[Eug]

Apple has purchased the source code for the Common UNIX Printing System

In February of 2007, Apple Inc. acquired ownership the CUPS source code and hired me (Michael R Sweet), the creator of CUPS.

CUPS will still be released under the existing GPL2/LGPL2 licensing terms, and I will continue to develop and support CUPS at Apple.

Should users of Linux and other *nixes be worried? Or will this have minimal effect on them? This will benefit Mac users though, as we'll get better OS X printer drivers out of this.

[nonhuman]

I'd say it's probably nothing to be worried about. Apple's been good about releasing the Darwin source, and the WebKit source. Why would they be any different about CUPS?

[Peter]

I hope this means networking printers will actually be usable.

[Eug]

Originally Posted by nonhuman 

I'd say it's probably nothing to be worried about. Apple's been good about releasing the Darwin source, and the WebKit source. Why would they be any different about CUPS?

Well, there were a lot of people P.O'd about the WebKit source for a long time, because Apple would release nothing for just about forever, and then BOOM! Apple would release volumes of source code all at once... and because of this, Apple's work couldn't be easily integrated into the main fork of KHTML.

Mind you, Apple can be more open about it this time.

OTOH, can Apple now incorporate completely closed source aspects to parts of CUPS now? I don't understand the licencing under this new arrangement. ie. If the licencing and source code releases remain exactly the same, why would Apple need to acquire the ownership to the source code in the first place? They could hire the guy without buying the source code, no?

[olePigeon]

Maybe they can fix the Lexmark printer drivers. I know Lexmark is the devil, but unfortunately a lot of people have them.

[Y3a]

Right after fixing the Epson drivers...

[Eug]

Yeah, printer drivers for OS X are one big Achilles' heel.

Mind you, the drivers for my Canon photo printer are not bad. However, they're not CUPS drivers. They are Canon's.

[ink]

CUPS is licensed under the GPL; why would Linux users be worried? Any source-code changes that Apple makes to CUPS must be made available to those that receive binary versions of it (OSX).

[OreoCookie]

CUPS has also announced that it will continue to release CUPS (the software) under the GPL … I don't think there's anything to worry about.

[Chuckit]

Originally Posted by ink 

CUPS is licensed under the GPL; why would Linux users be worried? Any source-code changes that Apple makes to CUPS must be made available to those that receive binary versions of it (OSX).

The GPL only applies to licensees, not the entity granting the license. Apple can do whatever the hell it wants to its own code.

[Big Mac]

Originally Posted by olePigeon 

Maybe they can fix the Lexmark printer drivers. I know Lexmark is the devil, but unfortunately a lot of people have them.

My little e120N works perfectly well for me and my home network, FWIW.

[ink]

Originally Posted by Chuckit 

The GPL only applies to licensees, not the entity granting the license. Apple can do whatever the hell it wants to its own code.

Right, but if they compile GPL code (CUPS) in their project, they must release their code as well.

[Chuckit]

Originally Posted by ink 

Right, but if they compile GPL code (CUPS) in their project, they must release their code as well.

Not if it's their GPL code. A license only applies to people who are not the copyright holder.

[nonhuman]

Originally Posted by ink 

Right, but if they compile GPL code (CUPS) in their project, they must release their code as well.

Not if it's their own code... They can do whatever they want with it, if they're the owners. It's people who are merely licensing it that have to abide by the restrictions of the GPL. Only licensees are bound by the license.

[Dork.]

Originally Posted by ink 

Right, but if they compile GPL code (CUPS) in their project, they must release their code as well.

Nope. If Apple owns the copyright to the code, they can do whatever they please with it. They can make all the modifications they want, and are not obligated to release a single thing. Heck, they can make modifications, and then package them up and sell them to the highest bidder! They own the copyright, after all.

But the code that is currently available is GPL'd forever, and can't be taken back. And if another party (which does not own the code) makes modifications to the code, and distributes those modifications, they are obligated to release those modifications under the GPL, as well. But just because the modification is released doesn't mean the maintainers of the project are obligated to include it in future releases. (and if a developer who doesn't own the code is unhappy about that, they can fork their own release and try and get people to use it.)

CUPS is different than other Open Source projects: while the software is available under the GPL, when the original developer saw a modification he liked, he didn't integrate it into the main CUPS distribution until he bought (or otherwise obtained) the copyright from the developer. So all the code in CUPS is owned by one entity, which is why Apple could buy it.

Other projects let the original authors retain their copyrights to their bits of the code. Which makes it impractical for a company to just swoop in and "buy" it, because they'd have to agree to terms with each individual contributor.

As to Eug's original question, any Linux developers who are worried can just fork the current CUPS distribution and go off on their merry way. So no, they shouldn't be worried at all.

[Eug]

Interesting. So basically this suggests to me that Apple may want to close source this all up for future releases, which essentially means that it may have already permanently forked.

If this is the case, this could affect other *nix users, in the sense that any public improvements won't have any of Apple's help. Was Apple actively contributing in the past for the global CUPS project? If Apple's contributions to CUPS before were negligible, then the impact to *nix users would also be negligible.

[Dork.]

It's more likely that they (Apple) object to GPL v3, and wanted to buy the project to insure it stays at GPL v2. Outside developers can modify the GPL'd code, but only the rights holders can modify the license terms. Go read the Slashdot thread, they're all over it. The article even has a "rmsproofed" tag right now, implying that Apple wanted to keep Stallman's patent language out of CUPS.

[Eug]

So what's wrong with GPL v3?

Also, do you think Apple will continue to keep all of Apple's future mods open? My guess is no, but they may continue to have some level open sourceness... and will buy mods they like, just like the original owner did.

[Ganesha]

Apple's already made a slight addition the the CUPS licensing.

Apple Operating System Development License Exception;
Software that is developed by any person or entity for an Apple Operating System ("Apple OS-Developed Software"), including but not limited to Apple and third party printer drivers, filters, and backends for an Apple Operating System, that is linked to the CUPS imaging library or based on any sample filters or backends provided with CUPS shall not be considered to be a derivative work or collective work based on the CUPS program and is exempt from the mandatory source code release clauses of the GNU GPL. You may therefore distribute linked combinations of the CUPS imaging library with Apple OS-Developed Software without releasing the source code of the Apple OS-Developed Software. You may also use sample filters and backends provided with CUPS to develop Apple OS-Developed Software without releasing the source code of the Apple OS-Developed Software.

[Chuckit]

Originally Posted by Eug 

So what's wrong with GPL v3?

Basically, Stallman wants to use the GPL to wage a war on patents and DRM, both of which Apple uses.

[ghporter]

"Excuse me, but I seem to have far too little to worry about in my life. Could I borrow some potential trouble from you? Thank you SO much!"

Has Apple EVER goofed over a software using community? I'm not aware of any such incident. Why should they start now?

And when are they going to fix those Brother drivers so I don't print two blank pages at the end of a print job when I print over my network?

[Big Mac]

Originally Posted by Eug 

So what's wrong with GPL v3?

It's complicated, and I haven't really taken the time to understand it that well, but in broad terms v3 closes some loopholes that commercial entities took advantage of to circumvent the code publication requirements of the license. Don't rely on the accuracy of this statement, but I think v3 of the license prevents those licensing code from incorporating said code into their close source hardware and software. TiVO is currently having a hard time with the change.

[ink]

Originally Posted by Chuckit 

Not if it's their GPL code. A license only applies to people who are not the copyright holder.

That's quite an assumption; that they were able to purchase the rights to all the code, and that no contributions have been made by other parties under the terms of the GPL...

In particular; the CREDITS.txt file lists the following (all of whom would be considered Copyright holders):

CREDITS.txt - 2007-02-05
------------------------

Few projects are completed by one person, and CUPS is no exception. We'd
like to thank the following individuals for their contributions:

Nathaniel Barbour - Lots of testing and feedback.
N. Becker - setsid().
Philippe Combes - French localization and buttons script.
Jean-Eric Cuendet - GhostScript filters for CUPS.
Van Dang - HTTP and IPP policeman.
L. Peter Deutsch - MD5 code.
Dr. ZP Han - setgid()/setuid().
Guy Harris - *BSD shared libraries and lots of other
fixes.
Bjoern Jacke - I18N stuff.
Wang Jian - CUPS RPM corrections.
Roderick Johnstone - Beta tester of the millenium.
Till Kamppeter - Bug fixes, beta testing, evangelism.
Kenshi Muto - Japanese localization, patches, and
testing.
Kiko - Bug fixes.
Sergey V. Kovalyov - ESP Print Pro and CUPS beta tester.
Marek Laane - Estonian translation.
Mark Lawrence - Microsoft interoperability testing.
Jeff Licquia - Bug fixes, beta testing, evangelism.
Jason McMullan - Original CUPS RPM distributions.
Wes Morgan - *BSD fixes.
Daniel Nylander - Swedish localization.
Giulio Orsero - Bug fixes and testing.
Michal Osowiecki - Polish localization.
Kurt Pfeifle - Bug fixes, beta testing, evangelism.
Vincenzo Reale - Italian localization.
Petter Reinholdtsen - HP-UX compiler stuff.
Juan Pablo Gonz<E1>lez Riopedre - Spanish localization.
Opher Shachar - Hebrew localization.
Stuart Stevens - HP JetDirect IPP information.
Andrea Suatoni - IRIX desktop integration and testing.
Tomohiro Kato - Japanese localization.
Tim Waugh - Lots of patches, testing, and Linux
integration.
Yugami - LDAP browsing support.

If I've missed someone, please let me know by sending an email to xxxxxxxx

[Eug]

Originally Posted by ink 

That's quite an assumption; that they were able to purchase the rights to all the code, and that no contributions have been made by other parties under the terms of the GPL...

According to the various posts I've seen around the net (including in this thread), all the significant additions to the main branch of CUPS are in fact now owned by Apple. The reason is MR Sweet owned it all already. He acquired the rights to whatever he needed, and turned around and sold it all to Apple.

I wonder how many of those CUPS developers are a bit p'o'd by this.

Originally Posted by Ganesha 

Apple's already made a slight addition the the CUPS licensing.

pple Operating System Development License Exception;
Software that is developed by any person or entity for an Apple Operating System ("Apple OS-Developed Software"), including but not limited to Apple and third party printer drivers, filters, and backends for an Apple Operating System, that is linked to the CUPS imaging library or based on any sample filters or backends provided with CUPS shall not be considered to be a derivative work or collective work based on the CUPS program and is exempt from the mandatory source code release clauses of the GNU GPL. You may therefore distribute linked combinations of the CUPS imaging library with Apple OS-Developed Software without releasing the source code of the Apple OS-Developed Software. You may also use sample filters and backends provided with CUPS to develop Apple OS-Developed Software without releasing the source code of the Apple OS-Developed Software.

This seems to me to be a way for Apple to encourage the companies to write closed source drivers for OS X. This bodes well for Leopard.

P.S. I remember back in the early days before I had a dedicated driver in OS X for my printer. Trying to use CUPS was a major PITA. I really hope Apple does it's GUI magic with generic CUPS drivers quickly.

Something tells me it will be only for Leopard though.

[Big Mac]

CUPS is better than it used to be, Eug. Perhaps they lack features (I'm not sure if that's nearly as true now), but they do work for regular printing. And even Windows users complain about how crappy HP drivers are, so I'm learning to expect less and less in terms of driver quality in general.

[residentEvil]

Originally Posted by ghporter 

"Excuse me, but I seem to have far too little to worry about in my life. Could I borrow some potential trouble from you? Thank you SO much!"

Has Apple EVER goofed over a software using community? I'm not aware of any such incident. Why should they start now?

And when are they going to fix those Brother drivers so I don't print two blank pages at the end of a print job when I print over my network?

i don't have a brother printer to test out, but...

for any printer on our windows based domain, i simply use it's IP address and select generic postscript. it has worked on Xerox, HP, Lanier and Kyorcia network printers without any problems. i'd like to try a make/model to see the problems.

[ChrisA]

Originally Posted by ink 

CUPS is licensed under the GPL; why would Linux users be worried? Any source-code changes that Apple makes to CUPS must be made available to those that receive binary versions of it (OSX).

No. CUPS was licensed under the GPL Now it is licensed under a modified GPL that says the Everyone else but Apple but give their modifications back. Apple has exempted itself from the GPl requirements. hey can do that because they own CUPS now.

Actually the exemption is bigger than that. It says the code written for use on Apple computers does not need to be shared under the terms of the GPL.

I sure wish the pepole who write SAMBA would make an"Apple exception" and the GCC people too. What happens if they said "This software may not be re-distributed by Apple Inc."

[peeb]

Dork summed it up right - none of this affects people developing CUPS. All this does is allows Apple to create a Mac only version, and not release the source for the new stuff they do.

[TETENAL]

The exception that drivers running on an Apple OS and linking to CUPS need not be open source was there since Apple used CUPS in 2002. It was granted by the previous owner of CUPS.

The SAMBA or GCC people could add an exception that forbids Apple from distributing their software for all future versions of their software provided they own the copyright of all that code. They then can decide at will under which license they distribute. Since part of the GCC code is written by Apple afaik it would be kind of hard to have all contributors to agree on that though. If they removed all of Apple's code maybe. Why would the SAMBA or GCC teams do something like that though? I don't see how that would make sense.

[Dork.]

Remember that the GPL only applies when you use other people's code. It basically gives you a license to use other people's code however you see fit, as long as you publish your contributions under the same terms. You still retain your copyright, though, and can do what you want with your code.

The folks that make QT actually distribute it under two licenses: It's distributed to everyone under the GPL, which requires users to distribute their modifications under the GPL, but parties that wish to develop proprietary applications without sharing their source code can buy a commercial license. Trolltech can do this because they make sure all the code in the "official" version is owned by them, either by writing it themselves or buying the rights off the author.

[besson3c]

Originally Posted by Eug 

Well, there were a lot of people P.O'd about the WebKit source for a long time, because Apple would release nothing for just about forever, and then BOOM! Apple would release volumes of source code all at once... and because of this, Apple's work couldn't be easily integrated into the main fork of KHTML.

Mind you, Apple can be more open about it this time.

OTOH, can Apple now incorporate completely closed source aspects to parts of CUPS now? I don't understand the licencing under this new arrangement. ie. If the licencing and source code releases remain exactly the same, why would Apple need to acquire the ownership to the source code in the first place? They could hire the guy without buying the source code, no?

Also, Apple has lagged behind in releasing Darwin kernel source code, making the project nearly useless/dead. I'm subscribed to Apple's Calendar Server list, and the list has been so quiet and so little activity on the project website that members started to ask whether the project was dead!

Apple is not so great about working with open source developers, unfortunately. This could potentially spell trouble for Linux users.

[besson3c]

As to Eug's original question, any Linux developers who are worried can just fork the current CUPS distribution and go off on their merry way. So no, they shouldn't be worried at all.

True, but it would kind of suck if it came to this. I used a gimp-print (now called Gutenprint) driver with OS X for a long time. If Apple craps on the developers, perhaps they will be much less interested in working with Apple on integrating some of their work into OS X?

It is in Apple's best interest to be on good terms with open source developers. Hell, I bet several of their own developers were at one point open source developers.

[besson3c]

Originally Posted by ghporter 

Has Apple EVER goofed over a software using community? I'm not aware of any such incident. Why should they start now?

Sure they have. Charging for .Mac, ripping off Watson, Dashboard, their relationship with KHTML developers, Darwin kernel developers, Mac OS X 10.0, etc.

[besson3c]

Here is the Slashdot article, I'm reading through it now:

Slashdot | CUPS Purchased By Apple Inc.

[Dork.]

Originally Posted by besson3c 

True, but it would kind of suck if it came to this. I used a gimp-print (now called Gutenprint) driver with OS X for a long time. If Apple craps on the developers, perhaps they will be much less interested in working with Apple on integrating some of their work into OS X?

It is in Apple's best interest to be on good terms with open source developers. Hell, I bet several of their own developers were at one point open source developers.

It wold suck less than if CUPS was a proprietary program that Apple bought. Then, you don't even have the option to fork it, and you'd be at the mercy of Apple for all support. At least now, if Apple decides to stop supporting other platforms, "The Community" can provide that.

[besson3c]

Originally Posted by Dork. 

It wold suck less than if CUPS was a proprietary program that Apple bought. Then, you don't even have the option to fork it, and you'd be at the mercy of Apple for all support. At least now, if Apple decides to stop supporting other platforms, "The Community" can provide that.

True...

Originally Posted by besson3c 

Sure they have. Charging for .Mac, ripping off Watson, Dashboard, their relationship with KHTML developers, Darwin kernel developers, Mac OS X 10.0, etc.

Watson? Are you serious. Apple ripped off an app that ripped off Sherlock? Dashboard ripped off Konfabulator that ripped off Apple's Desk Accessories. Give me a break.

[besson3c]

Originally Posted by hayesk 

Watson? Are you serious. Apple ripped off an app that ripped off Sherlock? Dashboard ripped off Konfabulator that ripped off Apple's Desk Accessories. Give me a break.

Just what is original? Desk Accessories? Every operating system has mini single-purpose applications such as this, how can you pinpoint the source of such an innovation, and why does it even matter? There is no little in an operating these days that is truly new and innovative.

There is no way of knowing whether Watson or Apple decided to extend upon the basic functionality of Sherlock, but obviously Watson was inspired by Sherlock.

Either way, my point is that Apple has upset software communities, that was my point.

[Chuckit]

Originally Posted by besson3c 

Charging for .Mac

Point.

Originally Posted by besson3c 

ripping off Watson

Arguably helped the software-using community, harmed one software provider.

Originally Posted by besson3c 

Dashboard

See above.

Originally Posted by besson3c 

their relationship with KHTML developers

You'd have to be a complete drama queen to say they screwed anyone over there. That's like saying the Debian guys screwed over Torvalds by creating their own distro.

Originally Posted by besson3c 

Darwin kernel developers

Because they were a little late releasing newer internal revisions of the code? That seems like a bit of an exaggeration.

Originally Posted by besson3c 

Mac OS X 10.0

Point.

[besson3c]

Originally Posted by Chuckit 

Arguably helped the software-using community, harmed one software provider.

Harmed all of the people that paid for Watson, and arguably discouraged software developers to write cool stuff given the possibility that Apple will release something identical and put them out of business.

You'd have to be a complete drama queen to say they screwed anyone over there. That's like saying the Debian guys screwed over Torvalds by creating their own distro.

They took all of the KHTML base originally, and didn't give back their code in a useful and documented format to be reabsorbed by the community. I don't know what the license terms of KHTML were, but if it were GPL this would mean that technically Apple confirmed to the agreement by giving back code, but not initially in a courteous manner. Now that the project has forked, this is now no longer relevant.

Because they were a little late releasing newer internal revisions of the code? That seems like a bit of an exaggeration.

The Darwin kernel source was out of sync with the OS X kernel source. This made it difficult for people who wanted to develop kernel extensions, as they could not really test their work against what end users were going to be running.

[Chuckit]

Originally Posted by besson3c 

Harmed all of the people that paid for Watson

Even though they still got what they paid for? And I suppose The Prestige harmed all the people who had paid money to go see The Illusionist?

Originally Posted by besson3c 

and arguably discouraged software developers to write cool stuff given the possibility that Apple will release something identical and put them out of business.

I don't think you could make a very strong argument along those lines. Even Dan Wood has moved on, made another product (in an area Apple already occupied, no less) and seems to be doing just fine.

Originally Posted by besson3c 

They took all of the KHTML base originally, and didn't give back their code in a useful and documented format to be reabsorbed by the community. I don't know what the license terms of KHTML were, but if it were GPL this would mean that technically Apple confirmed to the agreement by giving back code, but not initially in a courteous manner. Now that the project has forked, this is now no longer relevant.

It was just sniveling by a few FOSS nuts from the start, not a legitimate instance of harm to end users (or even to the KHTML team).

Originally Posted by besson3c 

The Darwin kernel source was out of sync with the OS X kernel source. This made it difficult for people who wanted to develop kernel extensions, as they could not really test their work against what end users were going to be running.

That's not really true. It was possible to test kernel extensions on OS X all along, as far as I can recall. This isn't Linux; recompiling the kernel isn't necessary for that sort of thing. I've always assumed Apple was slow about releasing kernel snapshots because they didn't think anybody had an urgent need for them.

[besson3c]

Chuckit,

I know you want to play Apple fanboy and find a way to prove me wrong and defend Apple to the bitter end just for the sake of it, but I haven't the time nor the energy, especially since you don't seem to have the time and energy to read up on these cases to make an argument that addresses the real underlying issues.

I will say that Apple was slow on releasing kernel snapshots because, IIRC, they were afraid that people would figure out how to get the kernel to run on Intel hardware. As a compromise to this, I believe they stripped out their own kernel extension that monitors for OS X running on Mac hardware. This code may have originally not been a kernel extension at all, although I don't remember the details. Because kernel extensions do not run entirely independent of the kernel and need to be tested against the currently running kernel, this was a problem. Ditto for some user space programs.

Read up on all of this if you really want to continue this conversation... I'm assuming you are just looking for an argument though.

[Eug]

Chuckit,

To claim that the above is just par for the course is cutting Apple a little too much slack. The truth is Apple did piss off a fair number of people. Apple could have been a lot more friendly in the way they conducted themselves with the open source community, etc, even if they technically didn't do anything wrong in the legal sense.

[besson3c]

Originally Posted by Eug 

Chuckit,

To claim that the above is just par for the course is cutting Apple a little too much slack. The truth is Apple did piss off a fair number of people. Apple could have been a lot more friendly in the way they conducted themselves with the open source community, etc, even if they technically didn't do anything wrong in the legal sense.

And it's probably because of these sorts of stunts, incl. the KHTML deal, the relationship with the FreeBSD community, the Darwin source deal, etc. that people are hesitant to participate in Apple open source projects such as their Calendar Server, which appears to be dead.

It's bizarre what is in it for Apple at all, unless they recognize the errors in their past ways and are hoping to repair this damage. Hosting projects such as MacPorts may be a good start, although I've fallen out of touch with this community so that I don't know how responsive and cooperative Apple has been here.

My perception is that Bonjour/Zero-conf and Webkit have been at least fairly successful open source projects though. I don't know of anybody that is interested in using launchd though (including myself, don't really like it).

[Chuckit]

Originally Posted by besson3c 

I know you want to play Apple fanboy and find a way to prove me wrong and defend Apple to the bitter end just for the sake of it, but I haven't the time nor the energy, especially since you don't seem to have the time and energy to read up on these cases to make an argument that addresses the real underlying issues.

You have yet to demonstrate any gaps in my knowledge. You have, however, started attacking me personally, which doesn't seem like a good sign that you have a legitimate argument.

Originally Posted by besson3c 

Because kernel extensions do not run entirely independent of the kernel and need to be tested against the currently running kernel, this was a problem. Ditto for some user space programs.

I can assure you, each and every copy of OS X includes the version of the kernel that goes with that version of OS X. Testing against the currently running kernel is not a problem.

Originally Posted by Eug 

To claim that the above is just par for the course is cutting Apple a little too much slack. The truth is Apple did piss off a fair number of people. Apple could have been a lot more friendly in the way they conducted themselves with the open source community, etc, even if they technically didn't do anything wrong in the legal sense.

I'm not denying Apple has pissed off a few people. I'm saying that most cases of Apple "screwing over" large communities are wild exaggerations at best. Apple ripping off Watson and Kaleidoscope was shitty — feel free to Google for my previous posts on the subject and you'll see that I agree with this. But as far as I can tell, that's just a couple of people who found themselves inadvertently competing with Apple, not a large software-using community getting intentionally screwed over.

Yes, Apple has done things that pissed people off before. A fair number of people are pissed off by the GPL as well. Another fair number of people are pissed off by the BSD license. I'm not claiming, however, that the FSF or Berkeley are screwing people over left and right — and there's a more legitimate claim there than, say, with Apple and KHTML.

[besson3c]

Originally Posted by Chuckit 

You have yet to demonstrate any gaps in my knowledge. You have, however, started attacking me personally, which doesn't seem like a good sign that you have a legitimate argument.


I can assure you, each and every copy of OS X includes the version of the kernel that goes with that version of OS X. Testing against the currently running kernel is not a problem.

For starters, here is your gap...

In order to develop a kernel extension, it is most likely that you will need to see the kernel source. If the kernel source you have access to is not in parity with the shipping version of the OS, you have a problem.

You have not demonstrated that you really understand what is at the heart of some of the issues I have laid out, specifically this one, as well as quite likely not really knowing the grievances the KHTML team had with Apple. I'd have to research this myself to be precise about it, but there was more to it than some FOSS nuts sniveling, as you put it.

If memory serves me, their problem was that there was a long period of time where Apple was not releasing code back to the KHTML group, and then did so in one mad spurt, and it was largely undocumented and written in such a way that it would be difficult to merge back into the main branch. This pretty much forced a fork, mostly because it become clear that Apple was not really interested in being a courteous developer that works well with the existing developers, but simply a leech and a bully.

Normally when you work on an open source project, there is ongoing dialog... There is some democracy, and decisions made as a group. There are efforts to integrate changes contributed into the main branch so that there aren't different versions of the same product floating around. While Apple legally did nothing wrong, what they did was not terribly courteous, and the KHTML developers have every right to be pissed.

Yes, Apple has done things that pissed people off before. A fair number of people are pissed off by the GPL as well. Another fair number of people are pissed off by the BSD license. I'm not claiming, however, that the FSF or Berkeley are screwing people over left and right — and there's a more legitimate claim there than, say, with Apple and KHTML.

Which is?

[besson3c]

Chuckit,

While I'm attempting to fill in gaps (and hopefully succeeding at being accurate , you seem to think that Linux users have to recompile their kernels every 5 minutes. This is simply untrue.

Firstly, support for everything of use to an intended audience should be compiled in by default. It doesn't make sense to compile in SAN drivers for a consumer OS, for instance, so it is likely that Ubuntu Desktop does not include SAN drivers. The same is true for any operating system - an optimized, smaller sized kernel can perform faster, so there is usually some sort of attempt to come up with some sensible defaults rather than ship a completely bloated kernel full of useless cruft.

Beyond what is provided by the kernel, you can *supplement* the support provided by a BSD kernel with a kernel extension. This does not override the functionality of the included kernel, but simply adds to it. If you need to modify the way the kernel behaves with a certain kernel driver, you need to actually modify the kernel source. For example, in FreeBSD you need to modify the kernel source and recompile/reinstall if you want to use the Hauppauge drivers (which are kernel extensions) provided in the ports tree.

In Linux, if you want to supplement what is provided in the kernel, you either need to enable some additional options in the kernel and recompile, or build in support yourself and recompile. While this sounds less desirable, the advantage to the Linux approach is that programs cannot install rogue kernel extensions that do malicious things and load dynamically. Additionally, a kernel extension loaded into a BSD system may be unstable, uncertified, and bring down the system. I had some problems upgrading to FreeBSD 6.2 recently caused by kernel extensions I forgot were running.

In both a BSD or Linux system, it is always desirable to include support for the things you need in the base kernel, certified and within the radar of the kernel developers. Both BSD and Linux kernels (Mach and Micro, whichever is which, I always get the two confused) are restrained by the same sorts of decisions about what support is included in the default kernel.

[nonhuman]

For the record, you do not need to recompile the Linux kernel just to supplement it. Linux allows for kernel modules that do the same thing as Apple's kernel extensions.

[besson3c]

Originally Posted by nonhuman 

For the record, you do not need to recompile the Linux kernel just to supplement it. Linux allows for kernel modules that do the same thing as Apple's kernel extensions.

Didn't know this, it is so rare that I need to use any sort of kernel module in either a BSD or Linux based kernel.

Here is more info about this, Chuckit:

http://tldp.org/HOWTO/Module-HOWTO/x73.html

[ink]

Originally Posted by ChrisA 

No. CUPS was licensed under the GPL Now it is licensed under a modified GPL that says the Everyone else but Apple but give their modifications back. Apple has exempted itself from the GPl requirements. hey can do that because they own CUPS now.

That's my point, though. Apple can't change the license without consent from all the copyright holders. Perhaps they have received indemnification from the hundreds of people who have submitted code to CUPS; who knows.

I sure wish the pepole who write SAMBA would make an"Apple exception" and the GCC people too. What happens if they said "This software may not be re-distributed by Apple Inc."

Two wrongs don't make a right. Apple can be a prick about this, but I don't think Samba, et. all should stoop to that level.

Now, as to whether or not Novell/SuSE, RedHat, Gentoo, IBM, Debian, etc. will want to contribute back to Apple's CUPS project.. or fork a new project, we will see.