Microsoft Insists IE7 is Standards Compliant

ReadWriteWeb writes "Microsoft's Chris Wilson, the Group Program Manager for IE addresses the issue of whether IE7 is CSS and Web standards compliant. Last week a Slashdot post claimed that IE7 was basically non-compliant with CSS standards. But Chris Wilson says that isn't true and that standards improvements is a big part of IE7. He admits that there were a ton of bugs from IE6 that have caused web developers a lot of pain, but says that IE7 will address those and be standards compliant. He goes as far to say that IE7 supports Web standards even at the expense of more backwards compatibility."

Comments

Acid Test

[celardore]

I wonder if the browser will pass the Acid Test....

Re:Acid Test

[Anonymous Coward]

Passing the acid test would be a nice feat, but not even everybody's darling, the Gecko browser family, passes it. And it isn't really necessary. IE7 does most of the things right that webdevelopers have always wanted to use but couldn't because IE6 owns the market. The box model implementation is sane now, IE7 does semitransparent PNGs (without requiring non-standard hacks), it supports :hover on everything, and so on. I've recently designed a new site from scratch and tested only with Firefox. Then I loaded the site with IE7 and I didn't have to change a bit. Right now the deficiencies of CSS itself are the biggest hurdle, second only to the legacy browsers.

I'm mad at Microsoft for leaving us in the cold for so long, but even though I hate the IE7 user interface, I think the rendering engine really is good enough. Just make sure that IE7 gets pushed to each and every IE6 user out there. No bullshit like restricting it to Vista or XPSP2 please.

Re:Acid Test

[Anonymous Coward]

Even if it weren't biased, I think we can agree that a browser shouldn't be judged by its ability to render broken documents correctly. As a webauthor I can make sure that my pages are correct. It doesn't matter to me how a browser treats incorrect pages. There should be a separate test that only tests valid code, because if a page is invalid, you shouldn't rely on standards compliant browsers to make things right. Fix the damn page.

Standards Compliance at Cost

[salesgeek]

One major issue is that many sites do not render as nicely in IE7 as they do in IE6. This is going to be a headache for IT managers and marketing managers for quite some time...

and for the love of money, think of all the FrontPage sites...

The irony

[Bullfish]

The irony is, that whether you like it or not, when you control over 80% of the browser market, you are the standard. That they are willing(?) to try to accommodate other standards, is really a sop.

Re:The irony

[richdun]

Except that it isn't over 80% any more. The latest stats I saw (July 2006 I think) put IE at 73-75% and falling - still very high, but not nearly as dominant. Certain markets (universities, for instance) have much, much lower rates.

IE7 may change that, as many recent Firefox converts may switch back when it comes through as a security release. The real wildcard though is just how much marketshare Apple is really capturing - IE will never again be available for Mac, and if they (Apple) are to be believed, they had something like 15-20% of the laptop sales marketshare last quarter (or month...too many stats!), and are growing. It may be a case of too little, too late, but with Vista and Leopard we could see a swing in browser marketshare not seen since IE trounced Netscape.

Too easy to debunk

[erroneus]

Someone, or more likely several someones, will independantly enumerate every area of non-compliance that exists in MSIE7. (Has it been released yet? I haven't seen an installation for Linux yet... I have MSIE6 on my Linux laptop thanks to some very clever script writers: http://www.tatanka.com.br/ies4linux/index-en.html)

That said, I have read where even Firefox isn't yet 100% compliant. I'm usure of how much difficulty that causes web developers though. Actually, I don't know much of anything about the web except that I use Firefox pretty exclusively. If MSIE7 was made at least as compliant as Firefox, it would actually kinda bother me as it would give me a lot less leverage to keep my Firefox deployment where I work.

I can't use it anyway

[rucs_hack]

My one windows machine (authentic windows, purchased and everything, from HP), fails when trying to install IE7 beta.

It passes the genuine disadvantage test, then b0rks for an unknown reason.

Firefox, on the other hand, is perfect, so I don't feel it matters much anyhow. I only tried to install IE7 out of curiosity

Re:cut MS some slack

[ElleyKitten]

recommend anything but IE7 to anyone for many reasons

IE7 is about a billion times better than IE6. I don't know how anyone can stand it, yet I find IE7 rather tolerable for those few, IE-required situations. If someone ignores you when you talk about Firefox, you should recommend them IE7. Anything is better than IE6.

Fixing the fix

[RyoShin]

He admits that there were a ton of bugs from IE6 that have caused web developers a lot of pain, but says that IE7 will address those and be standards compliant.

By this, he doesn't mean that they're fixing the compatibility- far from it. By this, he means that they're closing up the holes in IE6 rendering that previously allowed developers to "hack" around IE. Since IE6 would or wouldn't recognize something it shouldn't or should, respectively, one could make one style sheet that contained CSS for both IE and FireFox, using various methods (such as inheritance) to hide CSS from IE. (I believe we've already had articles about IE7 Beta breaking websites created for IE6.)

(Granted, the best way to do it is to set up a broswer check and use a different CSS file for each browser. But when you have a tiny website, you don't really care to futz with it.)

This effectively means that when IE7 comes out, all the hacks made for IE6 will break, and many pages created by that "cousin in high school" will suddenly look like rubbish.

Of course, those that were made predominantly for FireFox and Opera will still continue to work unabated.

They are bowling googlies again!

[140Mandak262Jamuna]

One thing that the Trident engine that underlies Internet Explorer has had for many releases is editing support. A number of products have been built on top of this editing support in the past and it's quite a strong piece of our underlying infrastructure.

This is the key folks. So many corporate database products rely on IE as the rendering engine. If the backward compatibility is lost, most corporations' will see their Crystal Reports, and other SQL engines that use IE as their GUI/renderers will be broken. They will never allow that to happen. So they will sacrifice the standard compliance.

Of course they will claim their concern is the "not spoiling the user experience" of their old moms or breaking millions of websites. But the real concern is that all these products should continue to use IE as their rendering engine. Their hold on corporate desktops through MS-Office and IE is too dear and profitable for them to compromise.

Re:"no official CSS test suite"???

[_xeno_]

That's actually kinda neat. So far, of the seven tests I've run, Firefox 1.5.0.6 has passed one of them.

For the curious:

• float 0 - Fail
• The Height of Lines 0 - Fail
• The Height of Lines 1 - Fail
• The Height of Lines 2 - Fail
• The Height of Lines 3 - Fail
• Inline Elements 0 - Fail
• vertical-align 2 - Pass

Some time when I have more time, I'll have to go through all of them and see how Firefox does.

Quote from his blog

[kripkenstein]

In the meantime, Microsoft almost seems tentative in their position about standards compliance versus backwards compatibility.

Emphasis mine, changing the meaning a bit, but bear with me. If you read Chris Wilson's blog here, then you can see the following quote:

It's been frustrating, though, to be continually identified as the personal screw-up responsible for IE not supporting more standards today, when it's actually because of my personal influence that CSS is IMPLEMENTED in IE.

Again, emphasis mine (not the caps, though, just the boldface). So - if it weren't for this Chris guy, CSS wouldn't even have been implemented in IE. If he's right, that says a lot about Microsoft. I tend to believe him here.

Re:Expanding Box Bug

[rjstanford]

Actually, I just visited that link with IE7, and it rendered the page identically to Firefox. So yes, they have fixed that bug.

Re:Expanding Box Bug

[jhildo]

I just tried all of the examples on the page with IE7 that is part of 5472.winmain_idx01.060713.1900, and it appeared to render everything correctly. .Correctly meaning what the author expected the render to look like. I'd say that means IE7 has fixed the bug.

Re:Doesn't matter how complient they are

[abigsmurf]

Sorry I call BS on opensource browsers only taking a few months with a dozen devolopers to get to the levels of usability and compatibilty they are now.

The most used browsers have been mostly around for years AND almost all of them have been based off of existing code, whether it be konquerer or Mozilla. The only one I can think of that wasn't was Opera and that was designed from the ground up specifically to meet standards.

Re:Who cares?

[Anonymous Brave Guy]

I'm not saying there's no reason for those aspects of the test for anyone. I am saying that they are basically irrelevant to me as a practising web developer who validates the XHTML and CSS of all my pages using W3C tools anyway. Thus (as a web developer) I care very little about whether any given browser passes Acid2, because the non-compliance aspects will have no significance for the pages I produce.

Now, as a user, I might care more if it turned out that a browser was doing something particularly daft with bad input. But since I don't often find myself running into the sorts of things covered here with any browser today, W3C-friendly or otherwise, I just don't see this as a big problem. When there are still pages that just don't render at all sensibly on popular browsers like Firefox and Opera, and minor layout niggles on numerous sites as a result of things like box model gremlins, I'd far rather the IE team spent their time working on fixing those than some hypothetical problems that I've rarely, if ever, encountered.

Re:cut MS some slack

[russx2]

Why not just go on doing the same old, bad job on the pages rendered on quirks mode, and then render correctly and compliantly on pages that specify proper DOCTYPEs etc.?

Yes, it's true IE6 has a quirks mode and 'standards' mode, deciding which mode to use based on a valid doctype or not. However, the problem IE has now is that by following standards more closely in IE7 they potentially break compatibility with IE6 'standards' mode. Pages without a valid doctype can still be rendered as always by the quirks mode so they are not the problem.

Most of the problems will stem from all the inventive IE-targetted CSS hacks out there - tan hack, holly hack, star-html hack etc. that abused IEs improper understanding of CSS rules and will potentially break in IE7.

Hmmm.

[jd]

Objective figures. Ok, here's the way to come up with an objective figure:

• For each non-interacting standard defined...
  
  • ...For each defined non-interacting element in that standard...
    
    • Test that element with each individual attribute definable for it, with each well-defined corner-case and with a random selection of invalid cases.
      
    • For every valid case that tests correctly, score 1 for complance. For every invalid case that is rejected safely, score 1 for compliance. For every test that causes the browser to crash, score 1 for instability. Unimplemented cases are always treated as non-compliance, even when optional. (The instability value starts at 1, not 0. A stable browser doesn't have infinite quality.)
      
    
    
  • ...For each defined interacting element, repeat the above test with typical, corner-case and invalid combinations that test every element - but not necessarily every permutation - at least once in combination with another element that it can interact with.
    
  • ...For a random selection of totally invalid tags, repeat the above test with a selection of short and excessive invalid tag sequences.
    
  
  
• Normalize the results by dividing the totals by the number of tests that have been executed and multiply by 100 to convert to a percentage.
  

For each interacting standard, apply the above test program for typical permutations and corner-case permutations, such that all interacting standards are tested at least once in combination with another standard that it can interact with.

Sum up the totals and divide by the number of standards and standard interactions tested.

Divide the total compliance by the total instability to get the overall quality.

Calculate the theoretical values that would be obtained for a browser that met only the required elements of the specification, as a fraction, to get the compliance threshold value. Determine the ratio of the total compliance with the compliance threshold to get the baseline compliance.

The overall quality of the browser will tell you how reliable the browser is, when trying to follow the standards as defined. The baseline compliance will tell you how close the browser is to meeting the obligations of the specification. The total compliance will tell you how close the browser is to meeting the full specification.

It's a simple enough algorithm and is based on the usual testing procedures used by a million software engineers the world over. You test the typical, the corner-case and the error cases. In any specification, these cases are well-defined and should be easily tested.

Do these numbers mean anything? Yes. Due to the sheer volume of specifications out there, it is impossible to physically list every permutation that needs to be validated, but you CAN say what fraction of those permutations have been validated.

A superior method to this is to use an octal mask, where the value of each position represents the number of permutations (up to 7) that have been tested against a specific element, and each position represents one element. If you want to interpret half a screen of octal, go for it. It will give you more information, if you can process it, but will tell you less than the three suggested numbers will tell you unless you're prepared to do a lot of data crunching.

Re:cut MS some slack

[jawtheshark]

You're doing something wrong: I've converted all of my family including my wife to firefox. They are used to it at that point that they ask me "Hey, Jts, that webpage doesn't show right/do anthing". Then I surf around for a while, remember there is another browser and load it up in IE. Then it works. Happens about 2x a year, for all my users....

Firefox has become their browser, IE isn't even in the picture anymore.