Firefox Lead Engineer Scolds KDE Project

trent42 writes "Firefox lead developer Ben Goodger has had harsh words on his blog for the KDE project, in light of its public tiff with Apple over the KHTML rendering engine. Goodger says 'Safari's renderer is vastly superior to the KHTML used by Konqueror,' and that the KDE developers should follow Apple's lead and focus more on the needs of users, instead of insisting on software perfection."

Comments

Agile

[mfh]

So basically, KDE should read this.

Re:Agile

[Ozwald]

Other way around, dude. Agile methodologies come from open source ideas, like release often, listening to customer (or other developers).

Remember what it was like before Agile? Companies and consultants would develop big blocks of software, check it in, QA it, and show the customer who'd get pissy because it didn't work the way they expected. Yes, Agile prevents that. But seeing what is happening, I'd say that KDE does is very Agile, unlike what Apple just did.

Oz

Re:Agile

[robertjw]

Tricky thing, those numbers. Every machine Apple sells comes with OSX, 'for free'. They may not be selling their software, at least from an accounting perspective, but their hardware would be rather worthless if the price didn't include an OS to run on it.

Direct software sales may only be 4%, but software is a much larger part of their business than just the revenue percentages indicate.

Can't wait...

[ElGuapoGolf]

Personally I can't wait for the KDE response which scolds the Firefox developers for having such huge and stupid security holes in their browser.

Maybe the Firefox team should get rid of the glass walls before they start chucking stones at other people.

Re:Can't wait...

[LWATCDR]

Actually KHTML has the biggest security hole of all. It does not follow current standards so there are time when you HAVE to use another browser. So in effect you have any security holes in KHTML and what other browsers you are forced to use.

I actually like the KDE browser better than Firefox. I love the built in spell check and it is fast.
But I can not use it with Google maps or the full version of gmail.
Has Safari introduced any huge security holes? The latest Firefox hole seemed less that huge to me. Yes it could be exploited by a white listed site but the only white listed site I have is Mozilla.org
It was also patched very quickly.

In a way I agree

[baryon351]

the KDE developers should follow Apple's lead and focus more on the needs of users, instead of insisting on software perfection.

In a way, I agree. It's comforting to sit down, load an app, and have everything work. Knowing it's not quite perfectly written behind the scenes is a small worry sitting in the back of my mind, but it's smaller than when I have a slightly clumsy app that is otherwise technically correct.

Not that I think Konq is all that far behind in the user side of things.

Re:In a way I agree

[l3v1]

Knowing it's not quite perfectly written behind the scenes is a small worry sitting in the back of my mind

Ok, so that sounds like IE's early days. I say "early days" because its flaws are nothing less than eyepopping these days. Anyway, I don't care how well Safari works and how good or bad it is or isn't behind the scenes. What I care for is that Konqueror is very well written, very stable and very fast. I use Konqueror (for browsing) about as much as Firefox, maybe more. I really think the Konqueror guys deserve every bit of appreciation for their long great work. I wouldn't like KHTML being dropped in favour of an engine hacked together by Apple devs.

Re:In a way I agree

[mmeister]

I don't think there is any real evidence that Safari's WebCore engine is "hacked together" by Apple.

The patches submitted back to KHTML may be harder to integrated more because the changes made to Safari are greater and in a different direction than KHTML.

I think there is a bit of arrogance on the KHTML side to not even consider the aspect of WebCore.

The holier than thou attitude seems very pervasive in the Open Source Community. It's not unlike the Not Invented Here syndrome that many corporations suffer from.

Apple is offering up their changes but seem to have said "We've made major improvements that can't easily be patched in to the existing base. We offer the opportunity to use this new code as a basis for the future."

Re:In a way I agree

[Jason Hood]

Also keep in mind that KHTML and Konqueror are two separate projects. Konqueror is probably the most usable browser of all time from a functionality standpoint. User interface wise, it could be better. but its no worse than FF or IE. Once gecko is properly ported as a kpart, konqueror will be extremely powerful. being able to switch between renders on the fly, saving sessions will be awesome. It may actually kill KHTML, although KDE devs firmly deny that =)

This argument made more sense 3 years ago. Then KDE was laying the foundation for 3.x and focusing little on the UI (In my opinion). Now with Qt4 (seriously, read the new featureset) and KDE4 due out in a year, KDE/Konqueror/KHTML is going to get a major boost. In case anyone hasnt noticed, KDE and Gnome have been exploding in the last 2 years.

Now if I could only get a Mac clone theme for KDE ;)

Re:In a way I agree

[spencerogden]

It think so far the KDE frame of mind has paid huge dividends. 3-4 years ago it was entirely unclear who was building a better desktop, Gnome or KDE. KDE it seems went on a spree of doing hardcore behind the scenes work, with a focus on getting APIs and implementations correct. It felt like KDE was stagnating for a while. With 3.2 and now 3.4 KDE users are reaping the rewards of that effort. New applications are developed quickly, the interface is consistent, and system wide usabilty changes work for everyone.

I think one of the strengths of Open Source is that developers are not under economic presure to deliver it yesterday. They have usually taken the approach of getting it right. I think this means products are sometimes longer to market, but its a trade off. Its one that can be made in open source, but isn't always availible to a commercial developer, they often need code now, correct or not.

Re:In a way I agree

[BlueStraggler]

Your post would be much easier to understand if your sig came first.

Re:In a way I agree

[iamacat]

Along comes some big company that grabs your code, renames it, and puts it into their product publically. They submit some patches. Then, they basically stab you in the back.

Slow down cowboy, I think you are talking about SCO here, not Apple! Someone creating a new project based on your free code is a compliment, not a stab in the back. KDE developers should just take any portion of Apple's changes they find useful and let the rest stay in a separate project. If later Apple wants to take advantage of improvements in the new version of KHTML, they will then do the work of integrating their own dirty patches without any prodding.

In the meantime, the rest of us can hope for a GNUStep webcore-based browser and saying goodbye to bloated Linux desktop environments that try to look like Windows.

Re:In a way I agree

[codemachine]

The problem being that KHTML is not platform independant either. It is quite dependant on KDE. What should really happen is that Apple and the KHTML team together write a portable rendering engine that could be used by both projects (ala Gecko). Unfortunately Apple doesn't seem interested in doing that, and the KDE team isn't necessarily that interested in it either. Too bad, because it woud benefit both groups to have a standard rendering engine, just like it has benefited all of the Gecko-based browsers.

You know that if something renders in Mozilla, it'll probably work in Firefox, Galeon, Netscape, etc. This is great for both web developers and the browser teams, as it reduces the amount of testing needed (it especially helps the little-known Gecko-based projects that would never get tested against themselves). The KDE project could benefit hugely from having a truly shared HTML rendering core with Safari, as large developers such as Google already make their pages avaiable in Safari but not Konqueror. Fragmenting KHTML/WebCore only makes both less useful to test against, though this hurts KDE much more than Apple.

Re:In a way I agree

[AusG4]

Ok, so that sounds like IE's early days. I say "early days" because its flaws are nothing less than eyepopping these days. Anyway, I don't care how well Safari works and how good or bad it is or isn't behind the scenes. What I care for is that Konqueror is very well written, very stable and very fast. I use Konqueror (for browsing) about as much as Firefox, maybe more. I really think the Konqueror guys deserve every bit of appreciation for their long great work. I wouldn't like KHTML being dropped in favour of an engine hacked together by Apple devs.

I think you're missing the bigger point here....

Yes, KHTML is "well written, very stable and very fast". But so is WebCore, which is obviously derived from the same KHTML tree that you care for so deeply... but WebCore is vastly more capable. Sure, the KHTML guys deserve recoginition for their work, but to characterize Apple's fork as "hacked together" is a gross misunderstanding. The WebCore engine is clearly the superior technology and Apple's developers are clearly responsible for the progression that WebCore has made over KHTML.

The reality here is that this whole mess is nothing more than KHTML's developers wanting to have their cake and eat it to. They welcomed Apple to the table with the hopes of some full time developers helping out with KHTML, but then poo-poo'd Apple's efforts when they realised that Apple was foolishly committed to solving problems for their customers, rather then just writing pretty code.

This is one of those problems that happens time and time agian with open source projects - the developers become so consumed by making a technically superior product that they forget to deal with the fact that it's functionally underwhelming. There are a choice few exceptions to this rule... great sucess stories no doubt (Linux and Apache come to mind)... but they are certainly the exception, not the rule. Case in point... the Gimp. If I hear one more zealot even try to compare it to Photoshop.... No doubt, the code to the Gimp is probably cleaner,better written, and less prone to memory leaks.... but it doesn't change the fact that Photoshop is light years more advanced (4 letters: CMYK) and a lot more elegent to use.

Of course, what really bothers me is when these inadequecies are overlooked by zealots who disregard ease-of-use and functional elegence because they appreciate the idealogy of the developers. What kind of brain-dead reasoning is that? If "poorly" designed code -works better- for the end user, than it's not so poor afterall. This is the key point the KHTML people have missed.

At the end of the day... If the Konq guys absorbed Apple's changes, rather than crying about them, you certaintly wouldn't be complaining that suddenly Konq was a whole lot better than it was -before- Apple got involved, now would you?

Take Notes

[PaisteUser]

Now only if Microsoft would insist on software perfection....

Re:Take Notes

[ceeam]

Why, it is _perfectly_ crappy already.

Why not

[mattmentecky]

...and focus more on the needs of users, instead of insisting on software

Why not try something completely the opposite, like Microsoft, and focus on neither?

Ben Goodger's blog URL

[Manip]

http://weblogs.mozillazine.org/ben/ Who knows why the poster linked to a ZDNet article (Which incidentally can't handle a slashdotting) instead of the original blog.

Uh..

[Thyamine]

So the two are mutually exclusive? We can only have software that is perfectly written or software that addresses the needs of the users?

Can't we figure out what the users need, and then deliver excellently written software to do that?

Re:Uh..

[HomerJayS]

This is the classic software development dilema.

It can be developed quickly, cheaply, or correctly (but you may only pick two of the three options)

Re:Uh..

[Skye16]

I disagree. It's much harder to clean the code after it's already implemented and integrated. Do it right the first time and you don't have to worry about it later. In the mean time, you have a stable, secure product that people can rely on, even if they don't have the latest and greatest features.

Re:Uh..

[drew]

Of course, Apple hasn't really gotten into any significant agreement with the KDE people either. People who have been paying attention may have noticed the KDE developers saying that although Apple hasn't cooperated with them as much as the might have liked, they are within their rights to do that.

For all that has been said about a feud between KDE and Apple, the real feud is between the KDE developers and the users and slashbots who think that any new features in Safari should be in KDE too, and if they aren't it's because the KDE developers are slow, lazy, whatever.

It's worth noting that (from what I've heard, at least) the other open source projects that apple has used code from haven't gotten back much more in the way of useful contributions than the KDE team has.

Re:Uh..

[OglinTatas]

"...quickly, cheaply, or correctly..."

This is not quite that software dilemma. Lifting something quoted in an earlier post:

"the KDE developers should follow Apple's lead and focus more on the needs of users, instead of insisting on software perfection."

This is about writing something that is correct to spec by the design document (i.e. the needs of the user--or maybe it is about getting the design document right?) vs. technically correct (i.e. perfect software--correct to the language spec, elegant, robust, no flaws, easily readable and maintainable, etc.)

Ideally, one's software would be both of these, and thus fit in the third category "correct".

I therefore propose that this old adage be modified thus: "On schedule, on budget, on spec, or few flaws--pick two"

Feel free to change those words to make it flow better. OT: in my limited experience, the scope of the project is always changing, so really none of those apply, therefore one can only really try to achieve #4, as few flaws as possible in what actually does get done.

Re:Uh..

[EastCoastSurfer]

Do it right the first time and you don't have to worry about it later.

Assuming you know the right way the first time you do something. Software isn't static, and nearly all requirements are fluid. By the time it's done right the requirements have changed and what you've done is no longer what's needed.

User Needs vs Software Perfection

[ranson]

"KDE developers should follow Apple's lead and focus more on the needs of users, instead of insisting on software perfection."

Now I think back to 1995, when IE focused on user needs over software perfection and the following of published specifications. And look what a mess of incompatibility we have today of javascript, css, java VMs, etc. Mainly because M$ focused on 'the needs of users.' No thanks, I'll stick to the specs.

Re:User Needs vs Software Perfection

[mccalli]

Now I think back to 1995, when IE focused on user needs over software perfection and the following of published specifications. And look what a mess of incompatibility we have today of javascript, css, java VMs, etc. Mainly because M$ focused on 'the needs of users.' No thanks, I'll stick to the specs.

Utter nonsense. I too can think back to those days, and no browser was following the specs. "Netscape is the next Microsoft" was a common complaint, as Netscape piled proprietary tag after proprietary tag into their browser. And don't even think about their initial CSS stab, the web still suffers from that today.

Cheers,
Ian

Re:User Needs vs Software Perfection

[nine-times]

What about the fact that this all started because Apple created a build of webcore that passed the Acid2 test? I don't think the accusation was that Apple wasn't adhering to standards, but simply that the code was "messy" by KDE standards.

Re:User Needs vs Software Perfection

[Johnny Mozzarella]

Perhaps you missed the story about Safari passing the Acid2 test?

Safari's code is capable of performing to publish specifications.
Microsoft's objective was to create their own specification.

Entirely different thinking.

Re:User Needs vs Software Perfection

[masklinn]

I don't know why anyone informed would say this. The CSS in IE6 is kinda bad, but it's clearly supposed to be W3C CSS and not something proprietary.

1- It's not "supposed to be the W3C CSS", of the few properties that are implemented (CSS1, CSS2 is barely scratched) many are wrongly implemented [box model, only fixed in Strict mode IE6] and half the implementation is heavily bug ridden
2- You probably missed all the proprietary MS crap in their implementation of the CSS, such as scollbar shit, that was NOT implemented in a W3C-compliant style (W3C allows proprietary CSS properties, but you HAVE to use precise prefix of type "-name-", which is why you see such things as -moz-outline)
3- CSS in IE6 is not "kinda bad", it's an awful heap of crap and a pain to work with from a dev's point of view
4- And if we extend from W3C's CSS to W3C's everything, MSIE sucks at W3C's HTML (heaps of missing tags, or not completely implemented ones), XHTML (which it doesn't understand at all in fact), W3C DOM/DOM Events and their binding in ECMA-262 (also known as ECMAScript/Javascript), XSL/XSLT...

Re:User Needs vs Software Perfection

[Anonymous Coward]

And don't even think about their initial CSS stab, the web still suffers from that today.

Practically everybody has dropped Netscape 4 support by now.

Their "initial CSS stab" was a rushed job. They had decided that JSSSL, a stylesheet language based on Javascript, was a better choice than CSS. The W3C decided to adopt CSS rather than JSSSL, and so they had to back-pedal and add support for CSS quickly.

Interestingly enough, one of the big criticisms of JSSSL is that it violated the "Principle of Least Power"; that it provided the full power of a procedural scripting language when only a declarative language was really needed. Of course, once Microsoft had killed Netscape, they added in a bunch of proprietary expression extensions to their implementation of CSS, neatly violating the principle themselves, after nobody could stop them.

But it wasn't the "initial stab" that was such a problem for the web. If Netscape 4 had have been followed up by a timely Netscape 5 instead of the Mozilla people going off and writing a whole fucking platform instead of a browser, then nobody would have really cared, because the Netscape 4 users would have moved on to Netscape 5 fairly quickly, which, presumably would have had decent CSS support.

It's the same problem that we are seeing with Internet Explorer today. It's not the fact that Internet Explorer 6's CSS support is so crappy that is the problem. It 's the fact that once Microsoft killed Netscape, they didn't have to bother with improving Internet Explorer, so they disbanded the development teams and discontinued the software. This means that there was no timely Internet Explorer 7 with improved CSS to upgrade to, and everybody using Internet Explorer is stuck with a version that is four years out of date.

This is one of the reasons why an abusive monopoly is so bad. The moment Firefox started to make serious inroads into the browser market, Microsoft paid attention and reformed the Internet Explorer development team. There will now be an Internet Explorer 7.0 beta for Windows XP at the end of the summer. But in the meantime, and until people upgrade, us web developers will be stuck with "the new Netscape 4" - Internet Explorer 6.

Re:User Needs vs Software Perfection

[Foolomon]

So fine. IE and Netscape were both guilty. That doesn't make the action correct.

I agree with the original poster's implication that sticking to the specifications is better than attending to the needs of users. The ISO and ANSI committees exist for a reason: when new features need to be added to an existing specification, the committees consider them and update the specification so that all implementers of the specification can do what's "best for the user" at the same time.

Granted, I'm living in a pipe dream but this is the ideal situation: where the focus on writing specification-constrained (note the qualifier) software is in its extensibility (read: ability to rapidly adapt to new changes to the specification) and performance.

(Note: this says nothing about portions of software or entire applications even that are not bound by the terms of an approved specification. For example, a POP3 email client has to follow the POP3 specification when it communicates with the mail server but it can add all sorts of features to the client that have nothing to do with the specification.)

Re:User Needs vs Software Perfection

[StormReaver]

"Software Perfection" isn't the point the KHTML developers were making. They aren't accusing Apple of violating a license (how the hell did that thread get started????).

They are saying that, despite all the media to the contrary, Apple's work is of no use to the KHTML developers. This is because Apple has been providing its changes in huge blobs without providing any clues what those changes are for or how they relate the the rest of the KHTML renderer.

Apple is following the license, which was never in doubt, but is being mean-spirited about it. The KDE devs just want people to stop the nonsensical meme that Apple is somehow helping KHTML development.

As I understand the situation, it is somewhat like an editor of a large book returning to the author only the errors in the book, but without any type of markup or explanation, in no particular order, and in a foreign language the author doesn't know. While the editor was (strictly speaking) doing his job (which is to find the errors in the book), he wasn't being very helpful.

The code could be sorted out, given enough time, but it's just as productive to ignore Apple's changes and do it from scratch.

This has nothing to do with focusing on the needs of the users, and it has nothing to do with software perfection. It is purely an issue of Apple being credited with helping KHTML, when it is doing nothing of the sort.

Oh holy shit

[ErichTheWebGuy]

Do we really need to start another flamewar between projects? Who benefits? Perhaps the KDE project and Firefox should *both* keep their collective mouths shut!

Heh...

[Otter]

[T]he KDE developers should follow Apple's lead and focus more on the needs of users, instead of insisting on software perfection.

I got on the KDE guys for their bit yesterday, so today I'll point out to the Mozilla side that the reason there was a decent browser for Linux in 1999 was that the Konqueror guys satisfied the needs of users while Mozilla went off constructing a whole new software platform...

Re:Blah...

[Skye16]

Why is this necessarily a fight? Why don't the Konquerer developers just say "you're ugly" and proceed to ignore the other guy? He can have his opinion, they can have theirs, and it's completely useless to argue about it. As a general rule, people don't like being told what to do, especially after they've made an informed decision.

Pissing contests

[goldspider]

Ya know, I can't help but wonder if it's silly little pissing contests like this that, at least in some way, prevents OSS from reaching its full potential.

Here we have several very adept programmers slapping at one another over how their respective web browsers work. Am I the only one out there that finds this kind of bickering trivial and unproductive?

Yes, people will have disagreements, and people will have different ways of doing things. Fine. But why not harness those different perspectives and create something better?

As long as OSS projects are afflicted by this kind of petty squabbling, developers' attention will be diverted from creating quality software. Now knock it off!

Re:Pissing contests

[Taladar]

I can't help but wonder if it's silly little pissing contests like this that, at least in some way, prevents OSS from reaching its full potential.

Did the thought ever occur to you that this is the Open Source Process? Discussing the best way to do something and then trying to prove one is right when words don't convince the other side is exacly the reason why quality in Open Source in so high. If you don't allow anyone to critize you your software will never be of optimal quality.

Konq vs FireFox vs world

[B5_geek]

I have always found Konq to be the best alternative to FireFox on sites that are "IE-only". (including my companies intranet.)

As a general web-browser I find Konq to be slow and kludgy, but it has never dissappointed me on the stubborn sites.

Anybody found similar situations?

Odd..

[0x461FAB0BD7D2]

Well maybe as a software engineer I should. But does anyone that isn't a software engineer care? Probably not. Case closed.
And guess what KHTML's team is? That's right. Full of software engineers. Which is why they care.

Secondly, developers should prioritise releasing their products on time, even if they "may have to cut corners".
Software developers in the open-source world make software because they love to. They want to make their project (note: not product) the best it can be. Releasing products on time is straight from the Marketing Department.

Goodger has every right to give an opinion, but no right to flame others for caring about their projects, much like Mozilla used to, before they gave up a large part of their community.

Love for a project, not releasing products in a timely fashion is what makes open-source different, and much appreciated.

Re:Blah...

[AllUsernamesAreGone]

And this is different from the normal how exactly?

Quite frankly, I'd rather have them arguing - when OSS developers disagree it often highlights issues that people should really be thinking about.

You might like the Solid Wall Of Unity approach but give me chaos any day.

No shit Einstein!

[10Ghz]

"Not everyone wants to change the world, but Apple does," he said, "and although they may have done the least required of them in accordance with the licences of the original source code, it was within their rights to do what they did, and no one should begrudge them for it."

Isn't that exactly what the KDE-developers said?? Sheesh!

I for one think that it's great that there are still people out there with a goal to create perfect code, and not just slap features together. It's interesting that Apple chose KHTML because the code was clean, fast and small. And now this guys suggests that KDE abandons those benefits and moves to Webcore (which has lost most of those benefits due to cutting corners and less than perfect code).

Is that it? Crummy code that is "good enough" is the way to go?

That just doesn't sound right

[rsax]

The KDE developers should follow Apple's lead and focus more on the needs of users, instead of insisting on software perfection.

I can't say I feel comfortable hearing that type of reasoning coming from a Lead Engineer of my favourite web browser. I'm not a Microsoft fan but if an IE developer made a comment like that then geeks would be cutting him or her up for that. I might be wrong since I am not a coder but wouldn't keeping software perfection a priority lead to less bugs in the future?

Some times a little conflict is good..

[Visceral Monkey]

Personally, I've always liked KHTML but have been frustrated by the lack of any real progress in it's use in Konqueror. Now, is this Apples fault? No, they just built a better mouse trap. This whole thing smacks of the same hurt feelings over the Debian vs. Ubuntu tift. The king is dead! Long live the king! and all that..

Also, if anyone has the "capital" to expend on criticizing KDE, it would and should be the people who have made one of the most successful browsers out there to put a dent in IE usage. See, people kind of listen to you when you are successful as opposed to when you sit and whine because your take on things just doesn't seem to be taking off (Debian/Konqueror I'm looking at you).

He has a point

[ShatteredDream]

A large part of the reason that Apple is still around with not even 5% of the market is that they do care about the user. With a user base that small for their platform, most vendors would be dead but Apple focuses heavily on the user experience. I don't see a lot of that at all coming from most open source projects.

Here's a little theory of mine: users are more concerned with having a great UI and having apps that work together than raw speed. Open source desktops used to have the speed advantage, but not anymore. Can anyone honestly say that GNOME is faster than Windows XP's desktop these days? Same for KDE and MacOS X.

For all of this bitching about Apple exploiting OSS, I don't see any recognition that the mere fact that OSX's underpinnings are OSS gives OSS a vote of confidence in the corporate world. For one of the two largest platforms in the world to switch to that foundation is a big endoresement and help lend legitimacy to OSS. The funniest part of this is that KDE's developers are finally discovering the fact that forks do happen. Imagine that, Apple actually forked KHTML for their own needs. Why is it OK for X.Org to fork and go off in one direction, but not OK for Apple to do the same thing? They give the patches back and excuse me if I am at a loss as to how a forked code base is going to maintain a lot of similarity with the original when both are going off in separate directions.

Re:He has a point

[cowscows]

You're talking about a double edged sword here, namely that the rapid advances in hardware has allowed software to be written much more sloppily, while still maintaining acceptable performance. It's good in some ways, because it allows things to get done faster and cheaper. You could also make an argument that some of the larger projects couldn't be realistically done at all beforehand. A project with dozens of programmers working on it becomes increasingly difficult to coordinate and perfect. Letting the specs of the hardware smooth over some of the bumps makes life easier.

Then the bad side is that sloppy coding is not only inferior performance-wise, it also leads to maintenance difficulties, as well as security issues. The most notable example being all of the legacy garbage that windows still carries around.

It sounds like the KHTML people are trying to buck the trend, and make a large, but solid piece of software. They're saying that it's not impossible, just that it takes a while. The "computer industry" has been moving at this incredible speed for a while, so fast in fact that it wasn't realizing a lot of the mistakes it was making. There are plenty of examples of how this is making life tougher now. The KHTML guys aren't interested in doing that anymore, they want to do something right, so they're doing it.

Maybe they're thinking of their renderer as more a piece of infrastructure or technology more than an end product for your everyday user. Try to draw a vague parallel to some guy writing code for the space shuttle. There's more at stake when you're sending humans up in a rocket, but the mentality can be the same. We want to get this right, on the first try. It's inherently complicated enough , no need to make things any denser with hastily added features and sloppy coding.

Re:He has a point

[As Seen On TV]

As much as I like to hear talk like that, a better explanation is that the market is freakin' huge, and 5% of the installed base is still anywhere between 15 and 40 million units, depending on which estimate you believe.

Everybody likes to talk about how Apple owns such a small share of the market, but in doing so y'all lose sign of the fact that Apple is the fourth-largest computer manufacturer in the entire world, and the second-largest developer of operating-system software. Considering how narrow our focus is, I'd say those are two pretty remarkable facts, wouldn't you?

Re:KDE should be grateful.

[Anonymous Coward]

Apple went with KHTML instead of Mozilla. Instead of gratitude, the KDE devs are angry that Apple isn't tailoring their patches for them?

First, KDE devs are grateful. Read one of the many linked blog entries about how Safari has done many good things for the project, if you don't believe me.

However, they are angry at something: people like you. Coming here on /. and making a completely backwards post that misrepresents everything they stand for. Sit down and STFU.

MoFo getting more like MS by the day, it seems

[McDutchie]

From TFA:

Goodger went on to say the open source community could not accuse Apple of breaching any licences.

I would not be so sure of that. I seem to recall that the GPL defines source code as the "preferred form" of the program for making modifications of it. If Apple "comments" its patches by referring to numbers in a proprietary bug database to which only they have access, Apple could be accused of intentionally obfuscating its source code, which is a violation of the "preferred form" clause in the GPL. In any case, it's ethically wrong because the free-software concept is meaningless if the provided source code is not realistically usable without having access to essential information about what it does.

It was important, he said, realise that "no software is ever perfect".

Secondly, developers should prioritise releasing their products on time, even if they "may have to cut corners".

Gee, that sounds eerily familiar. Where have I heard it before, that "give Joe Sixpack what he wants and damn software quality" attitude? Marketing fluff at the expense of solidity and security? Oh right, of course, that's the attitude that brought us the virus propagation engine that is Microsoft Internet Explorer. Is it any wonder that Firefox is now on its way along the same route?

"Most developers probably don't alienate people intentionally ... Over time, software has come to demand an impossibly high level of computer literacy," the Firefox creator wrote.

Ridiculous. The use of software is demanding less computer literacy by the year -- compare today to the MS-DOS days of twenty years back. But that is in fact a big part of the problem. People should learn to accept that using a computer requires some basic form of clue. If people are not willing to acquire such clue, they should watch TV instead so that they won't harm anybody with the viruses, spam and DDoS attacks perpetrated through their zombified computers.

Re:KDE should be grateful.

[10Ghz]

Apple went with KHTML instead of Mozilla. Instead of gratitude, the KDE devs are angry that Apple isn't tailoring their patches for them?

KDE-guys did not complain about Apple as such. They even specificly mentioned that Apple is abiding by the license. what they complained about were the USERS who whined when KHTML took time to incorporate improvements made in WebCore!

Do you "get it" now, or do I have to hit you with a clue-by-four?

Re:Blah...

[squiggleslash]

Even worse is that Ben doesn't even appear to know what he's criticising. He takes a quote out of context and puts the same spin on it that /. did a few weeks ago, treating it as a criticism of Apple when the thrust of the original piece was protesting that people were assuming that just because Apple had added something to WebKit, it follows that it'd be in the next release of KHTML, and were getting pissed at the KHMTL people when that didn't happen.

I'm not 100% surprised, given the degree to which the original post was misrepresented, but given some replies to his blog entry pointed this out and Ben's single response to them has been dismissive, it'd be nice to see a sign of good faith.

Re:Blah...

[hostyle]

How would you like it if you had a real nice and clean well documented codebase and you gave it to someone for free, the only stipulation - if you make some changes please give them back to us also. The guys you give your code to do make changes and do give them back. Problem is the code they give back is all over the place and badly (if even) commented. Then other people (your users) start complaining "this other guys software is better than yours, but hes using your code. Give us those features NOW." ?

Re:Blah...

[JohnFluxx]

Hmm, we work hard with Apple to give them the best possible access to our code. Apple does the minimum it can in giving the code back to us. Slashdotters praise Apple for the work on html, and so we just ask for people not to praise apple so much since they aren't exactly working with us - they don't use any of the resources we set up to try to encourage them to work with us.

And now _we_ are the pain to work with and aren't encouraging participation??

Re:Boy are you dumb

[BladeMelbourne]

Safari only passed the Acid 2 Test because the developer David Hyatt spent time over two weeks to make it pass.

I'm not saying that this is a bad thing, in fact it's an excellent thing. But the fact is Safari, Mozilla and MSIE all failed the Acid 2 test when it was released. Using MSIE I see red. lol.

Now Safari passes. And no doubt each would fail several more tough tests. No one test can prove a superior rendering engine, unless it was 10 MB big and tested every [X]HTML/CSS1,2,3/JavaScript specification in various scenarios.

I'm looking foward to getting a Mac Mini and seing how good Safari is. It will also allow me to develop web pages against Safari for the first time.

Safari and KHTML

[danalien]

Safari and KHTML
Submitted by carewolf on Fri, 05/13/2005 - 10:33.

• Notice how there isn?t a vs in the title?

  Hyatt and Maciej joined us on IRC yesterday, and we had some really good discussions. I might as well also admit that Maciejs comment was true (but out of context). Please notice that that implies we are discussing solutions and a common future. The idea of a common source tree is pretty much abandoned as we have very different goals and requirements, but we are discussing improved cooperation. With Apple just having released Tiger and us preparing for KDE4 we have a unique opportunity for bringing our source trees closer again.

  Since Apple is being a nice guy for the time being, I will let them announce how things will improve once we have a solution, but please, no more ?vs.? stories for the time being, we are working on solving it.

Safari and KHTML again
Submitted by carewolf on Sat, 04/30/2005 - 13:22.

• I just wish to weigh in on debacle to clear up some mistakes. First of all I would like to say I agree with Zack. The annoying part is not that Apple don?t cooperate as much as they could. They are actually helpfull in answering questions and _tries_ at least to separate OS X specific features in the code (allthough they fail miserably at it). No, our problem are users who think Apple does more and underestimate the effort it takes for us to implement patches from WebCore. We are doing this for free and for fun, all we really want is appreciation for our effort.

Emphasis added by me...

Re:Boy are you dumb

[nine-times]

Safari only passed the Acid 2 Test because the developer David Hyatt spent time over two weeks to make it pass.

Wasn't that the purpose of the Acid 2 test? To give an example of common rendering problems so that browser developers could see what their browser was doing wrong? Now, I'm not saying that passing the Acid2 test means the rendering is perfect, but the challenge was placed out there, and the Safari developers took it up. It's exactly what Mozilla and MS and KDE should do, too.

Now, if after all that, we can come up with a new series of serious rendering errors not addressed by the Acid2 test, then let's make an Acid3 test, or whatever. But I don't see the grounds for complaint. It's like saying, "well the only reason Firefox renders HTML properly is because the Mozilla team spent time to make it render HTML properly." Well, good. That's what they should be doing.

Re:Blah...

[10Ghz]

The KHTML guys are really shooting themselves in the foot with this. They certainly aren't encouraging companies to participate with open source projects. The only thing they're doing is reinforcing an existing conception about open source developers -- that they're a pain to work with.

The KDE-developers commented about the USERS who whine when Safari-patches don't get merged in to KHTML. They never whined about Apple as such. They even mentioned that Apple is abiding with the license.

How exactly are they "pain to work with"? Apple got a kick-ass HTML-code from them, with NO questions asked, no price being asked and with zero red tape! How exactly does that mean they are "pain to work with"? If anything, this incident shows that COMPANIES are "pain to work with". KDE-developers REALLY wanted to work with Apple, but Apple wasn't interested!

Where "free" software fails

[putaro]

Ben Goodger has hit on one of the major ways that "free" software can fail and that is that the people working on the project are doing so out of the goodness of their hearts and for their own reasons. Some developers, like Goodger probably, are writing free software for the kick of having as many people use it as possible. This will make them somewhat use oriented. Others, and the KHTML guys appear to be this, are writing code for the sheer joy of writing code. And it's not fun to write stuff that cuts corners just so you can get it out the door. Of course, you may not be meeting the users' needs. But then, there's no requirement to meet users' needs. It's free - if you don't like it, fix it yourself or don't use it. In this case, Apple chose to fix it themselves. The fact that they diverged from KHTML simply shows that they have different priorities and isn't any different than FreeBSD and NetBSD spliiting.

Re:Blah...

[ceeam]

It's actually a known phenomenon that "real nice and clean and well-documented codebase" can in fact be _evil_! Because everyone except really lousy coders are afraid to touch it. "It's so beautiful it's practically dead" one could say.

Re:Blah...

[mmeister]

There is no such thing as real nice clean, well documented codebase, at least not forever.

These attributes naturally go away as you add functionality to any code. That is a fact of software development.

Don't Follow Firefox Adivce: considered Bad!

[kupci]

What the Firefox guy is forgetting is that he wouldn't be where he is had he followed his own advice. Has everyone forgotten the beating Mozilla took when they scrapped the Netscape code and decided to rewrite the layout engine from scratch? While they strived for making their browser superior, IE blew past them in the marketplace, building on the existing (Spyglass?) foundation which was "good enough". Now finally the wisdom of that choice has come to fruition.

The original Mozilla browser, first released as Navigator 1.0, was developed rapidly by a small team that was passionate about creating the next killer app -- and they succeeded wildly. Now that the web has evolved, Netscape has assembed the finest team available to redesign and redevelop the next generation layout engine upon which it will build future products. Gecko enables a pioneering new class of dynamic content that is more interactive and offers greater presentation control to Web developers, using open and recommended Internet standards instead of proprietary APIs.

Re:Mutually Exclusive?

[cowscows]

They're mutually exclusive if you want to work at the pace that the computer industry tends to move at. Doubly so for a bunch of volunteers working for free.

I guess that makes the assumption that the needs of the users includes a rapidly expanding feature set and whatnot. And while that is important (particularly if you're going for marketshare), there are still users who'd rather have some good code. Not to mention that eventually the bad code may catch up to you, and cause the needs of the users to change. Windows needed a lot of usability enhancements until the Win95-98 era. Then stability became a big issue. MS ironed a lot of those problems out, and now security and spyware is the big problem. A lot of those issues could have been mitigated by better code at earlier stages. Fortunately for MS, their monopoly has allowed them to advertise their security and spyware solutions as new features, and so a mostly under-informed public still thinks they're paying for innovative work.

But returning to the original point, even for a big, well funded company like MS or Apple, it's not really possible to write perfect software fast enough to lead the market in features. You can dump more money into it, and hire more engineers, but that just makes it all the more complicated and harder to coordinate, leading to more mistakes.

The KHTML team can avoid that because they're not trying to keep a business profitable, they're writing this stuff because it's a hobby for them. Personally, I try and keep my hobbies as free of deadlines as is possible. And if anyone wants to criticize how I indulge in my hobbies in any sort of non-constructive way, they can go to hell, I'm not interested.

Re: Do have to admit though........

[SirTalon42]

KDE has a page (http://www.kde.org/info/requirements/3.4.php) listing all of their requirements, I'll provide you all the things marked as 'required'

QT
X Server
Perl
BZip2
ZLib
PCRE
LibPNG

There are other things marked as recommended (such as OpenGL and OGG Vorbis), and there are others marked as optional (such as LAME).

The problem isn't KDE is bloated, its the way the distros package it (huge monolithic packages that contain a load of different programs), though some distros like Gentoo now provide 1 package per app (which allows you to trim most packages off.

Also comparing KDE to XFCE makes no sense, XFCE is an extremely minimalistic desktop environment (its just a bit more than only a Window Manager). Only comparing KDE to GNOME would make any sense since both are complete desktop environments.

Tests are no substitute for good design

[expro]

I am sure to be modded OT as this whole thread is, but...

The Agile/XP movement is warped at best. Tests are no substitute for good design and they cannot prove any useful level conformance to a design (except in an extremely trivial application). Tests are useful in many cases, unless they are used to rationalize bad practices based on false notions.

And the more extremists you have trying to force it to be so, the worse the XP/Agile movement is percieved. Sure, they picked up on parts of a number of good practices that good programmers already followed, but when will they stop twisting them and advocating that experienced programmers abandon principles of adequate forward-looking design and methodology and follow the way which is what they ultimately believe to be The Only Right Extreme Way.

They resemble the pointy-haired managers who would like to think they can substitute their process for masterful programming and design.

I was attracted to XP by their advocacy of some of the more-reasonable principles until the fanatics showed why it was really called extreme programming. They need apologists to start really apologizing.

Re:??? HUH ???

[saider]

Both of those approaches have points of value, but they are both extremes. The Agile way seems to be more appropriate for contracts for time, whereas the "Fragile" approach is more applicable to fixed price contracts.

Agile works as long as the customer is willing to pay for the changes. Agile is good because the customer sees progress. Agile is bad because an indecisive customer can flip-flop on features and cause significant headaches.

Fragile works better where the customer needs a specific solution. They list their demands and your company realizes those demands for a price. Changes are discouraged, but that should be fine as long as the customer knows exactly what they want. The company's process ensures that what is contracted for is met. This is not good for research type projects, where the best solution is not known and needs some experimentation.

Note: One would have trouble applying the Agile paradigm to any kind of regulatory environment. Telecom, medical, and military/government contracts pretty much mandate the use of the "Fragile" system.

Pursuing Software Perfection is a GOOD THING...

[rben]

But a big part of that is supplying what the users need. I think all three sides of this discussion could learn a thing or two by listening to the others.

You don't get involved in an open source project to write crappy code. You get involved in order to fix a problem that bugs you, show off your coding skills, or do a little good for the community. One of the benefits of coding for open source is that you really can take the time to get it right.

Businesses often fail to pursue excellence in coding because they believe that by taking shortcuts they save money. That's almost always wrong. One of the reasons that Netscape got beat by IE was (and I know I'll get beaten for saying this) that IE was written in a modular way that allowed it to be used more flexibly than Netscape. The IE code was better planned and executed. The developers who joined the Mozilla project took the original Netscape code and hammered on it for a long time to produce the successful browser that we now have. Even so, it was bloated and in need of a lot of trimming. So the Firefox project fixed those problems and now Netscape is based on the Firefox core rather than the original Mozilla core. (Sharing is a good thing.)

Apparently Apple believes that by taking short cuts they save money because the FOSS community will come in behind their engineers and clean up the code for them the way they did for Netscape. Why shouldn't Apple take advantage of the same mechanism?

What happens too often in corporations and is apparently happening in this part of Apple, is that they have forgotten that they are dealing with people. The FOSS crowd seems to have more than it's fair share of idealists, and they dont' like being taken advantage of. Hopefully Apple has figured this out by now and is working on a plan to mend fences. Otherwise it might be very hard for them to get help from the community in the future. It would be a shame for OS/X to fail because of foolish management mistakes on the Safari side.

Open Source can learn a lot from what Appled did, though, even if we don't like how it turned out. Appled focused on fixing things that were causing problems for their customers. They also focused on becoming standards compliant. A lot of FOSS projects come up short in that area. What gets attention is whatever is cool to code or bugging a particular developer. Not enough of the FOSS projects have any real central focus.

Listen, learn, and move on. The best thing that could come out of this whole mess is a good discussion on how FOSS and regular software companies can work together to mutual benefit. Perhaps we need some kind of template agreement that makes responsibilities clear so that the companies involved don't make bad assumptions like the ones Apple seems to have made.

Apple offered, but KHTML didn't want to.

[Paradox]

Apple is on record for offering to jointly attempt to make the important parts of WebCore cross-platform, similar to the situation with Gecko.

The KHTML team turned them down. They probably did so because it would shift the focus away from the KHTML they know and love and more towards the more realistic (but messier) WebCore, which they don't seem to want to do.

The KHTML team doesn't even seem to want many of the changes. Apple makes a product, and they don't care if they break small things to make deadlines. KHTML is a product of the opposite school, preferring to make a very small, clean codebase. The price of this is feature deficit.

This isn't about Apple being evil, or KHTML being snobs. It's about a project being forked. As time goes on, Apple has less and less to offer to KHTML. WebCore and KHTML are diverging, and people seem to be upset about this. I can't imagine why, this sort of separation was inevitable. Apple's best interests are served by leveraging their own excellent environment, and every time they do, they further exclude the KDE project.

Mr. Sarcasm says...

[craXORjack]

'the KDE developers should follow Apple's lead and focus more on the needs of users, instead of insisting on software perfection.

But if we hadn't waited on software perfection, we wouldn't all be playing Duke Nukem Forever on top of the GNU Hurd.

Re:Blah...

[Jimithing DMB]

how has KDE/KHTML been hurt by Apple's actions?

Look around you

Bullshit. The KHTML developers brought this discussion on themselves. Apple rightfully took the code, incorporated it into their products, and not only made the modified version available to users of their products but also publicly for anyone wishing to download it.

It appears that the KHTML developers expect Apple to do all the integration work for them. There is nothing in the LGPL that requires this and my reading of both the LGPL and the GPL indicates that requiring modified code to be made available is the spirit of the license. Integrating it back to the mainline is a courtesy but the license specifically does NOT require this.

Sure, it's customary to help integrate it back into the mainline tree but there have been other instances where this hasn't happened. For example, read up on Lucid Emacs (XEmacs) vs. Emacs debacle. You can draw many parallels from that to this. In that case Stallman was working on releasing a new version of Emacs for years and hadn't done it. Meanwhile, Lucid needed certain features and implemented them. When they offered to integrate the work back to mainline Stallman rejected it because he had his own ideas of how it should be written. In addition (and this does not apply in this case) Stallman required copyright assignment to the FSF which was something Jamie Zawinski in particular did not agree with. After much back and forth Lucid gave up and thus the XEmacs fork was born.

The Lucid Emacs developers suggested that their code simply be incorporated into the next Emacs and that if Stallman wanted to rewrite it again that's fine but what they had was already working and better than what was in the tree. Stallman rejected it because he preferred to wait until their rewrite was complete at which point Lucid could try to integrate with the rewritten code.

It should be obvious that this is a *really* stupid decision. The KHTML developers should suck it up and realize that Apple now has a better rendering engine than they do. They should merge in the changes now (including the ones they don't like) and THEN decide to rewrite things if the code is problematic. In the meantime the KHTML users have a better browser. It will take just as long to write code regardless of whether they merge in the Apple changes or not.

This basically amounts to the KHTML developers having a serious case of Not Invented Here syndrome. After trying and failing to get Apple to do the merging work for them they cried foul and posted publicly about how Apple wasn't helping. I'm truly glad it has mostly backfired because it's up to the KHTML project to decide whether they want the code or not. It's not Apple's responsibility to take orders from the KHTML developers. If KHTML doesn't like it then WebCore can remain a fork of KHTML just like Lucid Emacs is a fork of Emacs.

And don't give me any of the shit floating around here about how the KHTML developers merely wanted to point out that Apple wasn't doing the merging work. There are claims in this thread that the KHTML developers are fine with that but they just aren't fine with it looking as if Apple is contributing to KHTML. Well, news flash, Apple *is* contributing to KHTML and if the KHTML developers don't like the way they contribute and didn't want a big media fuss about it then they should have been smart enough not to write about it publicly. It is for this very reason that I *don't* keep a journal on the web.