Safari vs. KHTML

Johnny Mnemonic writes "CNET has a story that describes the divergence between the code base of Safari and KHTML. Although there were high hopes that Apple would contribute significantly to the OSS project, that optimism has all but disappeared. Is an unrealized danger of OSS that others may take your project in a direction you didn't intend? Can OSS code and goals harmonize with the goals and needs of corporation designed code? Is it that Apple mismanaged the relationship, or that the KHTML guys expected too much? Interesting warning for other OSS-corporate marriages." We've previously reported on the frustration in the OSS community on this issue.

Comments

Its only the bad things we head about?

[Folmer]

Afaik the relationship between apple and freebsd is fine, and they use eachothers' patches etc. The problem seems to be that apple wanted to develop the browser in another direction than kde, and the communication stopped as they didnt use eachothers patches. As apple are having paid developers working on it, they should develop it their way and kde should maybe look at their methods to see if they are able to work in that way. If not, though luck.. I cant see that apple is the bad guy here.

Re:Its only the bad things we head about?

[Metzli]

That's a problem for KDE, but does Apple not have the right to do what they want with their patches? They wrote the code and they're releasing it to be viewed and used, so shouldn't that be lauded rather than complaining that they're not releasing things the way that the other guys want?

If Apple complained that the KDE guys were releasing code in a manner than didn't work for Apple, then people in the OSS community would say that the Big Bad Corporation (tm) is trying to control OSS and tell the developers what to do. Does Apple not get the same consideration? I thought the point of open sourcing code was to allow people to do what they want with it. Apple is, so either take what they're giving (for free) or shut up and write your own patches. It seems simple enough to me.

Re:Its only the bad things we head about?

[Anonymous Coward]

From the parent: "the problem with Apple and KDE is that apple doesn't make patches that are easy to apply to the khtml source

Noone seems to bitch about X.org changes not getting back into XFree86. Forks are not a bad thing. If Apple can move the software faster than the khtml guys did, they have my blessings. I wouldn't want to slow down Apple's progress by making them jump through hoops for the KHTML guys any more than I'd want to slow down X.org.

Is an unrealized danger of OSS that others may take your project in a direction you didn't intend?

For almost everyone in the world it's a "fully realized feature" not "unrealized danger" of open source that if a new team can advance the software faster than an old team, they're FREE to do so because the software is OPEN.

Re:Its only the bad things we head about?

[angst7]

As others will likely point out, they certainly are within their legal rights to do what they like where patches are concerned. But you're missing a crucial distinction between Must and Ought.

It is not required that Apple play nice with the way they release patches. That is to say there is not 'must' apart from the requirement that they make them available. But there's alot of 'ought' that comes into play when you use OSS code. This is basically a niceness test that says, in effect, if you use this code to make money, great, but you 'ought' to give back in such a way that we can make use of as well.

Having said all that I feel a bit bad about even responding to an obvious troll. There's very little 'insight' in your comment.

Re:Its only the bad things we head about?

[MoneyT]

Yes, because running it through the obfuscator is not how you work on code.

But the source files are just that.

If some 3 man team in the middle of nowhere started working on KHTML and just realeased the changed documents back to the public, would there be such a great outcry over the fact that they aren't providing the bug trackers?

Re:Its only the bad things we head about?

[Anonymous Coward]

Let me get this straight...

KDE gives Apple (and everyone else) access to a monolithic block of code that doesn't run on Mac OS X, and that's considered a big favour. Apple gives KDE (and everyone else) access to a monolithic block of code that doesn't run on Linux, and that is the moral equivalent of spitting in their face?

Error, error, does not compute.

Re:Its only the bad things we head about?

[turbidostato]

"Let me get this straight..."

Straigth but wrong.

"KDE gives Apple (and everyone else) access to a monolithic block of code that doesn't run on Mac OS X"

The question is, it is NOT a monolithic block of code: everybody has access to KDE source code repositories and they can be analized checkin by checkin.

"Apple gives KDE (and everyone else) access to a monolithic block of code that doesn't run on Linux, and that is the moral equivalent of spitting in their face?"

Yes.

It is not because it doesn't run in Linux (the problem is not Linux, either, but XWindow). In fact, if Apple would develop its safari port in a way compatible with XWindow it should have been seen as a GREAT FAVOUR, but it wouldn't been asked not taken as an offense if the code itself was exactly what it is now.

The offense comes from Apple not giving (read)access to their source code repositories so you only gain access to a bunch of incompatible inextricable code while it would cost NOTHING to them to allow read access to the repo so that very same code could be analyzed checkin by checkin thus making possible to segregate incompatibilities from useful code in understandable bits.

No one sane would say Apple is not doing what is legally bound to do, nor anyone sane would say Apple's behaviour to be unexpectable (so this news is kind of flamebait), but yes, they could do things much better without them costing nothing and they decided to go for the most difficult and unuseful path for the KDE team, so KDE guys are reasonably feeling insulted.

Re:Its only the bad things we head about?

[alienw]

Read The Farking License. It doesn't say "preferred form for the original developer". It says "preferred form of the work for making modifications to it". As in, ASCII source code. What they are saying is that you can't comply with the GPL by, say, releasing code only as a PDF file, or only as a hardcopy, or without the appropriate makefiles. Nowhere does it say you even _have_ to release a patch.

Re:Its only the bad things we head about?

[Anonymous Coward]

The thing is, the KDE guys did Apple a favor, and DID make it easy for them to get at the code.

This is the second time I've seen you post this bullshit. Quit it. It isn't true. The first the KHTML developers heard about it was when the first Safari beta was released, so they couldn't possibly have done anythingto help Apple out. Everybody was expecting a browser based on Gecko because of their hiring decisions.

If you still don't believe me, read the email yourself:

I'm the engineering manager of Safari, Apple Computer's new web browser built upon KHTML and KJS. I'm sending you this email to thank you for making such a great open source project and introduce myself and my development team. I also wish to explain why and how we've used your excellent technology.

Does that sound like the KHTML developers gave Apple any special help?

Again, people, stop modding this up, he's a clueless idiot spreading lies.

Re:Its only the bad things we head about?

[99BottlesOfBeerInMyF]

The thing is, the KDE guys did Apple a favor, and DID make it easy for them to get at the code. Apple just spit on their courtesy by just releasing their monolithic patches.

I'm not sure, what exactly did the KDE guys do to help Apple? Did they help them to incorporate the code into their Webcore fork of Konquerer?

Apple using Konquerer has already helped the KDE team in a number of ways. First they do have those patches to look through, in fact the acid test CSS patches from a previous article were even separately documented for the Konquerer codebase as much as possible given the divergence of the codebases. Next Apple using the Konquerer engine has made a lot of web sites compatible with Konquerer since Web designers are much more likely to test their pages against Safari than Konquerer. Also, they have advanced standardized HTML in general by promoting a browser that does not conform to either Gecko or IE's bugs and quirks.

I think it is unlikely that Apple is going to change versioning systems to make the KDE team's job easier, nor are they likely to implement changes on both browsers (even if they could which they can't since the Konquerer developers do not want to implement all of the same type of changes Apple has and don't use the same development tools or APIs). What they do, however, is provide their changes openly so that the KDE team can look through them and copy whatever fixes, improvements, or changes are useful to them. I'm sure both sets of developers are overworked and neither has enough time and both would like more people to do more for them. Maybe if the KDE developers asked for more granularity with the patches the Safari team would be willing to accommodate them. They would probably also be happy to answer questions and explain particular changes.

I guess I just don't see what anyone is mad about, and I'm not sure that anyone really is mad that is involved with Konquerer. This seems like a lot of people trying to make drama out of very little.

Re:Its only the bad things we head about?

[drakaan]

I'd imagine he (or she) was saying something along the lines of "after Apple rummaged around in KDE's browser code and started building something, they pretty much ignored repeated requests from the original developers to make things easier on them when dealing with patches".

It's not wrong per-se, but it's definitely rude.

It's easy to pick out one comment from a post that has some feeling in it and go sociology-101 on that poster. It's not much harder to try and see what the whole post is trying to get across.

Re:Its only the bad things we head about?

[Thomas Miconi]

That's a problem for KDE, but does Apple not have the right to do what they want with their patches

Of course they bloody do. That's called a fork ! And freedom to fork is the most important aspect of OSS - in fact enforcing and maintaining this freedom to fork is the central aim of the GPL.

Apple quite simply forked Safari. This happens all the time in the OSS world. Hello, does anyone really expect that X.org patches will remain 100% compatible with the XFree86 code structure ad aeternam ?

Could someone please tell me what exactly the problem is in the Apple-Safari case ?

Thomas-

Re:Its only the bad things we head about?

[MudflapSoftware]

The thing is, the KDE guys did Apple a favor, and DID make it easy for them to get at the code

Ummm... no. The GPL gave apple the access to the code. Apple has been compliant with the GPL by giving the changes back to the community.

The fact that the changes are not in the prefered format is completely irrelevant.

AFAIK, the GPL doesn't dictate what format changes need to be submitted back to the project in. In fact, that would hypocritical, as mandating a specific patch format would be a limitation of our freedoms to use what software we want.

Apple is checkpointing their releases. They start with a block of code, make a buttload of changes and then run diff across it.

Big deal, stop whining.

Re:Its only the bad things we head about?

[RodgerDodger]

the changes Apple's made involve porting what started out as KHTML to a completely different API, and no matter how frequently Apple released the patches, or in what sized chunks they released the patches, they'd still be full of changes related to the fact that Apple's APIs have evolved along a completely separate path from X11 for their entire lives.

Yes. That's KHTML's problem. They want changes being written for another platform, but they don't have an application or code-base structure that makes it easy to seperate platform-specific code. Nor have they taken the time or effort to introduce one, _after_ they saw the nature of the Apple patches.

This isn't a hard technical problem: it's a political and organisational problem, for the KHTML project.

IMHO, part of the problem here is that KHTML wasn't designed to be extended in piecemeal fashion. Look at Eclipse: totally plug-in based, and if one port goes in an undesired (by the general community) direction, you simply swap out the affected plugin. Very manageable.

OS Divergence

[Veinor]

Open Source is designed so that everyone can see the code. If you can see the code, then you should be able to tweak it and make your own version of it, as long as you still give credit where it's due. Indeed, look at all the variations on *nix/Linux:

• Suse
• RedHat
• *BSD
• Knoppix
• Mandrake

And there are definitely more that I haven't included. If Safari diverges form KHTML, it's fine with me.

Um

[IamTheRealMike]

Is an unrealized danger of OSS that others may take your project in a direction you didn't intend?

It's not unrealized, lots of projects have forked before. I think anybody who puts their code under a license that allows forking will realize that it can happen.

Can OSS code and goals harmonize with the goals and needs of corporation designed code?

Of course it can, this happens every day. Look at the kernel, GCC, Wine, etc.

Is it that Apple mismanaged the relationship, or that the KHTML guys expected too much?

I don't think expecting documented patches or a shared bug tracker is asking too much - this is the pretty much the minimal level of co-operation most projects would expect from a corporate good citizen. Some companies go even further than that, and hire some of the core developers, sponsor conferences, provide hosting facilities etc. There are plenty of examples in the Linux community of companies doing that.

So did Apple mismanage the relationship? Arguably there is no relationship. They certainly mismanaged expectations - if they'd come straight out and the beginning and said "we're not going to co-operate" a lot of frustration would have been avoided. That would have harmed their (mostly imaginary) pro-open source image though.

I doubt there's some kind of Evil Plan to screw over KDE here, it's more likely that Apple don't care or want to help the open source community, it's just a convenient place to take code from (go see how much FreeBSD has got back from them, for instance). Open source and Linux specifically are primary competitors and they'd be foolish to help the community more than they have to. After all, they're in the business of selling proprietary operating systems.

Re:Um

[danigiri]

"After all, they're in the business of selling proprietary operating systems."

No, really. They don't. They're in the business of selling computers and peripherals.

Having those computers and peripherals work well (or at least up to their expectations) incidentally needs of propietary operating systems.

Dani++

PS: look on the changelog of Bash, recently there has been some significant Apple contributions, reported on /., even.

Re:Um

[bahamat]

If they are going to feel betrayed when the project forks they shouldn't put it under a free license anyway.

Apple publishes changes they make to khtml. They have to because khtml is [L]GPLed. If anybody bothered to check, WebCore is licensed under the LGPL 2.1. There's absolutely nothing preventing KDE (or any other Linux destkop for that matter) from incorporating WebCore into the system.

What if someone wrote a new VM subsystem or scheduler for the Linux kernel, and then published patches and the new vm/scheduler? Would they still be villians? Even if they sold it commercially? Even if Linus didn't use it?

All you KDE developers, quitcherbitchen. Why GPL your code and then get pissed when someone uses or forks it? Don't snivle that Apple's version of khtml links against WebCore. Use WebCore. If you don't know how or don't have time to learn it, that's not Apple's problem, is it?

Re:Another question

[Spy der Mann]

As long as they're abiding by the terms of the license, does Apple, any corporation, or any entity for that matter, have any obligation to contribute anything back to the project?

If you don't like that people JUST obey the license, then change the license!

i.e. If a company decides to launch a similar product based on this source code, they're obligued to keep a revision history in a previously agreed format (i.e. CVS, SVN, etc) so that the authors can track down their improvements.

Ta-da!

Open Source Is A License

[American AC in Paris]

Folks, if you want your code used in a certain way, you put the terms in the damned license. That's what it's there for.

Besides, last I checked, the KHTML folks don't have a beef with Apple. They do have a beef with the fanbois who can't seem to grasp the fact that Apple using KHTML's Open Source code does not immediately mean that they're best buddies.

All it means is that Apple is using Open Source code. Period. Apple isn't violating anybody's trust.

This is just stupid

[Otter]

The original story just had Zach Rusin saying that Apple's contributions to KHTML were exaggerated. OK, fine.

This stuff is just stupid. Apple has done absolutely nothing illegal; arguably they've done nothing inappropriate. KDE and KHTML are not in any way any less well-off, and if this story accurately reflects the attitude of the primary KHTML developers, honestly, they're being jackasses.

What all this demonstrates is why using free code (especially GPL/LGPL code) is much more of a minefield than a reading of the license would suggest. You can comply to every last detail, and it doesn't do you any good against the negative publicity when someone decides you "owe something to the community".

Nothing to see here... It's just a fork

[Jarnis]

Apple has followed the obligation of the license.

It's just a fork. Forks happen. Move along. If KDE guys think KHTML sucks compared to WebCore/Safari, they are free to fork THAT and start from there (backporting it to KDE). The source is open. Whine less, code more :)

Childish Spat

[cyberlotnet]

This is nothing but a childish spat between 2 diffrent groups of developers.

Apple published the patches, and changes and KHTML cries about them having to much OSX specific code in them? Thats just crap..

Apple is acting in good faith, they are basically asking Apple to make sure all patches are 100% compatible with the current code base.

The KHTML team might as well just ask Apple to take over the project in full.

Open Source does not mean "Anything you do must conform and work with our project or your not doing it right"

Open Source is "If you make changes please give back to the community with the understanding that your changes might not be compatible with ours, Your code changes may not be what we want, but we can't complain about that"

Re:Childish Spat

[TravisWatkins]

They aren't complaining about Apple at all. They're complaining about Apple fanboys who say Apple is helping KHTML and about KDE users who think the KHTML devs are lazy for not merging the patches from Apple right away.

KHTML Developers Assumed Too Much

[mgbaron]

I think Apple made a decision that it needed to switch cores and at that moment has every right to do so and never look back. The fact that they are putting any effort into KHTML at all should be looked at as a mere bonus for the KTHML developers at this point. Apple never claimed to be the white night funding the KHTML project or that they would be the dominant developer for the future. This is not an example of IBM taking over a project. I think some KHTML guys read way too much into this relationship. It was pretty clear from the start that they were being used (but the nature of their license allows for this). It was great that they showed trust and attempted to built a relationship, but they should not have become in anyway dependent. I'm not saying this is the case, but the bitterness of their response seems to suggst this sort of dependence.

Sound like KHTML team doesn't want to play either

[MoneyT]

"One thing you may want to consider eventually is back-porting (WebCore) to work on top of (KDE), and merging your changes into that," Apple engineer Maciej Stachowiak wrote in an e-mail dated May 5. "I think the Apple trees have seen a lot more change since the two trees diverged, although both have useful changes. We'd be open to making our tree multi-platform."

The suggestion, which KHTML developers said they were unlikely to accept,


So Apple is open to making the tree cross platform, and presumeably to them back porting web core (which is nessesary to implement some of the things Apple has done since) and KHTML doesn't want to.

So by choice KHTML has already limited the changes they can use.

"In open source, everything's supposed to be done the right way, but sometimes the less correct way is faster," Rusin said. "In fixing one problem, they were breaking a whole bunch of other things. Apple developers were focused on fixing bugs in such a way that we could not merge them back into KHTML. Those fixes were never an option for us."

Ignoring for a moment the fact that OSS is not done the "right way" many times, Apple has an obligation to turn out code and to do it fast. They have obligations to their customers. The fact that KHTML wants to take their sweet time and Apple wants to get the patches done fast and out the door shows where the divergence is. Apple can't afford to take the open source approach of spending 5 years in beta before releasing the next version.

Once again a choice by KHTML. The patches are there, but they choose to do the patches their way, thus eliminating Apple patches.

KDE volunteers said they suddenly found themselves dealing with bug reports Apple deemed too sensitive to share, new requirements for auditing code before releasing it, and demands that developers sign nondisclosure agreements before looking at some Apple code.

So you mean once KHTML devs wanted access to code that wasn't part of KHTML, they had to play by Apple's rules? Say it isn't so! Apple plays by their rules for their code, but KHTML doesn't want to play by Apple's rules for Apple code. Again, choices by KHTML to limit their own options.

"As long as they needed us, they used us, but when they gained enough knowledge they had no reason to keep sending us reviews and patches," Rusin said. "At a certain point they decided it was a waste of time for them, and at that point the communication just stopped...We had hopes that they would pour resources into KHTML. But that never happened."

No, it did happen, but they're pouring resources in to the ways that allow them to serve their customers best too, and that means leveraging OS X technologies. KHTML has chosen to be just as uncooperative as Apple.

Re:Here's a quote from Zack Rusin

[Elwood P Dowd]

And even during his rant:

They do the very, very minimum required by LGPL.

And you know what? That's their right. They made a conscious decision about not working with KDE developers. All I'm asking for is that all the clueless people stop talking about the cooperation between Safari/Konqueror developers and how great it is.

So really, the only people they were ever particularly annoyed with were us. Here on Slashdot.

Free donuts

[flaming-opus]

you can't put donuts on the table in the break room and then complain when someone eats them.

Complaining about it shows a great lack of grace.

OSS project mantra

[93 Escort Wagon]

How many times have you filed a RFE on an OSS project and gotten this back? "If you don't like how that works, feel free to submit a patch".

Okay, if you don't like how Apple provides its patches back to the KHTML guys, please feel free to write a tool that converts their patches into the form you prefer.

Apple should start...

[emil]

...by getting everybody on the same source control software.

AFAIK, KHTML uses CVS, and Apple internally uses Perforce.

Nothing constructive can be done until everything is on the same platform.

Apple, offer to buy licenses of your source control software for the KHTML core. Even if they still spurn you, it will appear to the rest of us that you at least tried. You will look more and more of a villan until you make some effort at a reconciliation.

p.s. The KHTML team will need to be conversant with OSX to the point that they can remove GUI calls to it and replace them with QT. If this is a current problem, then some books might be in order.

No problem here

[booch]

This definitely isn't a GPL violation, and doesn't even violate the spirit of Free/Open Source Software. The Apple developers are making their resulting branch of the code available in compliance with the KDE license. They're even trying to work to contribute their changes back to KHTML. Even if the patches don't apply cleanly, the KHTML developers are more than free to look at Apple's changes and add them by hand. Apple is even offering to give back their entire branch, to make it the new official KHTML, since their branch has advanced faster.

This really seems to be a case of the Apple guys offering their changes (or at the very least, making them available), and the KDE guys not being interested in them, or unable to use them for various reasons. It's really hard to blame Apple for that.

Re:Learning about Apple

[slipstick]

And the KHTML guys have decided to "leave it" and explain why. If Apple gets a black eye out of it so be it.

Apple could have tried to be a little more community spirited rather than just ignoring the needs of the very people they relied on to save them millions in development cost. How hard would it have been to include real comments in their patches rather than pointing to a bug database number?

This is being blown *way* out of proportion

[Illissius]

What actually happened, was this: KHTML developer Zack Rusin read one too many uninformed comment on the internets about how awesome the cooperation between KHTML and Apple is; being on the recieving end of the very not awesome cooperation, he understandably got a bit pissed off, and blogged about it. The thing to note here is his ire *was not directed at Apple* (recognizing that they were fulfilling their legal obligations, and were required to do no more), but rather at the uninformed idiots. This has now been spun, in part by those same uninformed idiots, into the KHTML devs being whiny Apple-haters, and the whole legality question has also been quite predictably confused into it all as well, which was never a part of it.
So far, I have seen exactly one comment on this thread with some understanding of this. it'd be sad, if it weren't so fucking ironic...

Re:Safari on Windows?

[Chucker23N]

Um.

Because
1) It's easy to port.
(It's not. Windows doesn't even have native APIs to support Objective C, let alone Objective C++. Porting this means porting large parts of Cocoa.)
2) Safari is just a little front-end for WebCore.
(It's not. Writing a WebCore front-end using WebKit doesn't require, I shit you not, a single line of code. Safari, on the other hand, has many.)
3) Apple would profit tons from this.
(They wouldn't. Giving away a browser for free for a foreign operating system without any other benefits is a honorary decision at best, not a business one.)

What's with the spin?

[eison]

How is this a 'danger', that other people do other things with a project that you have intentionally given the world the right to work on?

Re:I see nothing wrong with it

[argent]

If they would just shut the hell up then we could look at Apple as what they are, a passive OSS user.

You have obviously never looked at how companies that are really passive open source software users behave. Microsoft, for example.

Right now Apple is putting some really interesting code out there for people to pick up. They've provided a framework in darwin for completely changing the way UNIX systems are managed, in nicely packaged tools like launchd, complete with "configure" scripts ready for dropping in peicemeal or packaging for debian or Red Hat.

There's a whole damned revolution waiting on opensource.apple.com and nobody's paying attention to it. Why? because it's from Apple? because nobody knows it's there? Because it's not the Linux Way? I don't know. You tell me.

Re:Sound like KHTML team doesn't want to play eith

[molnarcs]

This is unbelievable (I mean the +5 insightful moderation). How can one claim that KDE developers have been just as uncooperative. They set up a mailing list specifically for Apple devs. They gave CVS access. All they asked was incremental changelogs instead of 60Mb code dumps. They even offered to sign NDAs with apple (and their offer was completely ignored) in order to get them.

"The fact that KHTML wants to take their sweet time and Apple wants to get the patches done fast and out the door shows where the divergence is. Apple can't afford to take the open source approach of spending 5 years in beta before releasing the next version."

This is quite ignorant. There are, admittedly some OSS project that are perpetually at a BETA stage. KDE is not one of them. KDE 3.4 had a few weeks of beta testing, and then it was released as final. Just as Tiger. Yes, there were a few bugs found since RELEASE - just as there were bugs in Tiger, and probably there will be more till the next release.

KDE developers did everything they could to help cooperation - in vain. And they don't even regret that as much as they regret that there are clueless users who overestimate APPLE's contributions.

And this makes hardly any sense:"Once again a choice by KHTML. The patches are there, but they choose to do the patches their way, thus eliminating Apple patches." Excuse me? What were you trying to say?

Mods: congrats!