Join 3,433 readers in helping fund MetaFilter (Hide)


WebKit build scores 100/100 on Acid3
March 26, 2008 8:46 PM   Subscribe

WebKit, the rendering engine of Apple's increasingly popular Safari web browser becomes the first "publicly available rendering engine to achieve 100/100" on the Acid3 web standards test. The Opera browser is expected to have an experimental build that passes soon, as well.
posted by aletheia (72 comments total) 4 users marked this as a favorite

 
Note that both browsers are working on additional improvements that constitute a "full pass" beyond the 100/100 score itself.
posted by aletheia at 8:49 PM on March 26, 2008


Didnt Opera post its screenshots of passing Acid3 HOURS before Apples announcement?

Seems like some people are trying to steal Opera's thunder.
posted by IronWolve at 8:54 PM on March 26, 2008


Didnt Opera post its screenshots of passing Acid3 HOURS before Apples announcement?

Seems like some people are trying to steal Opera's thunder.


It's called competition. Opera touted their engine, Apple's developers came back with a public piece of software instead of a screenshot.

Why snark over this? Who gives a shit about "thunder"? Competition is good for all of us.

And how do you steal the thunder of unavailable software?
posted by Mikey-San at 8:58 PM on March 26, 2008


Note that both browsers are working on additional improvements that constitute a "full pass" beyond the 100/100 score itself.

A very important point to make. Passing Acid 3 is more than achieving a 100/100 score. Via the Acid 3 page:

To pass the test, a browser must use its default settings, the animation has to be smooth, the score has to end on 100/100, and the final page has to look exactly, pixel for pixel, like this reference rendering.
posted by Mikey-San at 9:00 PM on March 26, 2008


Not that it really matters, but apparently the test creator, Ian Hickson, fixed a bug in the test that would have knocked Opera's internal build back down to 99 before the WebKit team scored 100. Obviously there will be two camps bashing each other on this, it's unavoidable. But remember that 1) the test isn't fully passed until they both work out additional rendering kinks and 2) only WebKit is available publicly. The latest version available for Opera does not score 100/100. The important point is that having browsers (all of them) work so aggressively to pass tests like this (even the edge cases) is good for everyone.
posted by aletheia at 9:00 PM on March 26, 2008


To be fair, Opera have released a nice screen shot but the Webkit folks apparently have the source publically available and today's nightly build will allow you to try it for yourself.

Personally I don't care about who is first, it is great having two widely available (if not widely used) browsers that track the standard so closely. Have first-class support for SVG available will open up lots of cool stuff. I also like what webkit is doing with the CSS transform property, even if it isn't actually part of the standard yet. CSS Animations + CSS Transforms == ROTATING WIN.
posted by AndrewStephens at 9:02 PM on March 26, 2008


Too soon!
posted by mazola at 9:03 PM on March 26, 2008


But will it go to 111.
posted by stbalbach at 9:04 PM on March 26, 2008 [5 favorites]


The important point is that having browsers (all of them) work so aggressively to pass tests like this (even the edge cases) is good for everyone.

Personally I don't care about who is first, it is great having two widely available (if not widely used) browsers that track the standard so closely.

Absolutely. If we're going to have a second "browser war", this is the kind of war I want.
posted by Mikey-San at 9:09 PM on March 26, 2008


(shrug) Cool. As long as Safari beats Firefox, I'm pleased.

I have no real beef with Firefox, but it's the rabid fanboyism of some of its supporters, who seem to think it's The One True Browser, that tends to turn me off. That and the fact that its performance lags behind Safari on OS X. The Firefox diehards seem to either overlook that, or to be entirely unaware of it, focusing on its superiority over IE on Windows (not that that's such an impressive feat).

And I'm happy to have it around, in any event; insert "competition is good" platitudes here.
posted by CommonSense at 9:11 PM on March 26, 2008 [1 favorite]


What I don't like about the Acid tests is that they've degenerated from interesting benchmarks to Hakon Wium Lie's personal vendetta against MSFT.

Hakon Wium Lie: IE will NEVER pass Acid2!

Chris Wilson: We got an RC build of IE8 to pass Acid2.

Hakon Wium Lie: Oh yeah? BEHOLD! You will never pass Acid3!

Chris Wilson: We spent the weekend making massive changes to IE8. IE8 beta 2 passes Acid3.

Hakon Wium Lie: No! You cannot defeat me! Here is Acid4, which is nothing but SVG and canvas tags embedded in HTML5!

Chris Wilson: IE9 passes Acid4.

Hakon Wium Lie: NOW YOU WILL PAY! Acid5 uses EXSLT and Flex to create a living unicorn that craps rainbow colored marshmallows that each read "USE OPERA GOD'S FINAL BROWSER!"

Chris Wilson: OK, you got us there. IE10 doesn't support rainbow colored marshmallows.

Hakon Wium Lie: YES!!!! VICTORY IS FINALLY MINE!!!!! OPERA AND LINUX HAVE DESTROYED MICROSOFT FOREVER!!!!!

Web developers everywhere: Yeah, we stopped writing to Opera after all those marshmallows fried our motherboards one too many times.

Hakon Wium Lie: FOOLS!!!
posted by dw at 9:12 PM on March 26, 2008 [20 favorites]


Yeah, for the first time ever we have FOUR totally different browsers that are all viable options for browsing the web. Even IE seems to be lifting its game slightly with IE8 coming out soon.
posted by AndrewStephens at 9:13 PM on March 26, 2008


Update: WebKit has just been updated again and now passes the Acid3 test "pixel-exact."
posted by aletheia at 9:15 PM on March 26, 2008


Joel Spolsky was talking about the problems faced in IE 8 by implementing the standard very closely:

In practice, with the web, there’s a bit of a problem: no way to test a web page against the standard, because there’s no reference implementation that guarantees that if it works, all the browsers work. This just doesn’t exist.

So a lot of people coded what they thought were standards, and ended up creating things that don't work in 'true' implementations. But it's pretty cool that there is now something that will actually render this perfectly.

That said, I think Joel is kind of a self-important windbag with more ego then sense at this point. His article goes on and on about some tortured analogy about headphone plugs on mars. It's kind of a slog.

HTML and CSS are really kind of a kludge. It started with simple semantic markup, and now the goal is pixel precision, and animation, interactivity. It's trying to be everything to everyone.

I was thinking rather then using a soup of tags to try to specify what rendering engine to use just let web page authors embed their own rendering engine in an intermediate language (so there wouldn't be a security issue).

And, now that we have the <canvas> tag, you can actually write your own rendering engine in Javascript and include it with your page, if you really want to do that. If you stripped out the features you don't need, you could probably come up with one that was pretty small.
posted by delmoi at 9:19 PM on March 26, 2008 [1 favorite]


AndrewStephens' point is a good one. With Apple and Opera in this position, Mozilla presumably won't be too far behind with FireFox 3's score, but with IE still the #1 browser out there, one hopes they will join the bandwagon (despite dw's rant).
posted by aletheia at 9:20 PM on March 26, 2008


I have no real beef with Firefox, but [...] the rabid fanboyism of some of its supporters

A somewhat ironic comment given that you're touting an Apple product, no?
posted by Slothrup at 9:25 PM on March 26, 2008 [6 favorites]


Pot meets kettle, CommonSense.
posted by puke & cry at 9:38 PM on March 26, 2008


delmoi, I actually usually enjoy Joel Spolsky's writings (more form his role as entrepreneur than technologist), but WTF was that article? Sent-up in classic Daring Fireball style by Mark Pilgrim
posted by misterbrandt at 9:40 PM on March 26, 2008


i don't care about acid3 but i switched to safari a few weeks ago
it's so much faster and doesn't crash or eat up CPU the way firefox does. (at least on mac)

i've been using mozilla/firefox since the beginning. it seems to have gotten worse.
posted by bhnyc at 9:47 PM on March 26, 2008


I was thinking rather then using a soup of tags to try to specify what rendering engine to use just let web page authors embed their own rendering engine in an intermediate language (so there wouldn't be a security issue).
Maybe I'm missing something, but allowing web pages to embed arbitrary rendering engines sounds like exactly the sort of thing that would cause security issues. Attempts to create an 'intermediate language' that doesn't put computers at risk have resulted in JS + DOM, Flash, SVG, and so on. To a large degree, the ACID tests are about handling that 'intermediate language' -- all the browsers have to support the intermediate, after all, or we're back to the land of embedding java applets and building everything in flash.
posted by verb at 9:48 PM on March 26, 2008


But will it go to 111.

It's like, how much more compliant could this be? And the answer is none.

None more compliant.
posted by cortex at 9:55 PM on March 26, 2008 [12 favorites]


I have no real beef with Firefox, but [...] the rabid fanboyism of some of its supporters

A somewhat ironic comment given that you're touting an Apple product, no?</em

and Safari has to run stably on the windows platform before it can actually compete as a production browser.

Don't get me wrong, I like Safari, it's fast, firefox is a memory hog- but is quite reliable compared to Safari.

posted by mattoxic at 10:01 PM on March 26, 2008


Oh, and </em>
posted by mattoxic at 10:02 PM on March 26, 2008


IE still the #1 browser out there, one hopes they will join the bandwagon (despite dw's rant).

Funny story... I was talking to my old boss from the dotcom days. We were HTML code monkeys back during the boom. After the bust, I went to academia, he floated from contract to contract at MSFT.

Well, a few months ago he gets a call out of the blue from elsewhere in MSFT that they're desperately looking for old-fashioned HTML coders and could he come work for this group? Next thing he knows, he's working on bringing all these MSFT websites up to standards. The money is good, he's happy, but there's something bugging him.

"I need to test for Firefox so I can feed it its own stylesheet," he said.

I looked at him a bit puzzled. "Why would you want to do that?"

"Well, I can't get the pages to look the same in IE, Firefox, and Safari."

"Let me guess. You're designing in IE, aren't you? You need to start in Firefox or Safari and then use conditional comments to just feed the hacks to the specific IE browsers."

"So why can't I just do it the other way? Aren't there filters and hacks for Firefox and Safari?"

"Yes, but they're so rarely used that I don't think you could get 1 in 100 code monkeys to give you at least one. On the other hand, at least half could tell you about the Holly Hack, or zoom: 1, or how to fix the double margin bug in IE6."

He shook his head.

"That's sad, you know? My own employer can't build a goddamn browser that works with everything else."

"Welcome to the last ten years of web design, boss."
posted by dw at 10:13 PM on March 26, 2008 [4 favorites]


What I don't like about the Acid tests is that they've degenerated from interesting benchmarks to Hakon Wium Lie's personal vendetta against MSFT.

Maybe if MSFT would condescend to mingle with the proles on the various HTML5 lists (as Hakon Wium Lie does) it could have some input into the Acid tests. It refuses to so condescend. Fuck it.

Kudos to Apple and Opera. Firefox, please stop playing catchup on this shit. You're better than that.
posted by enn at 10:17 PM on March 26, 2008


And on Firefox, I switched to the current beta of Firefox 3 a month ago. They've plugged most of the memory leaks and it's a lot more stable now. Much better than FF2.

Just wish Firebug could get updated so I don't have to keep switching back to FF2 when I want to run it.
posted by dw at 10:18 PM on March 26, 2008


Firefox, please stop playing catchup on this shit. You're better than that.

I held high hopes for Firefox, but with constant stability problems on Linux and OS X, I've had enough. Back to Safari and Opera for me.
posted by Blazecock Pileon at 10:21 PM on March 26, 2008


Maybe if MSFT would condescend to mingle with the proles on the various HTML5 lists (as Hakon Wium Lie does) it could have some input into the Acid tests.

Considering how condescending Hakon has acted on those lists and elsewhere, that's a bit pot.kettle.black.

I personally think HTML5 is going to be a massive disappointment. They're touting something that looks like a blog templating language as "the future?" Is "the future" is nothing but newspapers and blogs? Good grief, they're missing the boat even more than XTHML 2.0 is.
posted by dw at 10:27 PM on March 26, 2008


I would love to switch back to Safari but I depend too much on Firefox extensions nowdays.
posted by mrbill at 10:31 PM on March 26, 2008


Sour grapes.
posted by YoBananaBoy at 10:38 PM on March 26, 2008


The em tag is for emphasis, not quoting, mattoxic. Remember what kind of thread you're in! ;)
posted by Mikey-San at 10:44 PM on March 26, 2008


It is beyond awesome that Webkit (the Safari/Omniweb/KHTML/etc engine) is now competitive against Presto (the Opera engine). For the first time ever, Opera's going to have a standards-compliant contender. Very cool.

Congratulations to both teams.

I hope the Safari team comes up with some innovative UI. Opera has completely ruled the roost on that front: its keyboard shortcuts are unparalleled, it has invented several UI elements that are now found on other browsers, and its got amazing configuration capabilities (though to really do the deed, one has to dig into its (well-documented) configuration files).

I think good times are ahead for those of us who use the web intensively. Whoo-hoo! Go Opera, Go Webkit!
posted by five fresh fish at 11:00 PM on March 26, 2008


I'm using a build of WebKit downloaded today, and it's only scoring 96/100 on Acid3. Curious. Is there another build I should be using? (I'm on a Mac / Intel.)

Incidentally, Firefox 3b4 went tits-up one time too many today. This might not have bothered me so much - it IS a beta, after all - if I hadn't just read an article in which Beta 4 was declared by its Mozilla overlords as ready for prime time. Their reality and mine don't seem to be intersecting, so I made the switch to Safari. I've been tooling around with WebKit for weeks, and have found it fast and stable. Still waiting to discover a downside...
posted by bicyclefish at 11:04 PM on March 26, 2008


YoBananaBoy: Definitely a sour tone, yeah, but that view has been around. It's debatable, though I would prefer to see Mozilla not voicing it (why doesn't he allow comments?...) Firstly, I think there are certainly enough tests (and don't forget the numerous subtests within) that impact real world situations, from rendering standards to performance, to justify having the test. "Worthless" is a suspicious claim. But it isn't just about improving compatibility with current sites. From what I saw of the WebKit's team on this task, a great deal was learned about the current specs on various web technologies, flaws were exposed, and interest grew in improving them. Being challenged to support edge cases can also lead to strengthening the quality and flexibility of your codebase. Additionally, the progress on this test also means that developers can adopt some of the newer features sooner, because there is a common goal all the browsers can focus on to inspire adding support, and when they do, they can expect more reliable results between each engine. I think his blog post there is needlessly dramatic.
posted by aletheia at 11:05 PM on March 26, 2008


> allowing web pages to embed arbitrary rendering engines sounds like exactly the sort of thing that would cause security issues.

Not necessarily; although it all depends on how exactly you implement the virtual machine that runs the code. But most of the embedded-languages that we think of as security risks today are security risks exactly because their VMs were broken by design. E.g., the VM wasn't completely isolated from the host system, and was designed to allow code running in the browser's little sandbox to interact with your computer/OS/filesystem.

IE, a few versions back, effectively had little to no sandboxing with regard to ActiveX controls; one wrong click with ActiveX controls enabled and suddenly you're running somebody else's code right on your machine's bare metal, with access rights basically equivalent to any desktop application. Oops.

The problem there wasn't that running code embedded in a webpage is inherently a Bad Thing, although many people took it that way, but just that you shouldn't let it run all over your machine. People just don't expect a webpage to be able to access their hard drive, or start up persistent services, or basically do anything outside of the page they're looking at. Personally I think this is a pretty reasonable assumption -- if you were flipping channels on the TV, and suddenly your TV grew legs, walked around, and kicked your dog, you'd be pretty pissed. Why? Because while you're no doubt aware of there possibly being offensive stuff on TV, it's not supposed to actually physically come into the room and wreak havoc. That's exactly what Microsoft designed ActiveX to be able to do, however. It's like asking for holographic TV and getting something that can actually kick you in the balls.

Java and JS code, by contrast, run either in a sandbox in the case of JS, or in a VM in the case of Java. The Java VM has proven pretty robust over time (there have been a few issues but nothing really major that I'm aware of), and as computers get faster the performance penalty gets less and less significant.

So if you wanted freeform web rendering, you could probably do something like let code run but only in isolated VMs, where they'd be limited to a few types of input and output. E.g., only input is the user clicking directly on the applet's area, only output is turning pixels in that area on and off and changing the color, and only communication is back to the site they were downloaded from. Aside from someone jailbreaking the VM (always a possibility, although the less complex it is, the better) I think the damage from running that code would be pretty limited.

I think the lesson if you're designing a browser is that being proactive (like Java) and locking stuff down, is better than being initially permissive and then waiting for a problem before pushing out lame reactive fixes (like IE's ActiveX controls).

If you're interested in reading about the differences between Java and ActiveX's security models, there's a good article here. He takes it easier on ActiveX controls; I think the programmers involved should all be severely beaten.
posted by Kadin2048 at 11:28 PM on March 26, 2008 [2 favorites]


There is one crucial reason why I am fond of firefox and that is because it reignited the browser war- and made IE step up to the mark.

The simple fact is that most of the world's computing happens on PC, and until FF the choice was stark. Coders didn't have to worry about cross browser apps too much, the minimum (and terrible) benchmark was IE 5 MAC.

Due to firefox's penetration we now have excellent Safari, Opera which is also lovely and even IE sports tabbed browsing- but is still a pile of standards ignoring shit- but then again, when did MS ever give a crap about standards that weren't their own?
posted by mattoxic at 11:32 PM on March 26, 2008


Molly has expressed view a number of times about the Acid2/3 tests not being the be all and end all of web browser standards.

But her main concern is around the governance of the tests, in this case Ian Hixie is one one who's creating them, and is therefore holds a huge amount of power about how x or y is rendered in a browser.
posted by X-00 at 11:56 PM on March 26, 2008


That and the fact that its performance lags behind Safari on OS X...

My anecdotal two cents; my wife and I have a Mac running OS X, and Safari kept crashing all the time. We recently switched over to Firefox, and it seems to work a lot better.
posted by The Card Cheat at 12:51 AM on March 27, 2008


Nice little pickup in the link YoBananaBoy posted, here is a webkit patch that changes the way webkit renders the font used in Acid3 (but only that font) so that it looks right.
posted by markr at 1:50 AM on March 27, 2008


And yet the fucking thing STILL can't render table captions properly. I so rarely use Safari, it's as bad as IE, just in different directions.
posted by bouncebounce at 2:26 AM on March 27, 2008


i posted a video of webkit rendering the test on youtube for those who actually want to see the (relatively boring) render. The animation is part of the test, as it should animate smoothly according to that page.
posted by sxtxixtxcxh at 2:38 AM on March 27, 2008


I would love to switch back to Safari but I depend too much on Firefox extensions nowdays.
Me to, but the stunning ugliness and bad UI design decisions in FF3.0 have me re-thinking my options. I can put up with a few ads and the ugly metal theme if it means not having to put up with the new Firefox's stupid "search everywhere but the actual URL" URLbar search, the stunningly ugly & fecked-up new default OS X theme, and everything else they've changed for the sake of "bright! shiny! change!".

Give Safari a decent plug-in architecture & Adblock equivalent, and FF will be relegated back to a minority browser on Linux. Which is a pity, 'cos I've been using and praising FF since V0.6-something.
posted by Pinback at 3:17 AM on March 27, 2008 [1 favorite]


I would love to switch back to Safari but I depend too much on Firefox extensions nowdays.

Hear hear. I've started answering the "what browser do you use?" question with "Adblock" and any puzzling look after that with "...with Firefox as the rendering engine for that".
posted by DreamerFi at 3:27 AM on March 27, 2008


For the first time ever, Opera's going to have a standards-compliant contender.

So WebKit and Gecko aren't "standards-compliant contenders" until they pass this single, arbitrary test? Really? Opera may have led the pack back in the bad old days, but as far as I'm concerned, all three have been legitimate options for the standards-aware for quite some time now (and IE8 makes four).

I hope the Safari team comes up with some innovative UI. Opera has completely ruled the roost on that front: its keyboard shortcuts are unparalleled, it has invented several UI elements that are now found on other browsers, and its got amazing configuration capabilities

I sense that you are equating "UI" with "features," which is a misguided but sadly not uncommon view (see: Microsoft). To me, the quality of a UI is in its design, and by that standard, Opera fails miserably. From what I've seen, Opera doesn't even make an attempt at visual consistency on any platform, and its "amazing configuration capabilities" only contribute to a bloated and utterly confusing interface. Just compare Opera's preferences window to that of Firefox and Safari to see what I mean.

I consider myself a geek and I find Opera hard to use; I can only imagine how frustrating it might be for someone like my mom. These are people who could give a shit about keyboard shortcuts and configuration capabilities, and your listing these as selling points not only tells me that you're a geek and reinforces the notion that Opera appeals only to geeks, but shows a fundamental lack of understanding of the reasons for Firefox and Safari's success.

In fact, Firefox owes not only its success, but its very existence (as a fork from the bloated Mozilla suite), to the desire for a standards-compliant browser people could actually use. The key to Firefox's popularity has been its ability to appeal to non-technical and power users alike, by providing a well-designed, easy-to-use browser out of the box, as well as a powerful extension architecture for those who want more. The geeks adopted Firefox early on, and pushed it to their friends and relatives as it became more and more polished.

Why didn't Opera step up to the plate when they were the only viable standards-compliant browser around? Being first to the market with features like tabbed browsing is pointless without quality implementation and ease-of-use. I've checked out every major release of Opera since 5.0 and they seem to have made no significant progress in this regard, which is all the more disappointing when you consider that Opera is commercial software, and until fairly recently was a paid product. Shitty UI design is commonplace and at least understandable in open-source efforts, but even Firefox trumps Opera because its core developers understand the importance of good UI design. (Apple, by virtue of the fact that they're Apple, don't even have to try.)

I respect Opera for their pioneering efforts in standards support and I'm all for healthy competition, but until they decide to make user interface a priority and hire some decent designers, they will never be anything more than a niche player. Safari and Firefox aim to be mass-market browsers, not just tools for geeks, and that's why they will never feature what you consider to be "innovative UI."
posted by aqhong at 3:45 AM on March 27, 2008 [1 favorite]


Pinback: I agree with you that the default theme in 3.0 is atrocious (not that 2.0's was much better), but have you checked out Aronnax's themes? They're the only way to go for Firefox on Mac, IMO.
posted by aqhong at 3:52 AM on March 27, 2008


As for adblocking in Safari, I just use this user stylesheet, set in Preferences -> Advanced -> Style sheet. I did the same thing with userContent.css back when I used Firefox. I understand that the Adblock extension offers some features beyond the scope of a simple user stylesheet, but it works well enough for me.
posted by aqhong at 3:57 AM on March 27, 2008


As for adblocking in Safari, I just use this user stylesheet...
Or you could install the Safari Adblock plugin, which is pretty much an exact port of the Firefox extension.
posted by AndrewStephens at 4:23 AM on March 27, 2008 [1 favorite]


Yeah, but some of us prefer not to muck up our system with unsupported hacks (which all Safari "plugins" are).
posted by aqhong at 4:25 AM on March 27, 2008


I think one of the critical points in all is is the actual browser market share. Safari and Opera can happily fight one another out for their respective less than 5% share of the market. As much as people might rail against IE and Firefox, keep in mind that they have a much, much larger install base that are absolutely counting on both backwards compatibility and a near 100% site support margin (specifically IE in the corporate intranet market).

I'll all for standards support but I have to agree with the Mozilla teams stance on this. Simply throwing in support for the latest and greatest in the standards area and potentially breaking support for sites or adding additional security issues would be haphazard. Once FF3 is out the door, I'm sure they'll get back to working on this. Give them a couple of months.
posted by purephase at 5:17 AM on March 27, 2008


It's interesting that the Acid tests are actually creating a "standard reference implementation" as described in the Joel Spolsky article. Not a very useful one, since it's all edge cases and wacky stuff, but it's something.
posted by smackfu at 5:36 AM on March 27, 2008


> Yeah, but some of us prefer not to muck up our system with unsupported hacks

Depends on the nature of the unsupported hacks. I've been running a couple freeware IM plugins on Safari for years, across several major Mac OS X releases, with little or no stress. When an update breaks an input manager widget, remove the widget. No harm.

The problematic IM hacks involve mucking with system behaviors that are considerably more obscured than re-channeling your keystrokes in a web browser text field. Those are dangerous, if not in substance, then in putting you at the mercy of a a software house pre-emptively patching your installation before you change anything on your computer. Some of these sell for real money, or in at least one case were third-party components of a hardware company's device controller -- the hardware company didn't know this would be an issue, and the third party had no way of knowing who the end users were.
posted by ardgedee at 5:47 AM on March 27, 2008


"But her main concern is around the governance of the tests, in this case Ian Hixie is one one who's creating them"

It's my understanding that Ian accepts tests that have been submitted to him and will work them in. (I see a bunch of notes in the source suggesting this is the case.) If that is the case, it would probably be wiser for people making this claim to consider submitting their own tests, including real world examples some seem to think don't exist in Acid3, rather than moaning at the idea of standards tests and who creates them.
posted by aletheia at 7:26 AM on March 27, 2008


Did all the browsers start from a level playing field? Did they all fail every test to start with?
posted by smackfu at 8:09 AM on March 27, 2008


I'm running FF3b4 and vimperator and I'm somewhere between nearly and entirely satisfied. There are some quirks yes and its not entirely stable but it's very, very close to my ideal web browser.

Now if all three could agree on a reasonable standard for plugin development or make switching rendering engines trivial I would physically cry.
posted by Skorgu at 8:10 AM on March 27, 2008


Pinback writes "Give Safari a decent plug-in architecture & Adblock equivalent, and FF will be relegated back to a minority browser on Linux."

You overestimate people's ability to understand technology and fight inertia. I work for a small ISP, and it's all we could do to convince people to use Firefox. Past experience says that trying to get them to switch yet again will take considerable time and energy, and will probably cause confusion among many. Quite a few people still don't really understand what "browser" means.
posted by krinklyfig at 8:21 AM on March 27, 2008


And the implication that Safari is going to take marketshare from Firefox on Windows is just silly. Safari just is weird in Windows, like a tourist from another country who wandered down the wrong back alley into a gathering of locals.
posted by smackfu at 8:26 AM on March 27, 2008


Can we stop arguing about standards and start arguing about Safari installing itself on Windows machines when it violates the EULA, and has some massive security holes with no way to fix them?

Because frankly, I think that's a fight worth having.
posted by blue_beetle at 8:44 AM on March 27, 2008 [1 favorite]


Safari Adblock is a start, though the lack of NoScript for Safari is a deal-breaker. In this day and age, promiscuously executing scripts from any web server one connects to is asking for trouble.
posted by acb at 9:31 AM on March 27, 2008


If that is the case, it would probably be wiser for people making this claim to consider submitting their own tests, including real world examples some seem to think don't exist in Acid3, rather than moaning at the idea of standards tests and who creates them.

I think the problem with Acid3 is that it focuses too much on edge cases and not enough on things that really could move things forward, things like full CSS3 root class and pseudo-class support, SVG/canvas, and generated content. (Of course, part of the problem with CSS3 is we still are working from a 7 year old working draft, and there's been little progress forward on it in the last couple of years.) Unfortunately, IE8 isn't going to do all this, but at least there's hope now that IE9 will. They're starting to care about standards in Redmond, and if they'd spend more time working with the ones they have and not creating their own (I'm looking at you, OOXML), I think a lot of the anti-MS attitude in the web community would subside.

I see Acid3 as the web equivalent of Top Gear -- let's see how fast we can get this Bugati going and whether it can make Turn 3 at Indianapolis at 200 mph while I drink a proper cuppa. Oh look, this browser can do SVG fonts! Yeah, well, FF3 can't. It rates 62/100 in Acid3. Is that all that bad, though, when 99% of the world's browser market is using FF for the browser equivalent of driving to the store -- buying stuff online?

And the implication that Safari is going to take marketshare from Firefox on Windows is just silly. Safari just is weird in Windows, like a tourist from another country who wandered down the wrong back alley into a gathering of locals.

It's a very frustrating experience. I can't get the last version of 3.0 (not 3.1) to work with Vista. I installed it, and then one day it just stopped loading after throwing an error. And I haven't been able to find an effective solution on Apple's site for it.

I did get it working on my XP laptop, but even then I only use Safari on Windows for testing sites in Safari. It's not ready for prime time on Windows, and I think that's just fine with Jobs and Co. They're in the business of selling hardware, not software.
posted by dw at 9:50 AM on March 27, 2008


If I could get something like vimperator for Opera I would die a happy woman. I love Opera. It does everything (except for the vimperator thing, dammit) and it does it well and on my system at least it's a hell of a lot faster than FF3 (my second favorite browser.)

Now if all three could agree on a reasonable standard for plugin development

This.
posted by LeeJay at 9:51 AM on March 27, 2008


Safari 3.1 has a new Develop menu (apparently based on the old Debug menu).
posted by kirkaracha at 9:53 AM on March 27, 2008


Apple seems to think that every app should look like it's running on OSX, even if it's on Windows. So much for UI consistency. fail meh.
posted by blue_beetle at 10:41 AM on March 27, 2008


Apple seems to think that every app should look like it's running on OSX, even if it's on Windows. So much for UI consistency.

I have yet to see a consistent UI on any class of Windows applications. Why should Apple get singled out?
posted by Blazecock Pileon at 11:17 AM on March 27, 2008


He speaks the truth. Apple certainly isn't at all consistent with their apps on OS X. Doesn't Garageband still have some woodgrain in it?
posted by smackfu at 11:47 AM on March 27, 2008


Apple should get singled out for insisting that the UI for their platform is obviously so much better than the Windows UI that every application should ignore any Windows UI standards and just look like it was running on a Mac. This results in stupid crap like Quicktime originally having a brushed metal interface and iTunes using "clear" and "get info" rather than "delete" and "properties". Granted, these issues have been addressed, but they still look weird in Windows. You also get Safari using its own font smoothing rather than the built-in ClearType. Why? Who the hell knows, it's Steve assuming he's better than everyone else. Apple's UI scheme takes the same approach that Henry Ford did with his car colors back in the day. God forbid we know-nothing heathen end users want more options for personalizing the UI.

On the other hand, Microsoft seems to think that standard squarish windows and buttons are old school, and that the new hotness lies in rounded things, transparency and cartoony colorful crap. So take your pick.

I'm always surprised to hear other people having issues with Firefox. It gives me no problems on my Mac (running Beta 4). FF2 on my Windows system at home isn't the best, waiting for FF3 to land before updating, but constant crashing? I'm not seeing that, and I even have quite a few 2.x extensions in forced-compatibility mode. I have all kinds of little problems with Thunderbird, but that's what I get for using pre-alpha nightly builds.

As for the new FF3 skin, my long-time Mac-using cousin likes it, thinks it fits in well. I like it. If anything it's a little too Safari-like, I prefer being able to tell the difference between browsers at a glance when using Expose, and it isn't always easy. Will this please the Apple fanboys? Probably not. There are always some Mac users that will never like anything other than what Apple makes themselves, even though the Apple products are really good at being just good enough without being killer applications. I'm not saying these programs are bad, just that too many Apple users are too content to just use what's already there, and still have the gall to chastise Windows users for not using anything other than what comes pre-installed. (Also, Apple users are always quick to jump all over Mac users that have the audacity to say that all Apple products are NOT gifts from God. They aren't. Get over it. Start demanding better alternatives. Like a functional X11 OpenGL implementation, for starters.)

I'm a Firefox user, and will remain so. I use a Mac at work, Windows at home and Linux on my server. I want as many cross-platform programs as possible. I want to be able to use the same profile cross-platform (and I can, with Firefox). There are too many small issues with Safari that Apple refuses to address to make me want to use it (UI issues, not rendering issues, unless you count the refusal to use native ClearType on Windows a rendering issue). This doesn't mean that I'm totally happy with Firefox, though. There are always reasons to ask for improvements. No single web browser (or any program, really) is good enough for everyone. There will always be a reason to want alternatives. The Acid tests are helping to ensure that browser developers keep making improvements to ensure that alternatives are usable. So, hooray for the Acid tests, even though we all know that they're mostly just for bragging rights.
posted by caution live frogs at 12:19 PM on March 27, 2008


caution live frogs writes "On the other hand, Microsoft seems to think that standard squarish windows and buttons are old school, and that the new hotness lies in rounded things, transparency and cartoony colorful crap. So take your pick."

Well, a good rule of thumb is that the application should refer to the window manager for widgets and look and feel, rather than forcing its own widgets, etc. into an existing wm scheme. Apple is notoriously bad about forcing their own styles onto whatever you're using. But I think you're right: Apple probably feels that its look and feel is as much a part of the application as the rendering engine, because that's how they sell their products. I don't have to like it, however. On a Mac, it helps the whole OS look uniform and clean. On any other OS, the opposite effect is produced.
posted by krinklyfig at 12:39 PM on March 27, 2008


Apple should get singled out for insisting that the UI for their platform is obviously so much better than the Windows UI that every application should ignore any Windows UI standards and just look like it was running on a Mac.

Then Roxio and Adobe and every other software company should be given the same Metafilter hate, because their applications certainly don't follow a majority of the Windows UI guidelines, whatever they are.
posted by Blazecock Pileon at 12:45 PM on March 27, 2008


The old square Windows look (well, since Windows 95) is basically a low-rent knockoff of the UI of Steve Jobs' NeXT workstation. Why shouldn't Microsoft rip off Jobs' new baby as well?
posted by acb at 5:18 PM on March 27, 2008


... and now Safari's the home to an 0-day root privilege hack at CanWestCon (here)

Just wish Firebug could get updated so I don't have to keep switching back to FF2 when I want to run it.

The 1.1 beta runs fine.
posted by bonaldi at 6:33 PM on March 27, 2008


The 1.1 beta runs fine.

I had not noticed that. Thanks!
posted by dw at 9:11 PM on March 27, 2008


Blazecock - Adobe at least uses the standard Windows "keep all palettes in the main application window" thing, which annoys me to no end on a Mac as I'm not used to the floating bits. I hate clicking through to the desktop on accident and having my application lose focus.

But yeah, there are a lot that break the rules. And there are rules, really. (Of course those are for Vista; Microsoft has helpfully removed all information about WinXP UI guidelines, because they want developers to ignore WinXP.)

The programs that break the rules the most often (aside from Apple products) are media players, Windows Media included. The default skins for most media players are so godawful ugly and awkward that I can't understand why they were built in the first place. At least Apple is consistent with their weirdness To be fair, Adobe is too - Photoshop on a Mac has the same menu options and toolbars as the Windows version, in the same places, with only minor differences in keyboard shortcuts. More on topic here, that's one of the major reasons I like Firefox. I have the same options on any platform. The developers of the program have to build it such that it can look as much a part of the host OS as possible, without doing so in a way that makes it unusable on another OS. This means in practice that it may never completely fit in on one system; the cost of this is worth the cross-platform capability in my mind.
posted by caution live frogs at 8:24 AM on March 28, 2008


Adobe at least uses the standard Windows "keep all palettes in the main application window" thing, which annoys me to no end on a Mac as I'm not used to the floating bits. I hate clicking through to the desktop on accident and having my application lose focus.

Odd. It pisses me off when Windows doesn't work like my beloved OS X.

Maybe I shouldn't expect it to behave like OS X.
posted by five fresh fish at 6:43 PM on March 28, 2008


« Older MindHabits...  |  Dinner With Darwin. Scientists... Newer »


This thread has been archived and is closed to new comments