Acid3 Test Unleashed, Murders Every Current Browser

[polonyc2]

Just a few months after the announcement that Internet Explorer 8 successfully passed the Acid2 standards compliance test, the Web Standards Project (WaSP) announced last Monday that it unleashed Acid2’s successor, Acid3.

Acid2, with its focus on Cascading Style Sheets, seems quaint in comparison to Acid3’s objectives, which target major web standards expected to see use today and in the future. Tests are derived from many of the last few years’ development in the web’s control languages, including rendering graphics embedded in HTML code, CSS3 compliance, DOM compliance, CSS2 downloadable fonts, as well as handling new graphics formats and Unicode support.

Currently, no known browser is able to correctly render the Acid3 test...

DailyTech - Acid3 Test Unleashed, Murders Every Current Browser

[CrowdAds]

[k0NG0]

Second time 'round that an Acid test has been released just at the end of a major Firefox version's development cycle. Acid2 was released just as Fx2 was about to be released.

[polonyc2]

these Acid tests seem like the 3DMark of the browser world...no current hardware can do well on it and is geared towards 1-2 years down the road...is it really even relevant, since they seem to create these tests just so current hardware/software can fail...just as IE8 and Firefox 3 are released to pass the Acid2 test, then they release an Acid3 test...seems silly

[k0NG0]

It shines a light on standards support by browsers. Microsoft could've just not cared and let every IE version fail Acid2, but instead they went and made it pass, which means its CSS support is pretty good (as I said, Firefox 2 doesn't pass).

And it's a bit like a benchmark in that it pushes people, developers in this case, to try and get a perfect score. But in a way it isn't, since the Acid tests highlight standardized web features. If you fail, you're not implementing the standard. Standards are good for interoperability, which is what drives the web. Benchmarks mean less.

So I think it's good. It's unfortunate that the release almost coincides with the Firefox 3 release, but such is life. Firefox 4 will be a hugely different beast, and the platform changes going into it will make it easier to comply to standards by lessening the cognitive load on programmers, so that's nice.

The latest Fx nightlies pass 66/100 tests in Acid3 I believe. The IE8 beta passed 17/100 tests.

[polonyc2]

if that were the case then wouldn't it make more sense to make passing Acid2/3 a requirement prior to releasing a new browser...that's like releasing a new television set that doesn't come with all the standard channels or a car that doesn't come with 4 wheels...seems like the standards are not really standards in that if it were all programmers would feel compelled to make sure their browsers met these standards

Acid3 might not be a benchmark but like 3DMark it seems like an artificial measuring stick of performance

[k0NG0]

Standards are fully opt-in, though. The W3C is a consortium that carries no actual political weight beyond what those who consume its standards think of it.

There is no higher regulatory entity that can force a company to implement a certain standard. The EU has put a lot of pressure on Microsoft to force them to allow interoperability with their products, but that's because of Microsoft's monopoly status, and where making interoperability harder benefits them because it makes them more money. The EU can't say "implement HTML5 or get fined", but they can say "make it possible to work easily with your hidden but important API such-and-such".

The web would be nothing without standards, but any requirements come from within a development team (what are the goals?) and browser users (can it do what I want?).

In the first case, politics matter, if the company developing the browser has a competing platform, there are sound financial reasons (being a for-profit entity) to push that solution instead of a standard. There are also good reasons to be standards compliant, since you'll catch PR flak if you act like a dick, and your product will work either way, so maybe compete on having a better development platform instead.

The Acid tests aren't published by a browser company, it's separate from that. So the reason tests and browser releases aren't better coordinated is because they're not coordinated at all, and they probably shouldn't be. That would be like an auto manufacturer getting to decide what tests its products have to pass to certify as safe.

I'm pretty sure most developers want to build products that are standards compliant. But they also have differences of opinion ("this standard sucks"), and they're usually held back by management.

It's not an artificial measuring stick as it directly tests what the browser can do vis-à-vis the standards it's meant to test. It's not synthetic in any way, it's a 1-to-1 mapping of "what should happen according to standard X" and "what actually happens".

[polonyc2]

so why woudn't any browser want to make itself 100% standards compliant?...is it that difficult to implement?...seems like every browser has major difficulty accomplishing this and when they do it pretty much goes un-noticed by the general public because as you said most end users don't care as long as their browser of choice can do what they need it to do (browser their favorite sites, banking, shopping etc)

with nobody to enforce these 'standards' Acid tests will always be an optional set of rules that nobody takes seriously

[k0NG0]

Oh, but they are taken seriously. Very much so. If you read Mozilla Planet you'll often see discussions about standards, how to get the best coverage, future plans etc. Since there now is healthy competition between browsers, there is incentive to be better than your competitors, that includes standards compliance.

For example, passing Acid2 required Gecko (the Firefox base) to undergo quite substantial changes to the rendering engine. Because the way it was originally designed didn't take into account future CSS rules and their requirements on rendering, etc. You can get edge cases where what on the surface would seem to be a minute change will require huge structural changes. That's the nature of most software. And since you're not just a browser but a platform, changing how APIs work (even internal ones) will cause problems to platform consumers, so you can't do it all willy-nilly.

It's worth noting that there's never been someone around to enforce these standards. Yet they are supported, to varying degree. The rise of Firefox has been the best thing for the web eco system since its inception.

[SupDawg]

Kongo, do you develop Firefox? You seem to have a serious interest in browsers in general.

[k0NG0]

Originally Posted by SupDawg

Kongo, do you develop Firefox? You seem to have a serious interest in browsers in general.

I do some work on a couple of Mozilla side-projects.

I am interested in browsers and the web. I think it's a very cool way of using the Internet, and I think openness and the web is the way to go as compared to closed, proprietary systems.