Microsoft launches open source foundation (Linux-Watch) [LWN.net]

Microsoft launches open source foundation (Linux-Watch)

Posted Sep 12, 2009 12:12 UTC (Sat) by nix (subscriber, #2304)
In reply to: Microsoft launches open source foundation (Linux-Watch) by trasz
Parent article: Microsoft launches open source foundation (Linux-Watch)

Your reasoning here implies that you think a closed-source product has a
less restrictive license than BSD. This is... peculiar.

to post comments

Contributions

Posted Sep 12, 2009 12:37 UTC (Sat) by man_ls (guest, #15091) [Link] (14 responses)

I assume nix means "a closed-source product has a less restrictive license than GPL". It is a very perverse reasoning since proprietary products may give some code, money, employment and all kinds of financial propositions to the upstream BSD project. But they conceal the most important thing of all, and that which is rightly singled out in the GPL: knowledge about what the downstream project is doing, and the ability to share it. With the GPL you can know what the derivative product is doing, and rewrite it in BSD code if you feel the need. A dual license is also a very common proposition.

This is not a theoretical proposition; it happens all the time. Think about proprietary drivers for Mac OS X; BSD cannot touch those, but of course de Raadt can then complain about the lack of wireless drivers. With Linux however the drivers are all there for anyone to see.

Your (trasz's) reasoning about parasitic activities, however, is biased and based on possible but uncommon occurrences. Why can't the BSD-based GPL projects be commercialized and give BSD developers money, fame and glory? Where are the high profile projects that have been relicensed under the GPL?

Contributions

Posted Sep 12, 2009 13:06 UTC (Sat) by trasz (guest, #45786) [Link] (13 responses)

Lets forget for a moment that driver model in BSD and OSX is completely different, and the drivers couldn't be ported from OSX to BSD even if the license allowed for it; also, the drivers are not derived from BSD either. Also, lets forget about Theo not being representative for anything except his own projects, and that OSX is not based on any of them.

So, OSX has closed source drivers. OSX took them from BSD, and BSD license allowed Apple to close the code, so drivers have to be rewritten.

Now, look at the situation in Linux. OSX took them from BSD, and BSD license allowed Apple to close the code, so drivers have to be rewritten.

See? You missed one crucial factor - the fact that the vendor who wants to close their sources simply won't fork from GPL-ed project, unless they are really stupid. The only thing using GPL buys you here is the guarantee that they won't use your code and won't give you anything back.
In the theoretical situation described above, Apple could give back the driver code to the original project, like they gave away e.g. their audit subsystem implementation back to FreeBSD. However, Apple couldn't give anything back to Linux - simply because they chose other system to take code from instead.

Contributions

Posted Sep 12, 2009 15:05 UTC (Sat) by man_ls (guest, #15091) [Link] (2 responses)

Fine, drivers were not the best example in the world; after all there are lots of proprietary drivers for Linux and there is little we can do about it except complain.

But even so: your analogy is again biased. If we are speaking about companies basing their operating systems on Linux, I would say that the situation is rather:

Now, look at the situation in Linux. Red Hat (a $1B company) contributes drivers to Linux. Intel (a $30B company) contributes drivers to Linux. IBM, Novell contribute code to Linux. Manufacturers (Realtek, Broadcom) contribute drivers to Linux. About 194 companies contributed just to the last kernel iteration.

All these companies gave plenty back to Linux. And the code is there for everyone to see, share and expand upon. Other developers (closed or free) can reimplement them with a different license or for another language. Meanwhile, Apple (another $30B company) contributes nothing back to nobody since their code is opaque. Except of course when they feel like it. And BSD enthusiasts feel they are not parasitic.

Contributions

Posted Sep 13, 2009 18:49 UTC (Sun) by drag (guest, #31333) [Link] (1 responses)

> Fine, drivers were not the best example in the world; after all there are lots of proprietary drivers for Linux and there is little we can do about it except complain.

Well actually... Name off all these proprietary drivers?

The only one that people can't seem to do without, right now, would be Nvidia's proprietary drivers. That and some GSM stuff for mobile phones.

Otherwise it's pretty damn easy to do avoid proprietary drivers altogether and yet find inexpensive and high performing hardware for everything you'd want to do.

The main reason why people run proprietary drivers today, other then to get away from Linux/X.org poorly performing, overly complex, and archaic driver models, is because they did not pay enough attention when buying their hardware.

Contributions

Posted Sep 13, 2009 19:14 UTC (Sun) by man_ls (guest, #15091) [Link]

There are proprietary drivers for many pieces of hardware: graphics cards, wireless cards, even for the NSLU2's wired interface. Sure, sometimes there are free equivalents which may or may not have the same functionality; but the thing is that the GPL has not stopped proprietary drivers from sprouting all around, despite Linus' clarification.

Contributions

Posted Sep 13, 2009 9:48 UTC (Sun) by njs (subscriber, #40338) [Link] (9 responses)

I know you're sort of trolling, but talking about Apple as an example is particularly amusing. Apple's grudging contribution of Objective C support to GCC is one of the most famous historical examples of the GPL successfully forcing a company to contribute work back to the community.

Contributions

Posted Sep 13, 2009 15:53 UTC (Sun) by foom (subscriber, #14868) [Link]

Well, *NeXT*'s grudging contribution -- that happened way before NeXT somehow managed to get
Apple to pay to be taken over. :)

Contributions

Posted Sep 13, 2009 17:12 UTC (Sun) by quotemstr (subscriber, #45331) [Link] (2 responses)

And now we're going back to the bad old days of vendor lock-in with Clang and LLVM.

Contributions

Posted Sep 13, 2009 18:51 UTC (Sun) by drag (guest, #31333) [Link] (1 responses)

How is 'back to the bad old days'? Clang and LLVM are completely open source.

Although so far GCC still outperforms them...

Contributions

Posted Sep 13, 2009 19:02 UTC (Sun) by quotemstr (subscriber, #45331) [Link]

Yes, but organizations that extend them don't need to contribute their changes back into the compiler, which will allow them to create a universe of proprietary plug-ins and even forks. "Sure, you can develop for our platform using open-source tools, but if you want to use the latest version of our language, you'll have to fork out for our plug-in".

Contributions

Posted Sep 14, 2009 7:53 UTC (Mon) by trasz (guest, #45786) [Link] (4 responses)

Apple's "contribution" of Objective C back to GCC is one of the examples of how useless GPL actually is - it didn't bring anything actually useful to GCC. See this article for more details: http://www.informit.com/articles/article.aspx?p=1390172

Contributions

Posted Sep 14, 2009 8:30 UTC (Mon) by nix (subscriber, #2304) [Link]

'Cos an article with a title like 'The Failure of the GPL', well, that's
not going to be in any way biased. Oh no.

(And Objective C is not worthless. I use it quite a lot. The language is
worthwhile even without the enormous Apple runtime.)

Contributions

Posted Sep 14, 2009 14:18 UTC (Mon) by pboddie (guest, #50784) [Link] (2 responses)

How "useless" the GPL is? It brought about exactly its intended effects: the availability of the derived work. Whether the released sources were easily integrated upstream is somewhat independent of the licensing, despite what the author of that article would have you believe with his somewhat flimsy reasoning. At least everyone got to see the source code NeXT actually wrote.

Meanwhile, you're complaining (as I note you have done every time you have seen an opportunity to "counter" mentions of the GPL with advocacy of permissive licences - http://lwn.net/Articles/310925/ to provide just one example) that it somehow isn't fair that copyleft-licensed projects can make use of permissively-licensed code. Well, one would think that the numerous advocates of permissive licensing, as they so often urge others to use permissive licences so that they may "consume" such works and incorporate such code into larger works employing different (and frequently proprietary) licences, would be able to stomach their own medicine and live with the effects of releasing such code under such licences themselves. Following up words with actions - that kind of thing.

Or is the real issue here the regret that permissively-licensed code can have a life of its own out in the open, teasing its authors by being improved in full transparency rather than getting upgraded in some proprietary backroom?

Contributions

Posted Sep 18, 2009 21:57 UTC (Fri) by trasz (guest, #45786) [Link] (1 responses)

Intended effect would be for the code contribution to be actually useful. This example shows that forced contribution is not useful at all - who cares about the source code released, if it's missing one crucial part?

As for permissive licenses - I explained this already somewhere else: BSD projects often get important contributions from closed source projects (either code, money or employment), but they almost never get anything from the GPL-ed projects.

Contributions back

Posted Sep 19, 2009 9:09 UTC (Sat) by man_ls (guest, #15091) [Link]

One would have thought that people release code under the BSD so that others can use it, not so that others contribute anything back. After all there is no requirement to give anything back in the license (either in the text or in the spirit). But BSD people still feel free to complain when their users don't contribute anything back. For example you don't see people complaining in GPL projects that they don't get money or employment from proprietary companies, or when they use the code internally (and therefore they don't have to contribute back their changes).

In other words. Your dead horse sounds like "hey, I lend you my bike, no conditions -- hey, I lent you my bike, why don't you ever give me a ride on your car? The neighbor often does". At least with the GPL you are not hiding that you require code contributions in exchange.

Microsoft launches open source foundation (Linux-Watch)

Posted Sep 12, 2009 12:47 UTC (Sat) by trasz (guest, #45786) [Link] (7 responses)

No, let me repeat. When some company takes BSD code from some project and creates its own, closed-source fork, there is a non-zero probability that this company will hire developers from the project and return code that is not a selling point for them. That happens a lot - from PostgreSQL to FreeBSD to Xorg. That's symbiosis.

When GPL-ed project takes BSD code from some project and creates its own, GPL-ed fork, there is zero probability of the original project getting anything in return. This just doesn't happen - for variety of reasons, one of which is that the GPL license prohibits incorporating changes back to the BSD-licensed code. That's not a symbiosis.

Microsoft launches open source foundation (Linux-Watch)

Posted Sep 12, 2009 13:47 UTC (Sat) by xilun (guest, #50638) [Link] (1 responses)

Except that that does indeed sometimes happen, when BSD devs ask kindly (or not so kindly...)

Microsoft launches open source foundation (Linux-Watch)

Posted Sep 12, 2009 14:54 UTC (Sat) by trasz (guest, #45786) [Link]

Could you provide some example? I seem to remember problems with even more trivial stuff, for example GCC developers not wanting to relicense patch fixing some bug from GPLv3 to GPLv2.

Microsoft launches open source foundation (Linux-Watch)

Posted Sep 13, 2009 9:51 UTC (Sun) by njs (subscriber, #40338) [Link] (1 responses)

I've certainly contributed code back to BSD-licensed libraries under a BSD license, when I needed the functionality for my GPL'ed work on top.

I believe that's exactly what you're claiming only proprietary companies do?

Microsoft launches open source foundation (Linux-Watch)

Posted Sep 14, 2009 7:59 UTC (Mon) by trasz (guest, #45786) [Link]

No. I'm not talking about merely using BSD code by GPL-ed projects, just like I'm not talking about merely using BSD code in closed source projects. What I'm talking is created GPL-licensed forks of BSD code and closed-source forks of BSD code.

Microsoft launches open source foundation (Linux-Watch)

Posted Sep 13, 2009 14:55 UTC (Sun) by clump (subscriber, #27801) [Link]

No, let me repeat. When some company takes BSD code from some project and creates its own, closed-source fork, there is a non-zero probability that this company will hire developers from the project and return code that is not a selling point for them. That happens a lot - from PostgreSQL to FreeBSD to Xorg. That's symbiosis.

"This happens a lot" is a pretty thin broth to serve. It's nowhere close to being required by the license to make changes available.

When GPL-ed project takes BSD code from some project and creates its own, GPL-ed fork, there is zero probability of the original project getting anything in return. This just doesn't happen - for variety of reasons, one of which is that the GPL license prohibits incorporating changes back to the BSD-licensed code. That's not a symbiosis.

Zero probablility? This is completely false. You seem to be arguing that since companies that make proprietary code from BSD *can* return code, they *will* return code. Yet your anti-GPL arguments miss the point that GPL code *can* be dual-licensed... allowing them to be used by ... BSD-licensed projects. Need an example? Hostap. Look at NJS's comment as well.

You don't appear to be saying anything on-topic, I might add. It's as if you've been asleep for nine years and woke up fresh to bring the year 2000's favorite dead horse back for new kicks.

Microsoft launches open source foundation (Linux-Watch)

Posted Sep 14, 2009 8:39 UTC (Mon) by hppnq (guest, #14462) [Link]

What annoys me a bit about this whole discussion, is that it is not so difficult to understand the BSD and GPL licenses, and why they are there. In either case, getting anything back other than the warmness that comes from releasing Free Software is not part of the deal. If you are in this business then go write proprietary code, force people to love you or pay their respect or whatever it is that you want, and be happy.

Stop complaining that people make use of the freedom that you have yourself stipulated in your license. This is incredibly stupid, even if someone like Theo de Raadt seems to think differently. I say this because obviously, this is especially true for people who have chosen the BSD license to release code: it permits using it in a GPL project. Whatever happens in the real world is part of the process that leads to chosing the license, it's not the other way around.

Microsoft launches open source foundation (Linux-Watch)

Posted Sep 15, 2009 8:30 UTC (Tue) by dgm (subscriber, #49227) [Link]

I think this is the crux of the matter:
Many "BSD people" release code with the (not so) secret intention of getting hired and making money. The code is a means to that end.
Most "GPL people", on the other hand, are (A) forced to release because they are adding to a GPL'd base or (B) because they care about the code and want help from others.
That GPL'd projects enjoy a wider developer base seems a consequence of type A developers, and maybe the fact that it looks much more fair from an occasional contributor perspective (the ones that would not be hired).
And of course, users don't care much as long as the stuff works.