Just when Microsoft thought it was on target with its forthcoming Internet Explorer 8 browser, the goalposts have moved. The Web Standards Project (WaSP) has released its latest browser standards compliance test - Acid 3 - and every browser that WaSP tested failed. IE 8 is, of course, not available for test yet. But given the …
Opera made it to 43 the first time I tested, crashed the browser each time after.
As someone who writes a browser ...
.... I know of a couple of simple way to improve standards compatibility.
1) No standards without reference implementations or even closed source proven implementations, that means you W3. If you can't check the validity of your own specs in running code, why is everyone else supposed too?
This isn't rocket science, it's the method that IETF has used with its RFCs for years.
2) No standards without full test cases, when your standards feature non-local user content, the specs need to define the failure behaviour as well.
Since the mid 90's the W3 has made over 90 specifications available, maybe they should spend a little more time checking they really work.
While i welcome more rigorous web standards, surely the fact that no browser can pass this test makes it a rather redundant standard or more of simply a test than a standard.... or maybe i am being picky about the wording, but it smacks of the same lack of meaning as unlimited bandwidth thats limited!
Why Is This Standard So Hard To Comply With?
I like a good pile-onto-Microsoft as much as the next Linux user, but it seems to me that something is actually broken with the standard, if it is so hard to comply with that the Acid* tests consistently flunk every browser. What's the point of writing a standard if programming to it invariably requires a Moon-shot of a development effort?
Written by a fanboy
The article is just a Firefox fanboy rant. It makes no mention of other browsers such as Safari and comes months after Acid 3 appeared yet just hours after MS announced IE8 would pass the Acid 2 test by default. So it's just a desparate attempt to bash IE because it's getting a little too close Firefox. Afterall, Firefox currently doesn't pass Acid 2 either and if IE8 gets out before Firefox 3 then it'll be in the strange position of being the only common browser *not* to pass Acid 2. Other browsers have passed it for some time now, and I'm talking about official versions, not obcsure custom builds used by 3 people or internal prototypes of future versions.
So, if Microsoft "has to go back to the drawing board with IE 8 and start over" because it doesn't pass Acid 3 then it's a shame you couldn't say exactly the same of the highly profitable Mozilla Corporation because not only does Firefox not currenly pass it either, but it's trailing behind two other browsers in completeness. Firefox may be better than IE, but these days that's starting to become a rather literal statement - It's better than _only_ IE.
Don't even try Konqueror
...on my system (OpenSuse) ACID3 crashes it
I've always wondered
How do they test that their test works?
Have they got a super top-secret 100% compliant browser?
After all, you wouldn't release an untested piece of code then say, "it's your fault if it doesn't work, because we think it should even though we haven't tested it / can't test it", would you?
Bill, because he just might.
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9b3) Gecko/2008021416 Firefox/3.0b3
GOT 60 Out of !00
I'm sure this will be mentioned 500 times before this comment gets approved, but my beta of Firefox 3 gets up to 70/100.
IE 8 Beta 1
17/100 stutteringly, but much better than IE 7's output.
Not that ACID means owt
Opera, W3C vs Industry
Opera got to 46/100 on my system.
Given that Microsoft has so much of the browser market, why don't they have more say about the standards? Who says a committee should simply make rulings that pull the rug out from under 80 percent of the browser share?
Here is one example. W3C clearly hates the old <font> tag, and its official definition appears designed to herd people way from using it. In particular, the tag's scope doesn't go inside tables. IE and Opera both ignored this new "standard" because it broke a ton of old web content. Firefox obeys the standard and a lot of hold HTML 0 pages look like crap. So who's fault is that, Opera/Microsoft or W3C?
Standards committees can get out of control fast, especially if they are populated with academics and bureaucrats, instead of engineers and people who are in touch with users. Look back at the OSI debacle, where Eurpean committees invented a networking standard so complex and untried that it simply collapsed and never worked. We don't want this to happen to the web.
My version of the excellent editor known as vi renders HTML better than emacs.
Isn't that the point?
The whole idea behind the Acid tests is that it tests things that are currently badly implemented - so there's clear info on what needs done. If browsers worked successfully, there'd be no need for the test.
The latest Safari/WebKit nightly build gets 89/100 up from 65 a few weeks ago, so they're almost there :)
Presumably, the point of this test...
...is to make every browser fail all the time, thus giving policy dorks a constant supply of axes to grind? And if no browser passes the test, what the @#@# do they look at it with to test IT?
opera does better
Opera 9.50 beta gets 60 out 100.
FYI IE8 Beta is out
Fails both acid2 and acid3. Hopefully they'll get it up to snuff before release.
Opera 9.5 gets to 65%
Closest of them all...
Not bad for the lightest and fastest browser of the bunch...
my copy of Opera (9.26) got through 32 tests before failing. So are there any browsers that actually pass all the tests?
Aw, poor Microsoft...
Aww, it must be so annoying to have to re-do some work just because things have changed.
It's just like web development; develop a site that works fine in most browsers and then spend days getting it to work on Microsoft browsers.
Better standards-compliance tests are good for the whole community in the longer term. Anyone tried Safari & Opera?
9.26 manages 46/100
9.50 beta (build 9815) manages 65/100
So what super browser passes all the tests?
Funny this should be mentioned - having a reference implementation is one of the reason Opera has put so much effort into things like CSS3 selectors, and implementing the significant new portions of HTML5 ( http://en.wikipedia.org/wiki/Comparison_of_layout_engines_(HTML_5) ). If it can establish itself as the benchmark, it reduces the risk of having to work around other's implementations.
My Firefox (from ONE HOUR AGO) gets only 68
So my 'hourly smoketest' build is a couple of points worse than the version Chris tried (above). It might be because I left my "RIP" extension in place, and something important is getting RIP'd, and a mere 2 points isn't much.
OTOH, if code like *this* is causing errors, maybe they need to type better:
<p> <strong/> Parsing Test </strong> </p>
</strong> is a block-level element, and isn't self-closing.
Why the differences?
Can anyone explain why the same versions of browsers used by the people above score differently? I would have thought that a test for standards compliance should at least give a standard set of results....?
Safari on an iPhone gets to 39%, that's not too bad really. (Stirs up nest of Hornets and runs for cover)
Surely the point of making the font tag hard to use, is that HTML is for document struture NOT style?
Surely all the font info should be in a CSS sheet some place.
50/100 firefox 2 here, and I guess you can get different results on the same browser if your have different settings.
Feels weird but must say...
To be fair this article should have noted that MS have (unusually) reversed their decision about the controversial aspect of their version targetting, presumably in response to all the flak they've got, and IE8 will now be standards-compliant by default. This is a truly Good Thing!
I think they deserve some credit for listening, not many companies do. It's easy to get sucked into the traditional MS-bashing, but not very objective.
Personally I always use Opera anyway. It's lightweight and rapid, a joy to use.
Acid 3 Tests
I'm thought that for a test to be added to acid 3, one of firefox, opera or safari had to fail that test for it to be added in the first place....
The Acid tests are meant to test behaviour on correct AND incorrect code, hence this Parsing Test to check if the browser correctly deals with code it can't parse properly. And strong is not a block level tag, it is an inline one but it requires some text content (so is meaningless when self-closed).
@Eric - IE8 passes Acid2
I'm running the beta on Windows XP SP3 and it renders the Acid 2 page perfectly.
Nice to see Opera choking on something too. Hakum's soapbox seems strangely empty today...
version 5.6 only gets a 39, but doesn't crash. 5.7 is in public alpha, not sure if that does any better.
There's no chance that this test was specifically targeted at highlighting IE shortcomings, is there? I mean, I'm sure the test makers are completely impartial and have no personal allegiance to any particular product.
Paris, because even she could see through this attempt to cattle-prod Microsoft's development in a direction preferred by the wonkers at the Web Standards Project.