Comments on: the missed opportunity of acid 3

By: Jeff Walden

Jeff Walden — Tue, 01 Apr 2008 06:02:00 +0000

Samilar: people have indeed been consistently decrying acid3 on IRC for a long time now. Insinuate as much as you want, but he who has ears and a touch of willingness to actively seek out non-blogged opinions could have heard the complaints.

Personally, I think the criticism’s a touch overblown — there’s lots of good there (among which I would include font-face, even if we assume the “poorly thought of” part is accurate, because then it’ll force movement on the matter), but there’s also lots of cringe-worthiness. I don’t think SVG should have been in acid3, and many of the tests — including tests I personally fixed, so this isn’t fix-envy or something — were just inane and not relevant to the real web. It’s good we’re mostly past the point where we can point to entire unimplemented features to include in tests, but it reduces us to the edge-casey (sometimes to extremes) tests of acid3 if we’re not willing to work from newer standards (I don’t really understand the 2004 restriction).

I think that was probably me (Waldo) on IRC warning against E4X, not that it especially matters so long as nobody tries to run with E4X. :-)

By: Bruce Rindahl

Bruce Rindahl — Sat, 29 Mar 2008 22:33:27 +0000

I have to side with the Mozilla folks on this one. They are rightly focused on getting FF3 out the door, not on Acid3. When Acid3 came out officially less that a month ago, it became the primary focus of bug fixes for Safari and Opera. That is their choice and kudos to their work. However, I need FF3 as my sites will not work in FF2. The Mozilla team has been incredibly responsive to bugs and regressions I report to make FF3 better.

By: Lars Gunther

Lars Gunther — Sat, 29 Mar 2008 12:41:30 +0000

It was on IRC, two months ago, even before the competition to contribute to Acid3 had been announced. I do not remember the name of who it was I was talking to. It could have been you - it was a name I recognized. Whoever it was I was told exactly the same thing as you are saying now. And my point does not need to be addressed per se. It was just an example of a brave stance and proof of your commitment to standards!

I guess the real question is will there be an E4X 2.0 standard? Probably a question that will have to wait until ES 4 is reasonably stable.

By: dstorey

dstorey — Fri, 28 Mar 2008 20:50:29 +0000

And when it comes to commitment to standards Mozilla proved themselves ahead of Opera and Apple/Webkit on at least one account: CSS 3 selectors.

To be fair, Opera has full CSS3 selectors support in Opera 9.5, and added these long before Acid 3 was announced, so I'm not sure your point holds true. To be fully fair, Konqueror added these even before Opera.

By: Erik Dahlström

Erik Dahlström — Fri, 28 Mar 2008 20:50:03 +0000

"While this did protect us from tests requiring the worst of SVG’s excesses (1.2, with support for such key graphical capabilities as file upload and sockets, was promoted to CR in 2007)"

The draft that you link to has not made it to CR, it's an old working draft of SVG 1.2 Full. I fully agree that testing specs that are not even in CR would be inappropriate for the sake of an acid test. Perhaps you are referring to SVG Tiny 1.2, which is in CR, but which does not include file upload?

By: shaver

shaver — Fri, 28 Mar 2008 17:49:13 +0000

If you’re referring to me as one of the people who recommended against E4X (you don’t say who, so it’s hard to actually address your points), it’s not because of quality of implementation or competitive issues — we’re way ahead of any other browser on that score, as far as I can tell. It’s that the spec itself is badly flawed, and forcing more people to follow its missteps would be a disservice. Similarly, there are behaviours required by the letter of ES3 that are considered to be bad for security if implemented literally, and so we didn’t propose them for Acid 3, though we’ve had the by-the-letter behaviour for years — pushing other browsers to implement them would be a disservice to the web.

By: Lars Gunther

Lars Gunther — Fri, 28 Mar 2008 10:06:40 +0000

@samilar: Mozilla people did say similar things about the test as is stated in this blog post from the announcement of this test – on IRC when asked by me and in Bugzilla. Check your facts!

It is plainly obvious that Apple/Webkit wanted to pass Acid3 (their font “fix” is the final proof). They are still very much an underdog compared to Mozilla – and perhaps even Opera who rule the mobile market. (Yes, I know about the iPhone but its user base is still relatively small, especially outside of the US where more than 95 % of the world population lives.)

In all fairness the Webkit font “fix” bug has been reopened in anticipation of a better solution.

I planned to contribute an e4x test to Acid3, but the spec is too new, and when I inquired about the idea Mozilla people discouraged me! Since they did not think their current implementation is good enough. That speaks volumes about a commitment to good standards in preference to winning a race.

And when it comes to commitment to standards Mozilla proved themselves ahead of Opera and Apple/Webkit on at least one account: CSS 3 selectors. Instead of just implementing them in a way that pleases Acid3, Mozilla postponed implementing a few selectors in order to first discuss the spec per se, since it was ambiguous, knowing it would delay improvement of their score on Acid3. Improving the standard was seen as more important than winning the race. I would almost go so far as call that act sacrificial.

Now I do think it would be wise for Mozilla to do a Firefox 3.1, perhaps as early as 2 or 3 months after 3.0 and not on a different code branch! The reason? To add this stuff, provided the specs have matured: - - XHR X-site requests - CSS 3 selectors - media queries - “real” border radius And perhaps with a 5-10 point better Acid3 score… And perhaps with this bug fixed: https://bugzilla.mozilla.org/show_bug.cgi?id=424973 ;-)

By: Sean

Sean — Fri, 28 Mar 2008 00:14:47 +0000

It’s funny to see you making the same argument that IE made during IE7 about Acid2: http://blogs.msdn.com/ie/archive/2005/07/29/445242.aspx

By: cris

cris — Thu, 27 Mar 2008 17:52:50 +0000

agree

main point being. dont let acid 3 test just become a test, it should be helping web developers and users. the method webkit used to pass the test is just pointless. if u pass that testm but doesn't support the full standard of that test. why bother? to confuse web developers?

By: shaver

shaver — Thu, 27 Mar 2008 17:32:27 +0000

Firefox 3 won't be far behind the other shipping competition when we ship, maybe a few points, and we're something like 30 points higher than Fx2 was; I don't think there's anything shameful there. We're way ahead of the real competition, which is Internet Explorer, for those people who think that Acid 3 signifies something that it didn't.

Why now? Because I finally had time to finish the post, albeit at 2AM. I've been hammered trying to get fixes into beta5 and therefore Firefox 3 (and trying to learn more about the history of things like @font-face, or how Acid 3 seems to depend on HTML5-specified parsing rules, but not HTML5-specified .innerHTML rules). I'l confess that when I saw people taking hacks specifically for the test I was newly energized to finish it off, if only to convey my hopes that Mozilla wouldn't follow the same path. We need to not rush to implement the parts of a spec that are in Acid 3 and leave authors to guess whether having createNodeIterator means that they can use the API as specified, or just as covered by Acid 3. (We left SVG filters turned off in Fx 2 for that reason: the performance wasn't great, and we didn't want people to feel that they couldn't use filters until Fx2 had fallen off the web.)