmmmm

[Your.Master]

You just totally missed his point. If, hypothetically, IE8 is in some way better than Firefox on Windows, Firefox will have adapt to compete. This will help Firefox on Ubuntu, because Firefox is competing with IE in the marketplace, even if it is not competing on your OS of choice.

Re:I don't care about IE at all

[Khuffie]

Not sure about your part of the world, but where I live, quitting a job due to having to use IE for a payroll site is...stupid. Choice of employment is not up to the individual. If it were, I'd be "that-guy-that-sits-on-his-ass-playing-videogames-and-getting-paid-millions".

Acid3

[drewness]

Just for fun I tried Acid3 with a couple browsers (all MacOS 10.4.11):
Firefox3 nightly from March 3rd: 66/100. (Second closest to the reference rendering.)
Safari 3.0.4: 39/100.
Opera 9.26: 46/100. (Looked the least like the reference rendering though.)
Webkit nightly from March 4th: 87/100. (It also looked the closest to the reference rendering.)

Re:I don't care about IE at all

[mike_sucks]

No no, it's a hat made /out of/ zealots.

Comfortable and quite a nice looking number, too.

/Mike