This post made me cry.
Thank you mark.
— Masklinn 
I love Firefox to DEATH. They’ll have to pry my dead, cold hands off my Firefox, as they will with Ubuntu Linux. But I tell you, you hit the nail on the head here and Mozilla could do well to win this game rather than to run from it.
I’m personally much more interested in the memory and performance enhancements that are being made in Firefox 3. Sure css is important but usability comes first, and with all the memory issues plaguing Firefox 2, they sure as hell made the right choice on prioritizing that before Acid 3. I mean the rush to support acid 3 wasn’t even about standards compliance, it was just about beating acid 3. That’s just something to brag about, but has nothing to do with endurance. At the end of the day the user is going to care about the browser that runs better, not the one that renders css to the pixel.
— Tim
Tim: Maybe you should give IE7 a try!
— Jeroen
This is spot on. Mozilla is no longer leading the web standards race, and it is purely from a lack of trying. Then again, Mozilla has matured and this is the response from a mature organization.
Both Opera and Safari have rushed to implement a ton of new features, and I’m pretty sure the QA on these cannot have uncovered all the potential security implications. Personally, I’m excited about that part.
Jeroen: Now there’s a joke!
Have you compared Firefox 3 and IE7 lately? IE7 is the tortoise to Firefox’s hare.
— BizarroJeroen
C’mon Jeroen that Firefox 3 that crashes my computer so offen that I started using Opera?! It’s no use that is fast if you can’t use it!!! :(
— Cristian
Cristian, what computer are you using that a user-space application can crash it? My Principles of Operating System Design from as far back as 1970 says that this should be impossible.
— Shevek
Mozilla came “last but one” in the Acid 2 test as well, and exactly the same complaints were made at the time. There still isn’t a production-quality Mozilla release which passes Acid 2. And despite having petulantly ignored the last bad report completely, the project seems to have trundled on quite well, picking up its raises / promotions / bonuses just like before. Meanwhile, MoFo have concentrated on browser performance, which didn’t even get brought up in its last review, and made substantial progress.
This is a really Spolsky-style post. I assume that if I turned my Adblock off right now I’d see you’ve installed AdSense or the like.
– Chris
When I said crashing in fact I meant that the computer is unusable for 20-30 seconds and Firefox is not responding anymore. If I let Firefox in this state, I can see that is eating away my resources and I have to kill it.
Sorry for the misunderstanding.
— Shevek
Shevek when I said crashing in fact I meant that the computer is unusable for 20-30 seconds and Firefox is not responding anymore. If I let Firefox in this state, I can see that is eating away my resources and I have to kill it.
Sorry for the misunderstanding.
— Cristian
Cristian, Id like to point out that Firefox 3 is BETA and yes, it will likely crash.
— Cristian
It’s very ironic to see read the same excuses from Mozilla about Acid3 that we’ve heard from IE team about Acid2.
— kL
I’d like to point out that current webstandards are *imposible* to implement in a reasonable way. Spending time and efforts on acid test instead of revising current standards makes me cry.
— nsz
I’m glad someone said this. The Mozilla behavior around this was frankly disappointing.
There was also the minor manufactured controversy over Webkit turning on SMIL support in the trunk “just to pass ACID3″. A mozilla blogger pointed to this blog post and the SVG test suite results contained within it.
http://blog.codedread.com/archives/2008/03/26/webkit-nightly-not-smiling/
(see the comments and, notably unlinked, follow up here: http://blog.codedread.com/archives/2008/04/15/webkit-nightly-now-smiling/)
They were trying to show their better support for standards but apparently didn’t notice that the results showed that Webkit (and released versions of Safari) had maintained a steady lead over Gecko in those tests for quite a while.
http://www.codedread.com/svg-support.php
Doesn’t it say a great deal about the current state of the web that actually implementing the corner-cases of specs and checking that your implementation is interoperable is relegated to a once-a-year carnival side-show rather than being, you know, their job? And in Mozilla’s case, part of their Mission statement, principle number 6 in fact:
http://www.mozilla.org/about/mozilla-manifesto.html
— dave
So, I just checked. On Ubuntu, my firefox 3-beta-5 scored a 71, while Opera is stuck at 46. What’s all this commotion?
— danny
Regarding Chris Cunnigham’s response: funny you should mention that because when people (admittedly probably totally uninformed people) complained massively about Firefox having memory problems the original response seemed to be that they simply didn’t exist. Then it was just caches and therefore a “feature”, later the story was that it was only problem add-ons, and therefore it wasn’t technically a problem despite the add-ons being a major feature of Firefox and in wide use. Finally it appears that memory fragmentation has been pegged as the main cause.
Eventually the problem was admitted (the proverbial first step) and a lot of work and cleverness was applied and the results have apparently been very impressive. But again, the initial response wasn’t it was lots of techy quibbling about what technically consitutes a ‘leak’ that the end user had no interest in. Some have suggested that only when the techy challenge of the mobile space arrived did anyone actually start to pay attention to this issue. And maybe, just maybe, if those had been dealt with earlier then firefox 3 might have been the one with the ‘free time’ to pass Acid3. I don’t really understand why “I was too busy fixing our memory problems” is considered a valid excuse.
— dave
@danny: You should be comparing apples to apples. You are comparing a release version of Opera (9.2x) to a beta of Firefox. Opera 9.50 beta 2 gets 79. You can get a public build of the version that passes Acid3 here: http://labs.opera.com/news/2008/03/28/
— Arve
The reason why Firefox doesn’t do well is not because Mozilla is lazy. Firefox 3 is just about to be released, and not only do the developers have no time, but it is not the appropriate time in their development cycle to fix major things like what is needed to pass acid3.
— John Kloosterman
There was a time back in 2003 where I fell for the idea of standards, and I bought Zeldman’s book and I visited the Zen garden and I read CSS and design blogs. It was so cool. I was already 40 by that time, a veteran of Fortran IV and Turbo Pascal and countless other oldies, so please believe me when I say it takes quite a bit to impress me.
Five years have gone by. Ajax has impacted everything to the point that the hours-reporting app I’m forced to use, which re-loads the page every time I submit something, looks antiquated and disgusting. That’s one development.
The other lingering issue is that bank sites (and off-track betting software, etc) still require Internet Explorer and won’t function at all with any other browser.
And these are the ones that matter. The bank is requiring me to own a Windows+IE machine if I want to do business with them at all. And in the corporate work, most Java enterprise software also “requires” IE even if it will work with Firefox or Safari. Of course that’s just laziness and cost cutting on part of whoever is developing the stuff, but it’s a reality.
I’d appreciate it greatly if Firefox concentrates on market share, thus forcing corporations to support browsers other than IE. If this comes at some expense to the web standards cause, so be it: you must understand that some of the people defending you are deeply imperfect people.
— Salomon Toribio
Acid3 is pointless to pass, perhaps they’ll make a real test sometime this decade that might be of any use.
— Jackie D
Dave, using the term “admitted” implies you think that Mozilla knew about a memory leak and intentionally tried to it secret. Mozilla has a bunch of people working on a really complicated product. The way I saw the memory leak stuff was more like somebody looked at it and figured out one answer. Then people kept asking and someone at Mozilla looked more deeply into it and corrected the previous analysis. There’s nothing shady there, I work on software too and a lot of people come to me with wild-ass guesses about what’s causing a behaviour they’ve observed. I try to figure out what the problem is and sometimes it takes more than one try.
As to the topic at hand, the Acid 3 test, I agreed with Mike on the post Mark linked to and still do. The ability to throw bug fixes into code is inversely proportional to the size of the user base. There are a lot of users out there who rely on Firefox to be stable and tweaking something to meet a standard can have unintended consequences on other parts of the code. That’s why bug tracking software exists – it helps to keep priorities in order.
The best thing that the Acid test could do for Firefox in my opinion would be to get some outside people interested enough that they’d try to come up with patches to fix what looks like low-hanging fruit (like the bug Mark linked to). Then those patches could be evaluated just like any others for inclusion in the current development branch. I would’ve lost a lot of respect for Mozilla if they’d just suddenly pulled people off of features and fixes that they’ve committed to deliver just to jump at making a test score higher.
I don’t see any reason for Mozilla to be embarrassed. They made a plan and they’re executing well. Acid would be a good motivator to refocus on cleaning up the bugs that it demonstrates, sure, but that belongs in planning for the next release and work done by people who aren’t already committed to current plans.
The pace of Mozilla development does frequently seem to be rather glacial, and the following process seems all too common: the bug doesn’t get fixed, people get cranky, and the developers get all defensive and excuse-makey when something’s been hanging around for ten years. (And I’m not even going to start on the “I want to stab my face with a fork” sensation I get when looking at the infamous tooltip-wrapping bug.)
It should be noted that Mozilla is hardly alone in having bug handlers (supposedly paid bug handlers) who’d rather tell the users how mean they are for complaining and how very lucky they are to have such wonderful developers who are doing very important things. There’s a shiny example over at Red Hat’s tracker. (Someone lost his job over that one.)
I wonder if anyone keeps a catalog of the most forehead-slap-worthy bugs in the Mozilla DB. I’m sure there are plenty.
Bookmarked. That was that kind of post.
— Meneer R
> that belongs in planning for the next release and work done by people who aren’t already committed to current plans.
There are 2 axioms in politics:
1. You can’t get any real work done in an election year.
2. It’s always an election year.
Re-reading Shaver’s rant once more, I still don’t get the impression that their reaction would have been any different if the test had been released at any other point in their development cycle.
Look, nobody expected this “race” to perfection, least of all the people who compiled and published the test. (They’ve said as much, publicly. They didn’t expect anyone to pass fully for YEARS.) But it happened, and it accomplished exactly what the test set out to accomplish — getting people to talk about web standards instead of all that other shiny stuff that’s backed by powerful vendors and a worldwide PR machine. Standards don’t have that; they’re not sexy, and small bugs drive developers crazy, and big holes and missing features — which someone took the time to spec out and which lots of people think would be a good idea — make the web overall less competitive against those shiny vendor-specific runtimes with the big PR blitzes.
The Acid tests are brilliant PR for making standards buzzworthy, even if it’s just for a few minutes. And this is how Mozilla reacted, publicly. *That’s* the missed opportunity here.
— Mark
dude, you just pissed my firefox off so bad, it’s not eating it’s food, and it growls when I get near it.
Rob Russell,
I believe you missed the point in Dave’s post. The point is that users complained about something and the Firefox development community got defensive and argued that the user was at fault by either not understanding how memory usage is reported or by installing bad 3rd party add-ons. As Dave noted, we are now told that Firefox 3 fixes memory problems, the same problems we were told weren’t problems and that we were just dumb.
I think the point of all this discussion is that some of the key Mozilla developers need to stop denying that problems exist.
— Travis
I have to say Mozilla made the right call here. Yes, the standards issue is important and Firefox needs to be able to handle anything the web designers of the world can throw at it, but this will affect far fewer users than its tendency to lock up and cause huge memory leaks. Additionally, standards compliance is exactly the kind of issue that can best be fixed with test data gathered in a genuine working environment on as many computers as possible, which is the whole point of beta versions; it’s neither fair nor helpful to judge them against production software in a test like this.
Besides, seeing as there’s no apparent consensus on the right way to implement standards compliance, I can’t really blame Mozilla for being somewhat disparaging of a test that’s designed according to someone else’s idea of best practice and apparently specifically intended to make those taking it look bad.
— Jake
>The funny thing is, I don’t even know who won.
Easy… Safari team won.
— mikeb
Please Microsoft, please pass Acid3 before Mozilla. It would be the best publicity trick you could pull for years.
— Peter
When I was in college, there were always those few people in CS classes that were brilliant but couldn’t take a test to save their lives. This wouldn’t have been so bad if they didn’t spend the few days after getting a graded test back complaining about how stupid or unfair the test was. It never occurred to them that this made THEM look bad instead of the test, especially when nobody else seemed to have a problem doing well on it.
Had Mozilla simply said, “Congratulations to these other teams for all of their hard work. Unfortunately we’ve got some more pressing issues to work out on Mozilla right now, but we’ll join them at that party soon enough.” or even simply kept their mouths shut, they could have saved face. Instead they drew more attention to their own shortcomings, some of which apparently have more to do with personalities than code or project management.
— Adam
“I’d like to point out that current webstandards are *imposible* to implement in a reasonable way.”
This lovely little piece of BS comes from the ambiguities inherent in human languages, which are used to write specifications out of necessity. In most cases, a single common-sense interpretation of an ostensibly ambiguous phrase is obvious, but not always. Test suites clear up this ambiguity, by providing a single and clear interpretation of what the standard intends: either you pass or you do not. These are the sort of things developers can depend on.
This is is why future standards will be defined not by specifications, but by the test suites based on those specifications. Test suites like the Acid series, which still have some growing to do but which have been vitally important in the advancement of standards thus far, will only become more prominent. Our standards bodies are still lacking in their understanding of this, but at least some of the browser developers are starting to understand.
— Millennium
If these count for anything, it’s for personal growth.
I love this line..
“And yet, if you rounded up all the managers in the world and shot them… no wait, that’s not where I was going with this.”
Great paragraph opener.
What’s so great about Acid 3 is now there are two browsers that account for 5-10% of web traffic that have partial support for a bunch of things like SMIL and SVG, and will break when people try to implement something that uses the entire feature-set. These breakages will just as likely turn people *OFF* from using standards and send them into Adobe’s and Microsoft’s monocultures that “just work.”
The arguments I’ve seen coming from Mozilla people is not that the Acid 3 is a bad test. Their anger is directed towards the implementors who are simply coding to the test and leaving out much of the full spec. Once Firefox 3 is released, expect to see work towards full implementation of these standards that will allow people to actually use them and actually increase adoption.
PS – I only include Safari’s and Opera’s market share as an explanation for why content creators might not spend much time testing in these browsers.
— dustin
when people (admittedly probably totally uninformed people) complained massively about Firefox having memory problems the original response seemed to be that they simply didn’t exist. Then it was just caches and therefore a “feature”, later the story was that it was only problem add-ons, and therefore it wasn’t technically a problem despite the add-ons being a major feature of Firefox and in wide use. Finally it appears that memory fragmentation has been pegged as the main cause.
Actually, the people whose jobs involved reducing Mozilla’s memory consumption never denied that Mozilla was RAM-hungry. Random people on Planet Mozilla did, which is of course good enough for the window-lickers who make up most of the commentariat. The real logical fault here is assuming that the people who whined about memory leaks for years without ever being able to reproduce them were “right” in some way that made the direction taken by Mozilla “wrong”.
It is in fact arguable that because Mozilla’s historically high memory usage pushed Apple into adopting khtml and bringing serious manpower to it in the form of Webkit that development vitality in the browser space has increased because of the choices made in the past. But this would still be completely incidental to why those decisions were made, and should be ignored for the same reason.
Re-reading Shaver’s rant once more, I still don’t get the impression that their reaction would have been any different if the test had been released at any other point in their development cycle.
If only there were some way of checking whether that stood up to scrutiny, such as roughly seven years of rhythmic development cycles which followed exactly the same approach to tackling new rendering problems as was professed by the current project driver. It’s such a pity that such things don’t exist, or do so only on the hard disks of dusty workstations somewhere in Silicon Valley. Maybe if one day these workstations were attached somehow to a wider network – possibly even an international one – it would be possible to gain a better idea of how truthful such ranting was.
– Chris
Hmmm. That first paragraph above was definitely blockquoted when I posted it.
dustin: Once Firefox 3 is released, expect to see work towards full implementation of these standards that will allow people to actually use them and actually increase adoption.
Ooh! Will there be ponies too?
I’ll remain unconvinced unless you can provide me with some compelling reason that this release will be the one where, for example, @font-face will work. That bug is over seven years old. Since early 2001, Gecko has gone through eight releases (1.2, 1.3, 1.5, 1.6, 1.7, 1.8, 1.8.1, 1.9); Firefox was created and has gone through four major release-quality builds (1.0, 1.5, 2.0, 3.0).
Your attitude is far too Friedmanesque for me to stomach. All you need to do now is assure everyone that the next four to six months of Mozilla development will be critical.
The Safari people are so focused on padding the acid3 test, that they’re neglecting features that end users would actually find useful, like anti-phishing, which IE, Firefox, and Opera all have.
— Sean
Peter: Please Microsoft, please pass Acid3 before Mozilla. It would be the best publicity trick you could pull for years.
Ha. Anyone want to write up Mozilla’s likely reaction should that happen? I suppose it wouldn’t look that different from Shaver’s rant.
Thanks for doing this, Mark. These things really needed to be pointed out. Mozilla has definitely been improving, but compared to Safari, it’s been stagnating.
I don’t understand the commotion
— oomu
Yes, there’s a big difference between Mozilla’s official responses and what would be desirable. It’s one thing to say “wow, we’re at the final bug-squashing moments of a major release, so we can’t make Acid a big priority today, but it’s great to see Opera and Webkit moving so quickly on standards, something we’ve always considered critical to keeping the web browser-independent” and what they did say, which was basically “wow, Acid3 is out and nobody cares about it and anyone who does care doesn’t know what they’re talking about and we’re really busy over here in case you didn’t notice so why don’t all of you plebes just shut up and be happy with what we give you?”
— Nathaniel
Just to remind you–Mozilla is open source–go fix it & stop bitchin’.
— jartur
grendelkhan, I can appreciate your frustration with the @font-face bug. It’s an interesting feature, and it seems like it could advance the web. However, there are some complexities there. Remember the reaction of users when they found out cookies took up their valuable hard drive space, and just to keep track of them? Custom fonts with potential security problems (recall that fonts go in a system folder on Windows) that take far more space than a cookie, and you’ve got an issue that must be treated with special care. Should the font be downloaded and stored on the system? Should downloaded fonts simply stay cached until the session ends?
— dustin
Shaver’s blog wasn’t the only Mozilla response to Acid3. I wrote http://dbaron.org/log/20080406-acid3 , and I said pretty clearly that we are working on the bugs. (And I have been since then.)
I personally have spent a lot of time over the past few years working on memory use — time I could perhaps have spent on new Web-facing features. And then in the past six months I’ve been spending most of my time working on bugs we needed to fix to get Firefox back to a quality level we’d be comfortable shipping (since we made the mistake of letting our development trunk get too far from that quality level — a mistake I think we’ve learned from). I take this criticism very seriously — but you haven’t convinced me that those were the wrong decisions for Firefox or for the Web. And I don’t think you should extrapolate our plans for the future from one release that took longer than it should have (though I think still very good in the end).
Finally, I strongly disagree with your statement that it’s the followup that matters. What you’re referring to as the followup is the public relations, not the substance of what we do. I think what really matters (as I said in the blog entry above) is our long-term commitment to moving the Web forward — not only in terms of features — but in terms of being built on clear, open standards that allow and encourage competition and thus innovation in the long run.
Who gives a damn about a test that proves NOTHING.
I still have margin issues between IE FF OP and Safari. Who appointed Ian Hickson or WASP as standards and practice gurus? Web standards and any Acid test are nothing more than “Well formed but not valid.” Which is turning out to be the motto of imperitive web developers everywhere.
— Don Ulrich
Don Ulrich:
The test ensures cross-browser compatibility. It’s basically one huge testcase, and testcases ensure an interoperable web.
This comment by “Millennium” hits the nail on the head:
In most cases, a single common-sense interpretation of an ostensibly ambiguous phrase is obvious, but not always. Test suites clear up this ambiguity, by providing a single and clear interpretation of what the standard intends: either you pass or you do not. These are the sort of things developers can depend on.
This is is why future standards will be defined not by specifications, but by the test suites based on those specifications.
QFT.
— Robert Firee
For the record (Mark doesn’t quite make this explicit), the Ahem-specific hack was removed. See http://ln.hixie.ch/?start=1206756775&count=1
— Dobb
I’d like to put Don Ulrich and Joel Spolsky in the same room. Perhaps there’d be some kind of mutual annihilation.
dustin: grendelkhan, I can appreciate your frustration with the @font-face bug. It’s an interesting feature, and it seems like it could advance the web. However, there are some complexities there.
Yeah, that’s really fascinating and all, but I’m not buying it when you tell me that the problem is so complex that it takes at least seven years to get it right, especially when very little actual time has been spent working on @font-face support. The Firefox dev team spent the last seven years ignoring @font-face (and don’t think that I have some kind of special affinity for it; I could pick a variety of other standards) and explaining that it’s just really hard while Webkit and Opera are actually implementing it.
I’ve read plenty of comments over at the bugzilla there. I’ve seen enough excuse-making to tide me over, thanks.
jartur: Just to remind you–Mozilla is open source–go fix it & stop bitchin’.
Or, in other words, if you’re not a hacker–no, scratch that, if you’re not a C++ hacker well versed in Mozilla internals–your opinions don’t matter, so go cry somewhere else, emo kid.
I appreciate the DIY ethic of open source software; I really do. I appreciate how it helps define a better way to interact with developers who are doing things in their spare time. But Mozilla has a professional development team with employees who are paid real money, and the vast majority of their end users are not hackers.
If Mozilla’s goal was to provide a web browser solely for experienced Mozilla hackers and for end users who never want their software to do something it doesn’t already do, I wouldn’t have a problem with what you’ve said. But that’s not their stated goal, and I don’t think the answer to complaints about standards and interoperability is to tell people to “stop bitchin’”.
“you should all be embarrassed with yourselves. But you’re not, so here I am stepping up, publicly being embarrassed on your behalf.”
While I agree that Mozilla as an organization has reason to be embarrassed by the statements made about the Acid3 test. I can’t help but think what an arrogant a$$ you’d have to be to embarrassed on their behalf. I don’t think you’ve earned that right.
— steve
> I can’t help but think what an arrogant a$$ you’d have to be to embarrassed on their behalf. I don’t think you’ve earned that right.
How many more years of work on Mozilla projects would Mark have to put in to earn that right?
— Kevin H
> I can’t help but think what an arrogant a$$ you’d have to be
You must be new here.
— Mark
What did you make of Eric Meyer’s piece ( http://meyerweb.com/eric/thoughts/2008/03/27/acid-redux/ )on Acid 3?
(For some reason I initially mistread it’s title as Acid Reflux)
— Sits
>>The funny thing is, I don’t even know who won.
>Easy… Safari team won.
Nope, Opera won. Opera is the first to have an internal build pass the test (even after the test being fixed last minute, they fixed the internal build within minutes, hours before WebKit). WebKit (not Safari) is the first to get a public release build pass the test. Safari does not pass the test.
— easymike