Microsoft Insists IE7 is Standards Compliant 389
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."
"no official CSS test suite"??? (Score:5, Informative)
Re:Acid Test (Score:3, Informative)
Re:Acid Test (Score:5, Informative)
"I said on the IE Blog that in IE7 we were not going to pass the Acid2 test"
He goes on to note that a number of the things used in the acid2 test are to not likey to be high on their priorities and would be focusing on more widely used CSS.
Hilarious (Score:4, Informative)
Their html editing control is crap crap crap. I'm talking about the control thats been used in Outlook 2003, MSIMN/Outlook Express etc, I assume the interviewee is too.
* It is very easy to get paragraphs that are indented to the right. Yet it can be absolutely impossible to remove the indentation and align the paragraph with the rest of the text in the email. I suspect it barfs when it has to deal with nested tables.
* Deleting some text or formatting can drastically alter the following paragraph.
* You can read in perfectly valid html then it refactors it into gibberish.
Anyway its absolutely effing hilarious that they think its a strong html editor control.
Re:cut MS some slack (Score:3, Informative)
It indicated something like ~60% for IE, approx. 90% for Firefox, and most for Opera.
Unfortunately I don't recall the URL, so that's the sloppy figures you'll get from me.
This post's summary is completely stupid (Score:2, Informative)
The usual crap by slashdot editors...
Re:Too easy to debunk (Score:3, Informative)
http://www.webdevout.net/browser_support_css.php [webdevout.net]
Translations from the managerese (Score:5, Informative)
"In IE7 we really are trying to support Web standards." Translation: we are not committing to being compliant with Web standards.
"We certainly spent a bunch of work trying to improve our standards support." Translation: We're over budget on standards support.
"I don't think we're at 90%, I think we're above 50% though." Translation: we're not compliant.
"Well as you saw I got a little frustrated with the Slashdot post." Translation: I can't point to factual inaccuracies in the Slashdot post, but I sure don't like the spin.
"The target for that was not just passing any one particular test." Translation: We don't pass that particular test.
Re:Acid Test (Score:3, Informative)
Re:"no official CSS test suite"??? (Score:5, Informative)
No. Not even close. Throw away the idea that he's arguing that Internet Explorer 7 is standards compliant. It's a complete fabrication; he never claimed that.
What he is saying is that they've done a lot of work in the area of standards compliance, there are moderate improvements, and that it doesn't really make sense to say that it supports 57.324% of the specification or whatever kind of number you can come up with, because there's really no sensible way of measuring something like that objectively.
Chris Wilson isn't a marketer, either. He's worked on Internet Explorer for years, he was on the W3C CSS working group and has his name in the acknowledgements of the specifications. I believe him when he claims to be working hard to bring Internet Explorer into compliance, but there's only so much a person can do without support from above.
Re:The problem is Microsofts creation (Score:3, Informative)
Every web developer must make a choice in the beginning which browsers he/she cares to support. IE, Firefox, Mozilla, Opera, Safari, Konqueror... etc. They all render differently. And different version of all of those render differently. However, standards compliance means you can at least depend on some things working all of the time. If you just pick one of those, no matter how big the market share you are shooting yourself in the foot. And IMNSHO, you'd deserve it too.
Re:Acid Test (Score:5, Informative)
Please do not spread this myth. It is simply not true. If you had actually read the Acid2 technical guide instead of relying on Slashdot hearsay, you would know this. From a previous comment of mine [slashdot.org]:
More like 6 out of 7 (Score:5, Informative)
Acid2 Test (Score:4, Informative)
Re:"no official CSS test suite"??? (Score:5, Informative)
A test suite cannot tell you if an implementation is compliant
Yes, it can. If the question is boolean, then the test pointed to by grandparent definitely can give an objective answer. Currently, every browser I know of would fail, but it can give an answer.
or to what degree an implementation is compliant.
Yes, it can. It can give you a consistent answer on the number of passed and failed tests. That number may be biased for a given single run, but it can give a consistent answer, so it can be used to test relative compliance. It may give a fuzzy response that is open to interpretation, but I would bet that when testing different versions of the same browser, the answer would almost always be clear. It would be something like this:
IE6: Passed; 77. Failed; 39.
IE7: Passed; 92, Failed; 24.
IE7 Relative to IE6:
Passed in IE6, Failed in IE7; 2.
Passed in IE7, Failed in IE6; 17.
"One of the things I said in my post is that I think it's very difficult, if not impossible, to have an analysis of exactly where we are as a number with supporting or complying with CSS"
Actually, it's easy. There's an example of the numbers above. Are the numbers fuzzy? Yes. Does it provide, "an analysis of exactly where you are"? Yes. It is, "an analysis" and it is based on "exactly where you are." Is the result fuzzy? Of course. Every test of every human endeavor in history has either had a loose question or given a fuzzy result. Now stop being a sissy and answer the damned question.
"we really only did standards improvements - particularly CSS and HTML improvements. That was really the largest focus of our platform work overall."
If that is the case (though when someone uses "really" twice in two sentences without providing supporting evidence I think he doth protest too much), and this guy really is involved in the project, then he should have those numbers tatooed on his inner thigh. If CSS and HTML compliance really was the largest focus of the largest software company ever, he would at least be able to say something. He would be able to say, "I don't think these numbers are perfect, but here's what we get on the official test suite."
If MS really were focusing on those tests, even if he really believed that taking number passed over number failed was such a great injustice, they would have those numbers printed in 120 point font and hung on the wall of the developer area. He could have said, "Well, I can't give you a percentage, because percentages are inherently subjective - they weight every test the same which does not necessarily reflect the value of each test to the total user experience. And though the results are subjective, the tests are objective, and we track them like a hawk. Hey - we're the biggest software company in the world. Software loves tests. So, while I reiterate that these numbers are't perfect, I can tell you that IE7 now passes 17 tests that IE6 did not, and failed 2 that IE6 passed, using the W3C's standard set of CSS 2.1 tests. You can get the specific list of test results at www.microsoft.com/ie7-css-test/."
Or he can say none of that, and I will remain unclear on Microsoft's current view on standards. I won't claim that I know them to be as bad as they have been, simply that what this guy says is sound and fury signifying nothing. Which is all that GP was saying.
He claims that this particular course is their "largest focus", a course which they have repeatedly faltered from, with apparent intent, in the past. He says there is no way of providing a number, without even acknowledging the existence of the official set of tests. Not even to say they suck. Either he doesn't know they exist, or he doesn't feel they are important, or he feels the results would leave the audience nonplussed. The one thing we can say for sure is not the case is what he is implicitly claiming; that they are deeply interested in passing the tests, that they always know the results for the latest build, and that they are proud of their accomplishments.
Re:Acid Test (Score:3, Informative)
Flip side is that every conformance test suite on the planet does stuff like this. The errors generated by mistakes are at least as important as the ability to work with valid input.
To take your C compiler example, if you wrote C code that left off every semicolon, you would not expect the compiler to say "oh, I'm going to add semicolons where needed". You would expect it to generate an error message. If, instead, it inserts semicolons where needed, it SHOULD fail the test because it will confuse the heck out of people trying to write and debug code.
In a similar way, detecting bad CSS and behaving in a consistent way is at least as important as behaving consistently for valid CSS. If one browser accepts slightly invalid CSS, that browser is no longer useful as a mechanism for testing your page because some other browser will choke on contents that this browser displays "correctly" (as intended instead of as written).
Re:More like 6 out of 7 (Score:5, Informative)
It would have been really nice of the W3C to mention that somewhere in the test suite. Unfortunately they neglect to mention that in the test suite, on the test suite "home page," and in the test suite documentation. They even neglect to offer a full download of the suite (you can use wget for that).
Even their own test suite documentation [w3.org] neglects to mention the font requirement. In fact, the only place they mention it is on the test authoring guidelines [w3.org] which is not something I'd expect to read when just running tests.
If I get the time, I think I'll try and bundle the test suite up WITH the Ahem font and try and run it again. Maybe even use the CSS font-embedding extension so you don't actually need to install the font.
But I can see why people may not like that test suite. Without a "full suite download" it's a bit of a pain to use.
Re:Acid Test (Score:5, Informative)
You would if the C standard explicitly required compilers to do so.
The fact is, it doesn't make much sense to compare CSS to C. One is an imperative programming language, the other is a declarative style language. You can't miss out bits of a C program and still have it work right, but CSS is designed around the idea that you can do just that.
I don't want to include a CSS 3 property in my stylesheet only to have every CSS 2 browser simply throw everything away, I want them to apply the CSS 2 properties they do understand and ignore the bits they don't. Fortunately, this is exactly what the CSS specifications require, which is why the Acid2 test includes invalid code - it's testing this part of the specification.
Re:Acid Test (Score:5, Informative)
Actually, gecko does pass it [nelchael.net]. The problem is that firefox 2.0 won't use that revision of the gecko core, only 3.0 will use it.
Now, even if current Firefox and future firefox 2.0 are not passing it, they're NEAR of passing it. IE7 rendering does not even look like a smiley [google.com].
I think the rendering engine really is good enough
Yeah, the software company number 1 of the world should be proud of shipping a widely used browser (IE is the most used application in the world) whose rendering engine is the worst one in the world, but that is "enought" only because IE defines what is "enought". If Firefox had 80% of market share, web developers would use lots features that IE does not even dreams to support until they ship IE8 in a couple of years. And nobody would use IE, because their engine is NOT "enought".
Firefox and Acid2 (Score:3, Informative)
Here's a good run-down of Acid2 status in major browsers [howtocreate.co.uk]. According to that, a "reflow" branch of Gecko alread passes the test, but the changes haven't been fed back into the trunk.
In short:
Safari: Passed
Konqueror: Passed
Opera: Passed
Firefox: Working on it, should be two releases away.
Internet Explorer: Ignoring it for now.
Iexplore my hd (Score:2, Informative)
Re:Acid Test (Score:3, Informative)
In my tables, I try to accurately describe exactly what features are handled incorrectly under which conditions. The tables are very much laid out as the features are in the specifications and therefore I don't see any legitimacy to his argument that I shouldn't note IE's lack of "inherit" support on every applicable CSS property. I maintain a complete public log of every change made to the information and you can get an RSS/Atom feed on it. If he believes that there is bias in my tables (aside from the fact that all features are weighted equally regardless of real-world usefulness, which is done to avoid bias), he should say exactly what the problems are rather than falling back on an ad hominem response. I made my tables to be useful to web developers and researchers, and I certainly don't want any bias in them.
I wrote the "Internet Explorer is dangerous" article mainly in response to IE's obviously poor standards support, as well as their poor record of fixing security issues. I don't have any specific anti-Microsoft agenda, but rather an anti-outdated-software agenda. As long as IE or any other browser with significant market share is seriously behind the rest of the major browsers in standards support, I will call it out in the interest of fair competition and progress. But if Microsoft pulls a miracle and makes IE even close to as standards-compliant as Firefox and Opera, I'll gladly remove the article and instead just encourage people to upgrade.
Re:cut MS some slack (Score:4, Informative)
At the time, I installed a late IE4 beta on NT 4 Workstation. After the obligatory shutdown it never booted again and I had to reinstall NT - this was before the days of Recovery Console or any of the nice rollback stuff that's present in XP. I figured if that was the kind of quality Microsoft considered late beta, Netscape had nothing to worry about. Boy was I wrong.
Re:Firefox failed the test (Score:2, Informative)
This indicates to me that the "uninstall" routine only changes the interface, not the upgrades it did to your OS in the background.
The moral of this story: All software has beta (or alpha) in the title for a reason and should only be installed on machines that do not actually need to work on a 24*7 basis.
(oh - and IE7 also fails the acid test and does not generate a neat smiley face)