One of the most exciting features we were expecting from Internet Explorer 8 was compliance with web standards. Months ago, Microsoft announced that IE8 would render the Acid2 test properly. So imagine our disappointment when we installed IE8 beta 1 today only to get an image that looks like the one to the right.Unperturbed, we tried again later, and the second time was a charm. IE8 passed the Acid2 test with flying colors. So what was the difference between attempt 1 and attempt 2?
The second time, we went to the official Acid2 test page. The first time we had gone to another site that was hosting a copy of the original test. While you'd think that a web browser that's capable of rendering one page properly should be able to handle the same content on another page, the folks at Microsoft's IEBlog point out that IE8 performs a cross-domain security check for ActiveX controls which will prevent it from passing the Acid2 test on any page but the official one.
So there you go, Internet Explorer 8 is standards compliant. Under the right circumstances.
Reader Comments (Page 1 of 2)
3-05-2008 @ 9:43PM
lauph said...
wow. isn't that just like a car that only works in the garage?
Reply
3-05-2008 @ 11:07PM
michas_pi said...
The scroll buttons on that image kinda, sorta look like eyes, right guys?
Guys?
Reply
3-05-2008 @ 11:07PM
Brian said...
I guess the real question is, why is the "standards mode" an ActiveX component in the first place. I like the idea that all ActiveX components require security to run, but I'm sorry, "standards mode" should be so fundamental to the browser, it should be impossible to disable it.
Reply
3-05-2008 @ 11:07PM
Marshall said...
This sounds like the "teaching to the test" methodology that makes so many people upset about education today. Great for statistics, totally worthless in the long-run.
Reply
3-11-2008 @ 8:50AM
Kris Thom White said...
The comment about "teaching to the test" isn't new for Microsoft. Way back when Word 6.0 was new (and I was an MS employee), one of the demonstrations of Word 6 was its WordPerfect importing capability and how "accurate" the imported document was compared to the original WP copy. WP came with this standard "demo document" at the time and we would import that document and it would be perfect in Word.. but an engineer told me that Word "recognized" the document in question was this standard test document for WP and there was "extra code" in the Word program to make sure it showed correctly. Results with your own WP documents, however, may not have been quite as accurate.
3-05-2008 @ 11:14PM
Tom said...
There is no activex. There is only one object that should not trigger activex:
There is no way this object should trigger AX, and my guess is it's exactly like Marshall said, it's designed only to work on the official test because they didn't really fix the problem.
Reply
3-05-2008 @ 11:20PM
Tom said...
Please link to this as well:
http://blogs.msdn.com/ie/archive/2008/03/05/why-isn-t-ie8-passing-acid2.aspx#8059324
It's a comment by Ian Hickson (designer of the Acid2 and Acid3 tests) saying it's still a bug.
Reply
3-06-2008 @ 7:58AM
kim israel said...
whew!!! try 2 view adobe website on ie8 beta...
Reply
3-06-2008 @ 9:05AM
mario said...
I tried Acid2 in Firefox (2.0.0.12) and the result was worse than your picture, tried IE7 and it was even worse.
Then tried Opera 9.24 (build 8816) and it looked just like your image.
I don't get it, are browsers supposed to show an undistorted happy face?
Reply
3-06-2008 @ 9:14AM
drewls said...
I tried Firefox and IE7 on that happy face thing and it's messed up in both. What gives?
Reply
3-06-2008 @ 9:23AM
el_bob said...
"prevent it from passing the Acid2 test on any page but the official one" sounds suspiciously like "make it so that it will pass on the official one when when it normally wouldn't elsewhere".
does this mean that all web pages have to be "official" pages for microsoft to render them properly?
Reply
3-06-2008 @ 9:40AM
quanta said...
Acid2 is basically a torture test with different web standards, so there is variable scoring from 0-100%.
Firefox 2 only partially passes Acid2 (it does handle it much better than IE7). Firefox 3, however, passes 100%.
Reply
3-06-2008 @ 10:31AM
kingkool68 said...
A standard compliant IE8 is nice and all but web developers are still going to have to make sure their pages look good in IE7 and possibly IE6 still for years. Until IE8 becomes the dominant Microsoft browser, it might as well not exist from a developers perspective. I will be using the IE7 meta tag for all of my websites that simply don't work in IE8.
Reply
3-06-2008 @ 2:40PM
JardaP said...
Maybe IE8 just displays a PNG picture embeded in the code, when it detect the right link to the ACID test. :-)
Reply
3-06-2008 @ 3:07PM
JardaP said...
Maybe IE8 just displays a PNG picture embeded in the code, when it detect the right link to the ACID test. :-)
Reply
3-06-2008 @ 9:33PM
michael said...
This is just a BETA version of IE8, you know. There could be a bug. I heard some people said it wouldn't work, while others claimed it did.
I'm going to wait until it's more finalized to give a call on Microsoft's promise to be more standards compliance.
Reply
3-06-2008 @ 11:51PM
Tom said...
I take it Firefox works well on both sites?
Reply
3-07-2008 @ 7:59AM
HardwareGuy said...
The Acid test is just a browser pissing contest. If standards compliance meant a damn everyone would be using Opera.
Reply
3-07-2008 @ 7:59AM
Tycho said...
Well, it's not quite standards compliant. It fails Acid 3 miserably. But more interestingly, Internet Explorer 5.5 beats IE 6 and 7 in the Acid 3 test.
http://www.anomalousanomaly.com/2008/03/06/acid-3/
Way to go, Microsoft.
Reply
3-07-2008 @ 12:55PM
James said...
It sounds like IE implements this *better* than the standard -- and since I always assume that IE makes the worst possible choices in all cases, this is pretty surprising to me.
For those who can't be bothered to read the MS blog explanation: the eyes are rendered using an embedded object tag with bad data, and IE8 by default will not allow an object to be pulled from a different domain than the page being rendered. I don't see how that's being billed as an "ActiveX" error (I thought ActiveX was dead now anyway?), it's just a perfectly reasonable thing to do, IMHO.
The point is, the "mirror" of the ACID2 test is in fact a substantially different page than the actual original test, because one points entirely to content on its own domain, and one points to content from a different domain. If they pulled over the original test and fixed the URLs so they were consistent (e.g. pointed to the mirror's domain), I suspect it would render the same as the original.
Reply