Battle for full Acid3 compliance won by Opera?

Regular listeners to the NeowinCAST podcasts will know I get more than a little passionate about web development, and a surprise announcement by Opera today at having scored 100/100 in the Acid3 test has caused more than a few heads to turn.

Acid3 is a test page from the Web Standards Project designed to check the compliance of a web browser to various DOM and javascript standards. Different browser manufacturers place varying amounts of importance on passing the tests in their nightly builds, but today Opera announced they have passed the full compliment of tests for the first time.

Previously the race had been expected to be won by WebKit, the rendering engine behind Safari with a current score of 98%; on the Surfin' Safari blog, daily updates of the edge towards passing each test had been given much to the anticipation of web developers.

It should be noted, however, that Opera's announcement has not been verified by the wider community as their preview build is not available for testing. Conversely, nightly builds of WebKit are freely downloadable for developers to test at their leisure.

Nightly builds of Firefox currently score 71%, while Internet Explorer 8 beta trails behind with 18% compliance.

View: Surfin' Safari blog for WebKit updates
View: Announcement from Opera
View: Regularly updated listings of Acid3 browser compliance
View: Neowin Forum discussion

Report a problem with article
Previous Story

Flaws found in Safari browser for Windows

Next Story

Camino 1.6 Beta 3 Released

64 Comments

Commenting is disabled on this article.

Since when are these 'Acid' tests the "only" benchmark for a web browser? I tried Opera and didn't like it so whether it passes an Acid test means nothing to me as it means nothing to most people using a web browser.

Here's the thing: If it displays the page properly (and I have yet to see any page in IE not display properly, all but one in FireFox displays correctly) then who cares???

I have seen plenty of pages not display properly in IE as I have designed websites before... these pages then have to be botched to render properly, without breaking functionality in browsers with better standards support (like Opera and Firefox). IE is still terrible when it comes to rendering pages. I whole heartedly congratulate the Opera developers for trying to improve the web experience and making designing web pages more straightforward. I like Opera but there are still some niggles that prevent me from using it as my default browser (not being able to drag pictures from the browser into a folder, for instance).

Whoo hoo. Colour me 'who gives a ****!'

The Acid tests are an artificial test. They mean nothing. What counts is can your browser display the pages that you want it to display? Nothing else matters.

(FF and IE user)

Agreed, a browser can pass acid3 at 100% but still unable to a proper render of real pages, i.e. Safari, Opera and in some cases firefox.

As Opera said, it did 100/100, but there are still some minor issues remaining for them to resolve.

A 100% pass doesn't have to mean everything is there, I believe there's a color animation that has to be "smooth" for example, which doesn't show up in a screenshot, for one thing. ;)

But both the Opera and Safari teams have made excellent jobs here!

If I were to pick today, it'd have to be 1) Firefox 3, 2) Opera 9.5, 3) Safari 3.1. Firefox at my first spot because of great extensions and the excellent memory usage and speed nowadays. Its customization AND performance is hard to beat. Additionally, Acid3 is not yet a major downside as websites won't be designed for this until IE supports the standards well, which won't even come in IE 8. That browser will do Acid2, which Firefox 3 also does.

Opera coming second for everything you get out of the box in a leaner install than Firefox's without extensions, and also being very fast and with great memory usage. Safari coming last because of its speed that also seem to rival the best, but consuming way too much memory on Windows at least to be comfortable after long sessions, and lacking even Opera's integrated adblocking support, as well as failing Windows UI guidelines miserably. :-( Yes, I know it's branding, but that doesn't make it less of a problem in some areas.

I'd be a hell of a lot more interested in Webkit, if it came in the form of a Windows browser.

For what good it may do, Safari is hardly a Windows browser. It's a MacOS app in look and feel clumsily dropped in the middle of Windows. It reminds me of the old X11 days, before GTK+ and similar popular toolkits, where every application looked completely different. Although, at least, the font rendering was the same across apps (aside from a few doozies like TeXmacs, which did it differently for a reason)

Exactly. Safari is simply a terrible browser - at least on Windows. It would be nice to see a proper Windows frontend released for Webkit, as then it could rival Firefox and Opera.

The problem, is that the windows port of Safari links to certain closed source libraries, so nobody else can make a windows app that uses the windows port of WebKit.

Luckily, there is a sub-port in the works that replaces the closed source stuff with open source stuff (Cairo instead of CoreGraphics and cURL instead of the networking stuff)

Cairo is LGPL (like WebKit) and cURL is MIT, So there shouldn't be any problems with using them in closed source apps.

WebKit nightly is at 99% now. Just missing the blue box.

Edit: Scratch that, after emptying the cache and reloading, it gets 100%.

I just downloaded the latest WebKit nightly, and it gets 100%.
It's the first publicly-available browser to get 100% in the test. I don't think Opera won this round.

as long as the browsers of choice out there display pages right, this acid tests are really meaningless. much like some of the certification that antivirus programs get because they use default settings and some developers choose to tweak their defaults for increased security instead of focusing on certificates.

but, nonetheless, this acid tests will help developers get the standards right more quickly and we all can benefit.

The fact that all these browsers are close to 100% in only a week or two shows you that must not be that many differences between acid2 vs 3 in the way the browser engine renders.

It took forever to get most browsers to render acid 2 let alone 3... so acid 3 honestly doesn't impress me as much.

Acid3 has focused on a wider range of things. It requires many new features such as SVG and web fonts that will make a massive difference to the internet in the future. If all the browsers support CSS, SVG and web fonts properly, web developers' lives will be made considerably easier and you'll have better looking web pages.

Acid3 is brilliant. We've seen browsers scrambling to get attention by trying to support it, which results in a better web browsing experience for users. Just because we have two browsers working hard on supporting it does not make the test any less valid - Firefox and IE are still lagging far behind.

(theyarecomingforyou said @ #13.2)
Acid3 is brilliant. We've seen browsers scrambling to get attention by trying to support it, which results in a better web browsing experience for users. Just because we have two browsers working hard on supporting it does not make the test any less valid - Firefox and IE are still lagging far behind.

I agree ACID is pretty good. But it seems to be a marketing scheme by browsers now a days, which still are riddled with other issues that affect us developers just as much

Not sure why that makes it a sham. If Mozilla or Opera found a bug, it'd be fixed too... Tests can have bugs as well as browsers can.

So yes, WebKit won. Opera almost won, but the tests were wrong.

I don't think these ACID tests actually mean much in complete compatibility. People think if it passes ACID 2, 3 that a correctly coded page will render 100%. That's simply not the case, and I can tell you this from current experiences. I have a page that is completely 100% XHTML 1.1 valid and valid CSS and Opera screws up while Firefox and IE work perfectly fine. Doesn't that seem oddly ironic considering Opera is supposed to be the best in this field?

Granted I'm not running the latest nightly build of Opera, that doesn't change the fact that I still won't have to write workarounds for compatibility.

The problem turned out that Opera messes up it's internal width values for liquid columns (in a table) when a class is added via Javascript and it takes up extra space that wasn't assigned to the column that the class was added onto.

(Tikitiki said @ #11)
I have a page that is completely 100% XHTML 1.1 valid and valid CSS and Opera screws up while Firefox and IE work perfectly fine. Doesn't that seem oddly ironic considering Opera is supposed to be the best in this field?

IE 7 doesn't support XHTML 1.1 well since it doesn't support the proper MIME type without using hacks. On IE, web servers normally have to be configured to send the XHTML page as HTML for IE to not try to render the whole page as a raw text document, and then you could just as well be using HTML 4.01 because you'd be violating the spec in this case. That's why using XHTML is considered harmful and to follow the specs without restorting to hacks, people would probably normally have to stick with HTML 4.01.

Here's more info on this subject:
http://www.w3.org/MarkUp/2004/xhtml-faq

Why is it disallowed to send XHTML 1.1 documents as text/html?

XHTML 1.1 is pure XML, and only intended to be XML. It cannot reliably be sent to legacy browsers. Therefore XHTML 1.1 documents must be sent with an XML-related media type, such as application/xhtml+xml.


Does Microsoft Internet Explorer accept the media type application/xhtml+xml?

No. However, there is a trick that allows you to serve XHTML1.0 documents to Internet Explorer as application/xml.

(Jugalator said @ #11.1)

IE 7 doesn't support XHTML 1.1 well since it doesn't support the proper MIME type without using hacks. On IE, web servers normally have to be configured to send the XHTML page as HTML for IE to not try to render the whole page as a raw text document, and then you could just as well be using HTML 4.01 because you'd be violating the spec in this case. That's why using XHTML is considered harmful and to follow the specs without restorting to hacks, people would probably normally have to stick with HTML 4.01.

Here's more info on this subject:
http://www.w3.org/MarkUp/2004/xhtml-faq

My mistake. It is XHTML 1.0 Transitional which I do believe works fine with IE 7.

Either way, my point still stands.

To bad its UI is a total crap shoot. and It cant get market share to save its life, (I think it was BARELY beating out Netscape last I saw)

who cares.... Firefox 3 is clearly much better at memory management than Firefox 2 was... and those test are more or less for 'future' stuff right?

nothing to really worry about.

uhh lol. What does Firefox have to do with this story?

/waits for a web developer to destroy ThaCrip's comment.

Congrats to the Opera team. I'm liking the 9.5 beta.

(ThaCrip said @ #8)
who cares.... Firefox 3 is clearly much better at memory management than Firefox 2 was... and those test are more or less for 'future' stuff right?

nothing to really worry about.

So you are saying that some memory management with some systems its more important than stablishing a standard for the web?

Congrats to the Opera team! It's nice to see all browsers working toward improving the web experience at such a brisk pace, it's nice to see the competition. With Opera, Safari, Firefox, and IE, there are lots of choices on all platforms, and even mobile offerings.

How do you know those "chunk of pages" aren't at fault themselves?
Perhaps the pages are faulty and Firefox is merely more tolerant of bad code than Opera?

Got any examples?

(TCLN Ryster said @ #6.1)
How do you know those "chunk of pages" aren't at fault themselves?
Perhaps the pages are faulty and Firefox is merely more tolerant of bad code than Opera?

Got any examples?

considering the fully released version of firefox (v2) doesn't even pass the acid2 test, i'd go with the opera renderings over the firefox ones

Well, the last time I used Opera it was not able to correctly render a <select>-tag, and I had to program lots of extra lines in order to get it to work correctly. That is, on a XHTML 1.1 strict validated page.

So, what sense does it make if it passes that freaking test, but fails at basics ?

Nothing wrong with a bit of friendly competition, especially if it benefits everyone =D
Now come on Mozilla, get your arses in gear!

Problem is, just like back in the day when Acid2 was released, Acid3 was released at the end of a major Gecko release cycle when it's too late to make big changes without significantly delaying the next release. Don't expect big advances on Mozilla Acid3 compliance until the Gecko 2.0 (Fx4 or whatever the next major release is called) timeframe.

Oh, and FWIW, a fair number of the outstanding Gecko Acid3 bugs actually have patches posted to them ready for review and landing after Firefox 3 is released, so expect the score for nightlies to start jumping again after that release.

(sanctified said @ #3)
Again, until its not released its meaningless.

It's not released right now so does it have meaning?

It is completely without meaning, even though they have had their coders working on this until they hit the 100% mark?

"Not trying" is meaningless. Putting out the effort has great meaning. Perhaps not practical until released to the public, but please don't dismiss their efforts in their entirety.


*disclaimer: I am not an Opera user. Nor am I entirely convinced that ACID testing is the most accurate way to tell the best browser. It is just a single benchmarking tool.

(Eis said @ #3.3)
Assuming you screwed up by saying "not" released sanctified, that's bull****. Why wouldn't it matter?

1.- Because the number indicator can display 100 despite other rendering errors.
2.- That could as well be the reference page.

As a whole we cant determine the results properly without a public build, even many members at the opera forums said that the real challenge is: "Who will release a public version: Opera or Webkit?"

(markjensen said @ #3.2)
It is completely without meaning, even though they have had their coders working on this until they hit the 100% mark?

"Not trying" is meaningless. Putting out the effort has great meaning. Perhaps not practical until released to the public, but please don't dismiss their efforts in their entirety.


*disclaimer: I am not an Opera user. Nor am I entirely convinced that ACID testing is the most accurate way to tell the best browser. It is just a single benchmarking tool.

Im not dismissing their efforts mark, I just say that we need something more concrete than a screenshot.

(sanctified said @ #3.7)

Im not dismissing their efforts mark, I just say that we need something more concrete than a screenshot.

Seriously dude... those are Opera's developers posting that not some leaked opera build on some chinese website

(Beastage said @ #3.8)

Seriously dude... those are Opera's developers posting that not some leaked opera build on some chinese website

So, you believe everything you read I suppose.

Well, that works for you I guess.

Believe me, when I see a public build I will be the first to congratulate them.

(sanctified said @ #3.9)


So, you believe everything you read I suppose.

Well, that works for you I guess.

Believe me, when I see a public build I will be the first to congratulate them.

Now you're just being argumentative. Saying that because he believes what the Opera developers say about their own program and saying he would believe everything he reads is just a ludicrous comparison.

(excalpius said @ #2.1)
Can anyone explain to me why anyone, I mean anyone at all, would care about this? I'm asking seriously.

The only people who care are the people who own the browsers. Only two letters matter to them: P-R (Press Release).