|
|
Log in / Subscribe / Register

Debian and Nexenta collide

[Posted November 9, 2005 by corbet]

When Sun Microsystems decided to release Solaris under the CDDL license, it did so with the knowledge that this code could not be combined with GPL-licensed code. That incompatibility was pretty much guaranteed to create some interesting conflicts at some time. That time appears to have arrived, thanks to the release of Nexenta, a Debian-based system built on top of the Solaris kernel and runtime libraries. How this conflict is resolved may set the tone for how the GPL and CDDL worlds intersect in the future.

The Nexenta developers got off to a bit of a bad start by announcing its existence while putting its entire web site behind a password gate. Once general access was allowed, developers discovered that binaries of their software were being distributed without the associated source. The Nexenta developers responded in a rather unhelpful manner:

Also some stuff not committed yet, beca[u]se we are testing them. In 2-3 months we are hoping to sort out all these "starting" issues with code browsing, scripts availability, etc.

Anybody who has hung around anywhere near the Debian community for any period of time will know immediately that this sort of answer is unlikely to go over well. Various developers responded with requests to delete the binaries immediately, and some even pondered the use of a DMCA takedown notice. The Nexenta developers appear to have taken the hint, and source availability has improved, though the occasional glitch still comes to light.

The hardest issue, however, remains unresolved. The Nexenta project uses, along with the Solaris kernel, a number of user-space libraries (including the core C library) from Solaris. These libraries, being licensed under the CDDL, are not compatible with GPL-licensed applications. But much of Nexenta's user space is GPL licensed, and is linked against Sun's libc. And, in particular, much of the management infrastructure which makes Nexenta a Debian-derived distribution is built this way.

Several Debian developers are claiming that distributing GPL-licensed applications linked to a CDDL libc constitutes copyright infringement and should be stopped. The Nexenta developers, instead, justify this distribution by citing the "system software" exemption in section 3 of the GPL:

However, as a special exception, the source code distributed need not include anything that is normally distributed (in either source or binary form) with the major components (compiler, kernel, and so on) of the operating system on which the executable runs, unless that component itself accompanies the executable.

This exemption has allowed the distribution of, for example, binaries of GPL applications for Solaris for many years. The Debian folks respond that this situation is different: since the libraries and the GPL applications are all part of the Nexenta distribution, the CDDL-licensed libc does, indeed, accompany the executable, and the exemption does not apply.

The Nexenta developers do not appear to entirely buy this argument. They have suggested, however, that Nexenta could be split into two pieces: the CDDL-licensed core, and the GPL-licensed applications. Once the core is installed, the applications could be brought in from a repository somewhere. The only problem there is that bringing in those applications requires the use of (GPL-licensed) tools like dpkg, which would thus have to be distributed with the core system. Getting past this little bootstrap issue could be a challenge.

Once again, Nexenta has not helped itself here: project developers have suggested that the Debian community might want to help them out by relicensing dpkg under CDDL-compatible terms. Suffice to say that this idea was not received enthusiastically. The idea of rewriting dpkg as a CDDL application has also been raised, though that raises some issues of its own.

A more plausible solution to this problem might be to get Sun to relicense its libraries in a GPL-compatible way. Nobody has asked Sun (publicly, at least) whether it would be willing to take this step, but, once again, Sun was certainly aware of the consequences of its licensing decisions when it made them. This situation could also be resolved by porting the GNU C library to the Solaris kernel and shipping it with Nexenta. This is evidently a big task, and the Nexenta developers (who seem to be fairly small in number) are not thrilled about taking it on.

The licensing issues are real, and need to be worked out. But many of the people involved in the debate appear to have lost track of the fact that the Nexenta project, while perhaps being occasionally arrogant and ignorant of how Debian does things, is trying to make a contribution to the free software world. It is a free software project. Anthony Towns has been almost the lone voice in calling for a higher degree of cooperation with Nexenta:

I'm amazed at the level of intolerance that's greeting a pretty major contribution to the free software community. There are, what, five major OS/kernels for PCs/workstations these days -- Windows, OS X, Solaris, BSD and Linux. How does it make any sense at all to be hostile to the fact that now four out of those five are free at their core?

He also points out that Debian's hands have not always been 100% clean, and that there is more to gain by helping a project like this toward free-software purity than by threatening legal action against it. With luck, the community will hear this message. What Nexenta is doing is very much within the spirit of free software licensing; with patience and help, they should be able to get within the letter of those licenses as well.

to post comments

Debian and Nexenta collide

Posted Nov 9, 2005 18:04 UTC (Wed) by ballombe (subscriber, #9523) [Link] (6 responses)

Nexenta would have helped itself by not assuming it was the first non-Linux Debian port. popcon.debian.org would have enlightened them. Cooperation works both way.

BSD libc already used in Debian

Posted Nov 9, 2005 21:28 UTC (Wed) by ncm (guest, #165) [Link] (5 responses)

Indeed, it ought to be easy enough to adapt one of the Debian/kBSD libc ports to call Solarix syscalls instead. Distance from glibc is no bad thing.

BSD libc already used in Debian

Posted Nov 9, 2005 21:47 UTC (Wed) by mel (guest, #5484) [Link] (3 responses)

How about applications like xmms that use glibc pretty heavily? Like the
glibc lists for example or is that something that is already addressed?

BSD libc already used in Debian

Posted Nov 9, 2005 22:19 UTC (Wed) by ncm (guest, #165) [Link]

XMMS is not part of the base system, so there are any number of ways to accommodate it -- if indeed it is needed at all. Most likely anybody running a Debian GNU/Solarix system will be operating it as a server, and have no use for XMMS.

Dpkg, by contrast, isn't terribly demanding.

BSD libc already used in Debian

Posted Nov 9, 2005 23:53 UTC (Wed) by jamesh (guest, #1159) [Link] (1 responses)

Are you perhaps confusing glib and glibc?

BSD libc already used in Debian

Posted Nov 10, 2005 12:06 UTC (Thu) by havardk (subscriber, #810) [Link]

Are you perhaps confusing glib and glibc?
Ah, probably, you had me confused there for a while. XMMS is reasonably portable, and runs just fine on Solaris with the Sun libc.

BSD libc already used in Debian

Posted Nov 17, 2005 22:57 UTC (Thu) by rqosa (subscriber, #24136) [Link]

Actually, Debian GNU/kFreeBSD uses a port of glibc. Maybe Debian GNU/NetBSD uses the NetBSD libc (the website doesn't say).

Debian and Nexenta collide

Posted Nov 9, 2005 18:09 UTC (Wed) by cventers (guest, #31465) [Link] (36 responses)

It does bother me that Nexenta and Debian are fighting in this way;
however, I can't blame either party. I think the party that needs to
resolve this issue is Sun Microsystems.

Sun's no stranger to the fact that open source is taking over the world.
They're willing to license OpenOffice LGPL, which is fantastic (and I
hypothesize is done simply to take shots at Microsoft's OS monopoly).

But their two big camps, Solaris and Java, are strange mutilations of
what many of us accept as real open source. When Solaris became available
under the CDDL, I was a bit miffed by their press blitz. They asserted
that their offerings were real open source and tried to gloss over the
details by making chest-pounding statements like "Now that the licensing
debate is over, we can start competing in terms of actual technology".
Does Sun expect volumes of open source developers to drop what they're
doing and go running with open arms to Sun, the CDDL, and their
bizarro-world of half-open technology?

If Solaris were made GPL, I'd know for a fact that Sun was interested in
open source as a viable way to do business. The CDDL tells me that Sun is
interested in exploiting the open source idea for their own gain alone.

Sun's stepped into the party; now it's time they take off their coat.

Debian and Nexenta collide

Posted Nov 9, 2005 18:29 UTC (Wed) by ajross (guest, #4563) [Link] (3 responses)

The CDDL tells me that Sun is interested in exploiting the open source idea for their own gain alone.

Well, that might be a bit much. Yes, Sun's decision to use a GPL-incompatible license for OpenSolaris was shortsighted and detrimental to the community at large. It disallows the most obvious opportunities for collaboration (driver and filesystem cross-pollination with linux) and basically fragments the community into "us" and "them", which is dumb and causes flamewars like this one.

That said, it should be remembered that the CDDL is a free software license. Regardless of Sun's motivation, the community is enriched by OpenSolaris's availability under the CDDL. Punishing Sun because they made a poor license choice and didn't enrich us as much as they potentially could have seems unfair to me.

Debian and Nexenta collide

Posted Nov 10, 2005 11:16 UTC (Thu) by slef (guest, #14720) [Link] (2 responses)

CDDL is not a free software licence. It restricts the actions of support agents and other similar people. At least as far as the debian free software guidelines go, that seems like discrimination against a significant field of endeavour to me. I don't know how that sits with FSF or OSI, but who cares? I don't want copyright licences trying to tell me any agent contracts can't cover the software I supply. These days, copyright law is so strong that I fear it would override my commercial agreements. Anyone know whether that's unfounded?

There is no good reason for the support disclaimer clause to be drafted in the obnoxious absolute terms that it is. It looks like a drafting bug. Why won't Sun fix it?

Debian and Nexenta collide

Posted Nov 15, 2005 11:06 UTC (Tue) by KotH (guest, #4660) [Link] (1 responses)

CDDL might not be a free license, but neither is GPL. Both restrict their users in a way that quite a lot of people do not like. Yes, you might now argue that GPL does that to ensure that the source will be always public. But that view is shortsighted. Just have a look how many problems the BSD people have with GPL software, although they write open source software too. Or as an other example, have a look at the problems the MPlayer project had with GPL until all code it used was relicensed under GPL. Or to cite Arpi on an never published interview for "Brave GNU World" on the question what his biggest problems were when developing MPlayer: "The GPL"

Debian and Nexenta collide

Posted Nov 15, 2005 12:44 UTC (Tue) by slef (guest, #14720) [Link]

I don't know what personal definition of "free software" you are using, but GPL is free software: that is, it follows the debian free software guidelines. By restricting support contracts, the CDDL is not. Sun want control of the support market to themselves and the CDDL is one tool to help them achieve that.

The MPlayer flameboys don't much interest me, but I wonder how you can cite something never published and expect anyone to accord it much weight.

Open is Open.

Posted Nov 9, 2005 21:18 UTC (Wed) by gregwilkins (guest, #515) [Link] (26 responses)

You are forgetting the whole point. The source to solaris is Open!
You can see it.
You can compile it.
You can modify it.
You can distribute it.
You can fork it.
If Sun disappear or change the license, you are free to continue
development yourself.

Sun have GIFTED their users and clients freedoms that few other
companies in our industry have. Their gift is no less a gift
because they also try to run a business and try to maximize the
good publicity from it.

They are betting the company on their belief that they offer
value add to their customers even though cheaper boxes are
available elsewhere and the software is available for free.

They have committing themselves to a path of inovation and
development rather than licensed lock it.

Bravo!

Open is Open.

Posted Nov 9, 2005 22:10 UTC (Wed) by cventers (guest, #31465) [Link] (24 responses)

Don't be so sure. The reason I actually raise objection (aside from the
confusion) is because of a critical difference between the GPL and the
CDDL that could having a stunningly bad impact on the ability of an open
source developer or team to fork either a CDDL project or a GPL one.

Quoting the GPL preamble:

>> Finally, any free program is threatened constantly by software
>> patents. We wish to avoid the danger that redistributors of a free
>> program will individually obtain patent licenses, in effect making the
>> program proprietary. To prevent this, we have made it clear that any
>> patent must be licensed for everyone's free use or not licensed at
>> all. Finally, any free program is threatened constantly by software
>> patents. We wish to avoid the danger that redistributors of a free
>> program will individually obtain patent licenses, in effect making the
>> program proprietary. To prevent this, we have made it clear that any
>> patent must be licensed for everyone's free use or not licensed at
>> all.

LWN itself said:

>> In other words, the CDDL does not license any patents for use in
>> derived products.

in an article about the CDDL. If you doubt the CDDL's stance on patents,
read it yourself:

http://www.sun.com/cddl/cddl.html

What does this mean? Sun wants you to contribute to their open source
offerings. But they have engineered specific clauses in the CDDL to make
sure that their patents are licensed in such a way that they are free to
take them back.

LWN previously reported that the GPL and the CDDL are very similar and
were only incompatible on technical reasons. I don't think Sun is dumb or
confused about licenses at all. I think the reason the CDDL exists is
because they did not believe in the patent freedom granted by the GPL -
in other words, they want to be able to back out it at any time. To me,
that is less than genuine, and discourages me from contributing any real
time to products licensed under the CDDL.

Open is Open.

Posted Nov 9, 2005 22:12 UTC (Wed) by cventers (guest, #31465) [Link]

Sorry, my post contains an inaccuracy I didn't catch in preview. I said:

>> is because of a critical difference between the GPL and the
>> CDDL that could having a stunningly bad impact on the ability of an
>> open source developer or team to fork either a CDDL project or a GPL
>> one.

That should read:

is because of a critical difference between the GPL and the CDDL that
could have a stunningly bad impact on the ability of an open source
developer or team to fork a CDDL project, as opposed to a GPL one.

Open is Open.

Posted Nov 10, 2005 2:01 UTC (Thu) by paulj (subscriber, #341) [Link] (4 responses)

>> In other words, the CDDL does not license any patents for use in >> derived products.

URL please? I don't think Jon ever wrote anything as misinformed as that. However, it does sound like it could be from the Solaris 10 "review" which ran here after release of S10, by Joe "Zonker (??)" Brockheimer, where, rather than reviewing S10 in any meaningful way, he instead used the article as a platform for general Sun bashing and FUD wrt to the (not yet released at that stage) OpenSolaris project.

That article still sticks in my throat ;).

Open is Open.

Posted Nov 10, 2005 2:15 UTC (Thu) by cventers (guest, #31465) [Link] (3 responses)

http://lwn.net/Articles/114839/

Seems pretty accurate to me. You have to get the terminology straight for
a straight reading of the CDDL. The patent grant is here:

>> (b) under Patent Claims infringed by the making, using or selling of
>> Original Software, to make, have made, use, practice, sell, and offer
>> for sale, and/or otherwise dispose of the Original Software (or
>> portions thereof).

"Original Software" is the unmodified software, released under the CDDL,
by Sun. They agree not to pursue you for using the unmodified software.

Open is Open.

Posted Nov 10, 2005 3:25 UTC (Thu) by paulj (subscriber, #341) [Link] (2 responses)

Ah, right. Yes, that is quite restrictive at first reading. Reasonably sure there are good legal reasons for that. The most obvious perhaps being the difficulty of determining at what point transformations of a work cross from derived to original. That'd be a huge grey area to get into, both for Sun, for contributors to the CDDL and for users.

From a practical POV, it's maybe not quite as restrictive as it may seem. Note that the grant extends to the Original Software and portions thereof, so you may (by my Joe Bloggs, IANAL, ordinary member of the public reading) strip down the Original Software and use the portion you want, under the terms of the CDDL and that grant. Functional, non-expressive modifications might still be allowed, as they don't affect copyright at least - however I'm uncertain if that extends both ways (ie an author can snarf such modifications at will, does it go the other way? I don't know).

Ie, my inclination would be to apply Occam's razor and take the view that its phrased as it is so as to provide clarity as to the exact reach of the patent grants, rather than as some sinister plot by Sun to sue people for patent breaches. I honestly don't think Sun are interested in suing people for modifying CDDL code as long as they honour the CDDL. Nor do Sun have any track record of patent litigation (well, other than being threatened or sued ;) ).

You might want to bring this up on the CAB-Discuss, might be good to explore it there further, to see what the intent was with that clause and whether or not, if the language does not properly reflect the intent, it would be possible to modify it in a future version of the licence.

Open is Open.

Posted Nov 10, 2005 4:32 UTC (Thu) by cventers (guest, #31465) [Link]

paulj, please see my other response.

Open is Open.

Posted Nov 10, 2005 8:56 UTC (Thu) by job (guest, #670) [Link]

Remeber the day last year when Microsoft and Sun became friends? "Microsoft is to pay $700m to Sun to settle anti-trust issues and $900m to resolve patent issues. It is also making an up-front payment of $350m to Sun to cover patent royalties for server products." It that's not a serious source of income for Sun, I don't know what is. And whether Sun would sue free software developers in the future is not something we have on paper, management change and Sun has a rocky ride ahead. Nobody knows what'll happen then.

Open is Open.

Posted Nov 10, 2005 2:14 UTC (Thu) by paulj (subscriber, #341) [Link] (17 responses)

replying further...

But they have engineered specific clauses in the CDDL to make sure that their patents are licensed in such a way that they are free to take them back.

Please, you've fallen for the FUD. You gave the URL for the CDDL and advised the poster you were replying to to read it. So I have to ask you to do the same. Where exactly does the CDDL give Sun the power to "take them back" wrt patents?

The CDDL obliges the initial developer (eg Sun), wrt any patents embodied in the covered software (eg OpenSolaris) to grant "world-wide, royalty-free, non-exclusive" licences to all users and contributors of/to the CDDLed code, conditional on following section 3.1 of the CDDL (the "must make source available" clause). Further, all contributors must similarly grant such a licence.

I know I've got big purple blinkers on (they're welded to our head as part of our indoctrination at Sun), so probably I can't see the evil "All your patents belong to Sun (evil laugh)" clauses, so maybe you'd be kind enough to point them out to me explicitely. ;)

Open is Open.

Posted Nov 10, 2005 2:46 UTC (Thu) by cventers (guest, #31465) [Link] (16 responses)

No, I've read no one's FUD. I read the CDDL myself, and I don't trust the
language. Look at section 2.1, License Grants.

>> (a) under intellectual property rights (other than patent or
>> trademark) Licensable by Initial Developer, to use, reproduce, modify,
>> display, perform, sublicense and distribute the Original Software (or
>> portions thereof), with or without Modifications, and/or as part of a
>> Larger Work; and

So far so good. Notice that Modifications is specifically mentioned here
- you can retain intellectual property rights (poor term) on modified
versions (except for patents and trademarks).

Question: Why are patents not simply rolled into 2.1 (a)? Let's look at
2.1 (b):

>> (b) under Patent Claims infringed by the making, using or selling of
>> Original Software, to make, have made, use, practice, sell, and offer
>> for sale, and/or otherwise dispose of the Original Software (or
>> portions thereof).

So if Sun wanted to close OpenSolaris some day, all they'd have to do is
stop releasing "Original Software" and whatever version they last
released couldn't be forked, because the patent grants specifically apply
to "Original Software" only.

Open is Open.

Posted Nov 10, 2005 3:55 UTC (Thu) by paulj (subscriber, #341) [Link] (15 responses)

So if Sun wanted to close OpenSolaris some day, all they'd have to do is stop releasing "Original Software" and whatever version they last released couldn't be forked, because the patent grants specifically apply to "Original Software" only.

The above is not quite true. You *can* fork OpenSolaris. The patent licence is not a patent claim of itself, merely a licence to any relevant patents (which may or may not be there). Sun don't have a track record of suing for patent infringement either.

You can fork away, Sun then could indeed sue for /modified/ files (AIUI). They still can not sue for the unmodified CDDL files, or portions thereof, to which the patent grant still applies. (The point you raised in your other post is wrong, I *think*, but could be wrong).

So if you fork, as you modify things your modifications are not exempt from patent claims by the original developer (Sun), even if you make them to files which did originally come with a licence from the Original Developer.

The problem here essentially is in an ideal world Sun would promise never to sue anyone "good" for infringing software patents, yet that is a very hard promise to make without Sun stripping itself of means to defend itself against patent claims. In the climate today, particularly in the USA, big corporations need to be able to assert patent counter-claims to protect themselves when sued. And Sun are quite familiar with being sued (got stung for $90M or so when a non-technical jury decided Sun had, in Java, infringed on patents by Kodak, of all companies.), and their executives are on record that Sun does not intend to use patents aggresively IIRC.

Further, it could dilute the patent MAD clauses in the CDDL. If Sun has broadly promised to not assert patent claims, then the 6.2 "threat" of the CDDL (patent MAD) becomes toothless.

So, yes, if you go way off on your own with OpenSolaris you also leave the realm of Suns' "patent peace" zone. You'll just have to take it on a combination of Suns' track record wrt patents and a bit of faith that Sun will not turn into a patent troll. (I think it's very unlikely).

That said, I wish too it could be phrased better. I suspect though that if it were easy to word it better, then it would have been done so originally. Do you have suggestions? :)

Open is Open.

Posted Nov 10, 2005 4:31 UTC (Thu) by cventers (guest, #31465) [Link] (14 responses)

Thank you for being fair about addressing my points. I really don't want
to write as if LWN is my platform to flame you or Sun but I feel as if my
concerns are legitimate, and I suspect many feel the same.

I think we can agree that the wording in the CDDL is questionable and
possibly open to interpretation (more so than either of us would like).

The remaining difference then, it seems, is that you trust Sun whereas
I'm not completely inclined to. Why couldn't Sun have chosen to
explicitly, and in plain english, license all of its patents to any
CDDL-licensed product derived from the original? Why would it care to
keep these patents locked up?

It's not anything to do with MAD, because Sun demonstrated (to my glee)
with OpenDocument that there are other approaches, by licensing their
relevant patents to all OpenDocument implementations except when an
adversary attempts to attack OpenDocument with their own patents.

You asked if I had suggestions. Please don't take this as a troll, and
this probably isn't what Sun wants, but my suggestion would have been to
simply use the GPL. The GPL has some big advantages. Bob Young of Red Hat
pointed one out in particular in a recent interview: the GPL is well
understood. If you license your product under the GPL and then attempt to
deploy it in a corporation, or perhaps ask the community for
participation, everyone knows that you aren't trying to sneak anything
through the back door.

In fact, it's the very reason that licenses are *hard* to understand and
interpret that makes it a big problem. Sun may have not been an abuser in
the past, but plenty of people have been burned by the phrase 'trust me'.

An HP exec publically called for the CDDL to be deprecated and replaced
with the GPL:

http://www.infoworld.com/article/05/08/09/HNhpfink_1.html

Sun responded:

>> "Instead of sniping from the outlands of participation in the open
>> source community it'd be real nice to see HP try and take a run at
>> dislodging Sun from the number one slot as contributor of code among
>> commercial companies. If that were to happen, perhaps Mr. Fink would
>> realize that there isn't one hammer for all nails and not one license
>> for all projects," Sun representative Russ Castronovosaid in a
>> statement.

Sun's behavior has been adversarial and confusing. Sun refuses to fully
open Java, and then it created the CDDL, half-opened Solaris and started
acting as if it were in the same open source league, license-wise, as
Linux. I don't think anything could be further from the truth. Linux is
an ideal example of the GPL working at its best. The code copyrights are
owned by whomever submitted the code, and since it's all submitted under
the GPL, it can never be taken away.

So if you don't have inherent trust for Sun, what obvious reasons might
Sun have for creating the CDDL? Either they didn't want anyone to be able
to mix Solaris code into Linux, or they wanted extra control over what
happens to Solaris in the "open" world, or perhaps even both. It doesn't
seem genuine, it doesn't read to me like real open source, and I don't
trust it.

Open is Open.

Posted Nov 10, 2005 8:14 UTC (Thu) by paulj (subscriber, #341) [Link] (13 responses)

Thank you for being fair about addressing my points. I really don't want to write as if LWN is my platform to flame you or Sun but I feel as if my concerns are legitimate, and I suspect many feel the same.

And same from me. This is LWN after all (and a locked article too!), we should be quite capable of lively, but reasonable, discussion. :). Thanks btw for opening my eyes a bit to other viable interpretations of some of the clauses we had discussed. I hadn't considered them (and on one point at least so far I stand corrected, thanks (patent grants extending to modified works)).

I think we can agree that the wording in the CDDL is questionable and possibly open to interpretation (more so than either of us would like).

I wouldn't say questionable. I agree it is less than ideal. However, as you can well imagine, the licencing issue had a *lot* of discussion internally at a high-level involving engineering, legal and executives. A lot of very smart people from those branches were involved in debating and thinking through the issues. I wasn't one of them (I'm a newbie junior staffer ;) ), but am quite confident there must have been good reasons for the wording as it is. My educated guess is that there wasn't a good way to word the patent grant broadly without making it very vague and diluted. Better to keep it tight and clear - if better wording is found in future, the licence can be modified. The flip-side is that you might not be able to undo over-broad patent grants if you make a mistake through vague wording. An engineer's approach to software-licencing perhaps.

The remaining difference then, it seems, is that you trust Sun whereas I'm not completely inclined to.

I do, but I can't convince you with that, other than asking you to take that in faith. I would ask though that you apply Occam's Razor sometimes, e.g. in deciding whether Sun have bad intentions in not granting broad patent rights or whether instead Sun simply wanted to keep the licence simple rather than risk unintended consequences with ill-defined wording to grant more liberal patent rights, consider whether maybe the latter explanation is the more simple and hence maybe the more likely.

Why couldn't Sun have chosen to explicitly, and in plain english, license all of its patents to any CDDL-licensed product derived from the original? Why would it care to keep these patents locked up?

Well, I really don't know to be honest. I can try to guess, from my POV as someone inclined to be quite sympathetic to Sun (and aware of the internal corporate culture). I would say firstly that there is the point above, it simply is easier to keep a licence 'tight' initially and to add more grants in later revisions if needs be, than the other way around. Also Sun considers itself responsible enough to be able to provide some stewardship over these patents and copyrights, goes without saying. Further, Sun is very much trying to devolve governance of OpenSolaris to a community consensus model. There is a OpenSolaris CAB in place, three of its five members are external to Sun. Formulating the governance process for OpenSolaris is still a work in progress, part of that process is figuring out how to specify the relationship between Sun and OpenSolaris in terms of stewardship, custodial duties, rights and responsibilities. Read the draft, it should give some flavour at least of the direction OpenSolaris might take. It's a learning process obviously though.

It's not anything to do with MAD, because Sun demonstrated (to my glee) with OpenDocument that there are other approaches, by licensing their relevant patents to all OpenDocument implementations except when an adversary attempts to attack OpenDocument with their own patents.

Hmm, but that is exactly patent MAD though. Also, ODT is a document format, not an entire Unix kernel and userland implementation. Not inconceivable there'd be different considerations.

my suggestion would have been to simply use the GPL. The GPL has some big advantages. Bob Young of Red Hat pointed one out in particular in a recent interview: the GPL is well understood.

ACK. Like I said, the GPL was a strong contender (I think one of the people involved in the opening of Solaris mentioned that in a blog shortly after the OpenSolaris launch). But, as mentioned in another post, there were reasons not to. E.g. the GPL does not allow linkage to/from closed code, an issue for drivers. Then there's the fact that the GPL does not (yet) try mitigate patent risks.

Sun may have not been an abuser in the past, but plenty of people have been burned by the phrase 'trust me'.

The main issue you've brought up though is patents, how exactly would the GPL have helped you have more trust wrt patent claims? Remember, the copyright holder doesn't have to follow the GPL, so the GPL does not stop them giving you their code under the GPL and then later asserting patent claims against you. You actually have a bit more protection from Sun suing You (in the CDDL "You" sense) with the CDDL than you would have if Sun had GPLed OpenSolaris. I find it a bit strange that this would be a problem ;). So Sun should have gone further by giving You *no* patent grants at all?

Again, I'm not trying to be "smart" or snide - I'm trying to understand where you're coming from, by getting you to expand on those parts of your position I find difficult to understand at present.

Sun responded:
>> "Instead of sniping from the outlands of participation in the open
>> source community it'd be real nice to see HP try and take a run at
>> dislodging Sun from the number one slot as contributor of code among
>> commercial companies.

So HP take PR pot-shots at Sun, and we respond. What's wrong with that exactly? They're a competitor of ours, they do their best to compete with us, we do our best to compete with them. FWIW, the point made by Sun above is not without merit. Sun hackers contribute to lots of stuff, both as part of their official Sun duties, and in their own time.

Sun's behavior has been adversarial and confusing.

Adversial, sure. Sun are a company, Sun has a duty to compete :). RedHats' executives have also taken PR pot-shots at us, and we at them. I guess that's part of what executives are supposed to do, I don't know. ;)

Confusing, I'll accept that actually (not speaking for Sun). I think Suns' "message" (to use some marketing speak) has been a bit mixed over the years. If you were kind, you might chalk that down to a company in the process of grappling with how to best adapt to new business realities of an open-source world. I think Sun has, in the last two years or so, finally decided on exactly what it needed to do, and the "message" should have been clearer since then (and stay clear).

Sun refuses to fully open Java

I'm not in that part of Sun, but again, you can bet there were a lot of discussions about it. In fairness, the specifications for the various parts of a JVM are open, and you can implement a compatible JVM if you wish. I don't really know enough about the issues here to give any useful comment though, sorry.

half-opened Solaris

Hmm, well, there's a lot to Solaris. I think Sun did mention it would take time and would be progressive, and that there'd be some bits which Sun simply could not release due to 3rd party interests, which would eventually have to be rewritten from scratch. See the CAB link above too.

So if you don't have inherent trust for Sun, what obvious reasons might Sun have for creating the CDDL? Either they didn't want anyone to be able to mix Solaris code into Linux, or they wanted extra control over what happens to Solaris in the "open" world, or perhaps even both. It doesn't seem genuine, it doesn't read to me like real open source, and I don't trust it.

Well, I hope then you're not using Mozilla in any shape or form, an even more untrustworthy licence than the CDDL! Seriously though, Sun didn't create the CDDL, they cleaned up the MPL. Mixing Solaris code into Linux? Well, they're quite different kernels, code simply is not going to be a straight-swap, nowhere close. Concepts from the Solaris kernel, those are not covered by copyright anyway. E.g. Linux has its own implementation of the Solaris slab allocator, quite a while already anyway (described in a USENIX paper by Jeff Bonwick a long time ago).

If instead you mean GNU/Linux, the userland applications, well yes of course you can swap (core utilities and libraries excepted, they tend to be kernel specific anyway). Why couldn't you? Linux and Solaris have *long* been running the same sets of applications anyway. E.g. You'd be very hard-pressed initially to tell whether you're logging into JDS on a Linux core, or on a Solaris core OS.

That's part of the beauty of Unix, that we have this independence of application from OS (not always perfectly, but better than any other system). Its open and good and allows for both strong cross-pollination and competition, in technical ideas and even in other areas like development models. Linux no doubt has learned things from Solaris, as it has from BSD and other systems. Solaris no doubt has things to learn from Linux, as Sun does from the community models built up around Linux. Indeed, one of the meetings CAB held was to investigate the models and processes these other systems used and to get hackers from those other communities to explain a bit about to them to the OpenSolaris CAB.

As a Linux user, you don't have to trust or like Sun at all, but maybe you don't have to be quite so suspicious of them. ;)

Regards,

Paul Jakma

Open is Open.

Posted Nov 10, 2005 15:25 UTC (Thu) by hppnq (guest, #14462) [Link]

E.g. the GPL does not allow linkage to/from closed code, an issue for drivers.

I think one of the reasons people do not easily put all their trust in Sun, is because of this kind of reasoning. Binary-only drivers are of course generally frowned upon in the Free Software world, but there's nothing that prevents you from getting things working: Linux has a long history with these drivers.

Obviously, things are quite a bit more complicated for a company such as Sun, but I'm quite sure that exactly this point would have convinced a lot of skeptics of Sun's good intentions. It is after all an essential aspect of Free Software.

Off-topic remark: excellent discussion, gentlemen. Thanks. ;-)

Open is Open.

Posted Nov 10, 2005 19:57 UTC (Thu) by bojan (subscriber, #14302) [Link] (11 responses)

> But, as mentioned in another post, there were reasons not to. E.g. the GPL does not allow linkage to/from closed code, an issue for drivers.

I'm sorry, but this is just nonsense. People provide exceptions for this kind of thing all the time. Linux kernel already does. Many userland apps also do. Given that Sun is the copyright holder here, Sun could have done the same.

Now, when it comes to the remaining issue of patents, I IBM don't seem to have any problems with contributing to a GPL-ed piece of software (Linux kernel), although they have a truckload of patents (something like 10 times more than Sun). So, Sun could have enforced their patents selectively, if they so wanted. Remember, to have a patent means to have the right to "exclude people from making, selling, [...]". If you let someone use your patents, that doesn't mean you cannot enforce them against people that cause you trouble.

My bet: CDDL was chosen precisely _because_ it was _incompatible_ with the GPL.

Open is Open.

Posted Nov 10, 2005 22:47 UTC (Thu) by paulj (subscriber, #341) [Link] (10 responses)

People provide exceptions for this kind of thing all the time.

Grand, but be clear: then we're no longer talking about a GPL licence.

Open is Open.

Posted Nov 10, 2005 23:30 UTC (Thu) by bojan (subscriber, #14302) [Link] (9 responses)

> Grand, but be clear: then we're no longer talking about a GPL licence.

Of course we are. For instance, here is a part of the text from such an exception (see: http://www.gnu.org/cgi-bin/license-quiz.cgi for an example use, as written by FSF):

-----------------------------------------
If you modify this file, you may extend these exceptions to your version of the file, but you are not obligated to do so. If you do not wish to do so, delete one or both of these exception statements from your version.
-----------------------------------------

So, the recipient of the software can relicense the software under the pure GPL if he/she so wishes. But the point remains - this software can be included, without any problems, in _any_ software licenced under the _pure_ GPL. In effect, it is licensed under the GPL.

PS. This text is from the copyright statement of my own software, so I didn't just pull it out of thin air.

Open is Open.

Posted Nov 11, 2005 9:22 UTC (Fri) by paulj (subscriber, #341) [Link] (8 responses)

Of course we are.

No.. You're talking about GPL + exceptions. Not at all disputing the validity of exceptions (I think I mentioned in another post the ability for GPL authors to give themselves exceptions so they could link to CDDL code).

However it is still not quite the GPL, it's "GPL + X". It's disingenuous to reply to me and claim that Sun could have chosen the GPL, when really you mean "GPL + X". Further, the "+ X" part would be the "specify patent pool and MAD on litigation", which would apply additional restrictions on patent rights not specified in the GPL. Hence that would make this "GPL + X" licence incompatible with the GPL.

So your suggestion essentially is no different to Sun having used the CDDL.

Open is Open.

Posted Nov 12, 2005 3:50 UTC (Sat) by bojan (subscriber, #14302) [Link] (7 responses)

> So your suggestion essentially is no different to Sun having used the CDDL.

On the contrary. My suggestion is _qualitatively_ different to what Sun have done. It is _illegal_ to include CDDL licensed code with the GPL licensed software. It is perfectly legal to do so with the code licensed under the GPL + Exception. And that is because GPL + Exception _is_ essentially GPL, for people that choose not to pass on the exceptions. Nothing like that is possible with the CDDL.

As for patents, I addressed that before to a degree, but I'll elaborate. Bottom line is this: Sun are unwilling to guarantee (in any legal sense, like IBM) that they won't sue open source developers over patents Sun holds. In other words, I can get in trouble for using patents from OpenSolaris in my open source software. I cannot get in trouble for using IBM's patents they promised not to sue over. So, I think this behaviour on behalf of Sun is more disingenuous that anything I wrote here, particularly coming from the company that supposedly not only understands but also "invented" open source (luckily, they didn't patent that invention :-).

We can all pretend here, but it is quite clear that CDDL was designed and chosen _precisely_ because it isn't compatible with the GPL, in order to prevent Solaris technology from going to Linux (which is likely, given the size of Linux v. OpenSolaris development communities). Ditto the patent promises, which only apply to CDDL licensed software. I just wish Sun were open about it and told us so in the first place.

I did watch the launch of Galaxy servers where Red Hat (and by extension Linux) supposedly became Sun's friend again. I just wonder why that was? Could it possibly have anything to do with the fact that most Sun AMD based servers run an operating system other than Solaris? Or was it a sign or "real" friendship? Hmm, I'm inclined to go with the former, but that's just my cynical self ;-)

Open is Open.

Posted Nov 12, 2005 15:41 UTC (Sat) by paulj (subscriber, #341) [Link] (6 responses)

On the contrary. My suggestion is _qualitatively_ different to what Sun have done. It is _illegal_ to include CDDL licensed code with the GPL licensed software. It is perfectly legal to do so with the code licensed under the GPL + Exception. And that is because GPL + Exception _is_ essentially GPL, for people that choose not to pass on the exceptions. Nothing like that is possible with the CDDL

This is about the patent-pool, patent MAD clauses of the CDDL. The problem is that if you add those to the GPL, as you suggest, then they are not exceptions but additional restrictions. Additional restrictions would make this theoretical "Sun GPL" incompatible with the GPL anyway. Authors of GPL software *can* make use of CDDL patent grants, just by giving themselves to link to the CDDL software to which the grant applies.

Bottom line is this: Sun are unwilling to guarantee (in any legal sense, like IBM) that they won't sue open source developers over patents Sun holds.

Ha, ha. Go and look at the patents IBM donated to the Linux *kernel* (not to opensource in general btw). I'm sure the linux kernel is far safer now that IBM won't sue it for things like patents on gel-packs, and such.

Sun executives BTW are on record that Suns' interest in patents is primarily defensive. And the patents Sun has given Free Software (which is what code under the CDDL is, according to *RMS*) are those relevant to the code at hand. Unlike IBMs publicity stunt.

We can all pretend here, but it is quite clear that CDDL was designed and chosen _precisely_ because it isn't compatible with the GPL, in order to prevent Solaris technology from going to Linux (which is likely, given the size of Linux v. OpenSolaris development communities).

Your powers of mind-reading are obviously far greater than mine. I don't see how that is clear at all. I see it as a possibility, a very unlikely possibility given various other factors, as I've already explained. Eg. that the CDDL simply has restrictions in it which can never be GPLv2 compatible (but maybe could be GPLv3 compatible). Or e.g. that OpenSolaris and Linux kernel code simply is *way* too different for code to transfer across anyway. As for OpenSolaris apps, you can port those across no problem, CDDL licence has no problem with that. If you want to implement DTrace on some other kernel, you're free to use the OpenSolaris Dtrace userland tools. (Someone is working on this for FreeBSD btw).

Are there reasonable things which you could criticise Sun for? Sure. Are there omissions or deficiencies in the CDDL, some claim there are (see other threads). I'm pretty sure Sun and the OpenSolaris community would like to hear them (join the lists), they might even be fixable if valid.

But if your arguments start to reduce down to mind-reading, I'll have to bow out I'm afraid.

Open is Open.

Posted Nov 13, 2005 3:03 UTC (Sun) by bojan (subscriber, #14302) [Link] (4 responses)

> if you add those to the GPL, as you suggest

I never suggested _that_. I said that Sun could have licensed with GPL + Exception for _linking_ with other software. I addressed patents before (twice) and I won't do it again.

> donated to the Linux *kernel* (not to opensource in general btw)

Please get your facts straight. IBM _explicitly_ promised not to sue _any_ open source project over about 500 patents they have (applies to all OSI approved licences). Not just the kernel.

> Sun executives BTW are on record that Suns' interest in patents is primarily defensive.

Which is legally meaningless.

> Unlike IBMs publicity stunt.

See above.

> Your powers of mind-reading are obviously far greater than mine.

I studied under Yoda himself ;-)

Open is Open.

Posted Nov 13, 2005 6:42 UTC (Sun) by paulj (subscriber, #341) [Link] (3 responses)

I never suggested _that_. I said that Sun could have licensed with GPL + Exception for _linking_ with other software.

Ok, fair enough. However, all my arguments here before about why the GPL would not have sufficed have revolved mostly around the patent pool+MAD clauses of the CDDL, linkage was just another example. Yes, the linkage thing possibly could be dealt with via exceptions. My point regarding the patent clauses stands though.

Please get your facts straight. IBM _explicitly_ promised not to sue _any_ open source project over about 500 patents they have (applies to all OSI approved licences). Not just the kernel.

Ah, fair enough. I must have been confused. However, go and *look* at the list of patents. Quite a few have absolutely no relevance to software (CPU, token ring, ATM, etc. hardware patents), others do not have much value (though, that's probably more a comment on the state of the patent system) and at least two of the patents don't have anything to do with computers (tamper proof screws, DNA reading assembly).

A unilateral patent grant still does not do anything to try setup a patent pool and patent "peace" around a software project, as the CDDL does. As I said in another post, be careful what you criticise the CDDL for, when there's a fair chance the GPLv3 will try do similar tomorrow (Eben Moglen has stated patent threats are something the GPLv3 may try address). The CDDL's clauses are intended to encourage the build-up of a pool of grants to a project on all relevent patents held by the initial and contributing developers. That's a worthwhile goal IMHO.

Which is legally meaningless.

IBM have *many* more patents than those 500. They can almost certainly dig up patents from their portfolio to sue almost anyone they wanted to, regardless of those 500. So their grant is not *that* meaningful either. So lets be honest: The primary reasons big corporates involved fruitfully in Free Software, like IBM and Sun, neither have nor likely will (in the short to mid-term at least) sue any Free Software developers or projects is because they have absolutely nothing to gain from it and a *lot* to lose.

Anyway, did Sun want to do something different than the GPL? Yes, I've listed some *good* reasons why. Disagreeing with the logic of those reasons is fine. However clinging to the view that Sun chose the CDDL out of sheer perverseness (reasoning I can never really refute, since it hinges implicitely on Sun having hidden motives), when it isn't that hard to come up with more rational reasons is perverse in itself, if not sheer paranoia. Particularly when the objectionable result of this "hidden motive" theory, incompatibility with the GPL, lies *not* in the CDDL but in the GPL and can easily be fixed in the next revision of the GPL, which is due in the next year or two. Further, any GPLv2 copyright holder can *today* avail of CDDL code by simply giving their own GPL code an exception. If you have suggestions on improving the CDDL, you are welcome to discuss them on the opensolaris-discuss list, the CDDL *can* be up-revved too.

I studied under Yoda himself ;-)

:^)

Open is Open.

Posted Nov 13, 2005 8:31 UTC (Sun) by cventers (guest, #31465) [Link]

>> A unilateral patent grant still does not do anything to try setup a
>> patent pool and patent "peace" around a software project, as the CDDL
>> does.

Patent "peace"? The points I pointed out earlier leave me feeling as if
there is much less patent "peace" from Sun than IBM, especially given
that not only did IBM open their patents in a non-discriminatory fashion,
but are participating in other patent projects to defend Linux.

>> As I said in another post, be careful what you criticise the CDDL for,
>> when there's a fair chance the GPLv3 will try do similar tomorrow
>> (Eben Moglen has stated patent threats are something the GPLv3 may try
>> address).

Anything at this point is pure speculation. There are a million ways to
address the patent issue. But what I will say is that given Stallman's
philosophy, if patent protection is somehow offered by GPLv3, it will be
unilaterally offered to all products derived from the original, something
that the CDDL *does not* do.

>> They can almost certainly dig up patents from their portfolio to sue
>> almost anyone they wanted to, regardless of those 500. So their grant
>> is not *that* meaningful either. So lets be honest: The primary
>> reasons big corporates involved fruitfully in Free Software, like IBM
>> and Sun, neither have nor likely will (in the short to mid-term at
>> least) sue any Free Software developers or projects is because they
>> have absolutely nothing to gain from it and a *lot* to lose.

I think IBM's grant, combined with their actions (such as participation
in the creation of the Open Inventions Alliance):

http://news.zdnet.com/2100-3513_22-5943781.html

Means that they are demonstrating much more *tangible* friendship and
support of open source. Sun's licensing of their patents to OpenSolaris,
and perhaps the original version only, imply that Sun doesn't care one
lick beyond their own projects. They could freely license their patents
to any OSI-approved license as IBM has done, and their patents would
*still* be capable of instigating MAD against any party that threatens
them.

>> Particularly when the objectionable result of this "hidden motive"
>> theory, incompatibility with the GPL, lies *not* in the CDDL but in
>> the GPL and can easily be fixed in the next revision of the GPL, which
>> is due in the next year or two.

I don't think any motives are hidden. I don't think Sun is about to go
after open source developers with their patents, be it Linux or
OpenSolaris - but it's clear to me that they didn't care to *yield* their
(legal) right to do so. They're not diving into this open source thing -
they're sticking a toe in and checking to see how the water feels, and
they'll probably be shivering all the way until the market finally shoves
them in.

Your assertion that the problem is with the GPL troubles me as well. The
GPL was intentionally designed to require that derived works be GPL as
well because that is simply the best way to ensure that no one can tamper
with the freedom of the system. No one can confuse the issue by
introducing new licenses. Given the GPL's massive adoption rate, and the
feelings of much of the community about the CDDL (including FSF's
suggestion to not use it), I think people understand this.

>> If you have suggestions on improving the CDDL, you are welcome to
>> discuss them on the opensolaris-discuss list, the CDDL *can* be
>> up-revved too.

It may be considered armchair punditry, but we simply don't have to. We
already have the GPL, LGPL, BSD license, etc, all of which suit our needs
well.

The CDDL was written by a business with business concepts and needs in
mind. The GPL was written by a philosopher with the good of the people in
mind. If Sun wants to volunteer to become a better player in the open
source field, they're free to do so, but we should *not* have to drag
them into it kicking and screaming. IBM along with a slew of other huge
companies walked right into the party.

Taking this a bit further (and please don't consider this to be a troll,
because it's not intended to be), but myself and I assume a lot of other
GNU/Linux users with servers, desktops, laptops, phones, TiVOs, etc
loaded full of GPL software, don't feel like we *need* Sun. Sun will join
the game if and when they choose. Speaking personally, if Sun closed up
shop tomorrow and started making microwaves and toaster ovens, I wouldn't
even blink. Sun might be liked by more if they didn't seem to carry such
an ego, implying in the undertones of their behavior that they are a
necessity to the computing community.

Open is Open.

Posted Nov 13, 2005 19:26 UTC (Sun) by bojan (subscriber, #14302) [Link] (1 responses)

> if not sheer paranoia

I actually agree with you here (and I'm not saying this to be sarcastic). I don't trust Sun after all the flip-flopping they've gone through in regards to open source and Linux. And it's all right there in Jonathan's blog (if he doesn't speak for the company, I don't know who does). All this change of heart makes me nervous. Others, particularly IBM (note: I don't work for IBM, don't own shares of it, never purchased a computer from them and where I work the whole organisation has zero IBM equipment) have been more straightforward, while still quite openly supporting other (proprietary) things and generally hedging their bets, but without "change of heart" every second day.

I would say that possibly the biggest trust busting factor of all is Java for me. If there ever was a vibrant open source effort, it is around Java (e.g. project Jakarta). Some of the best Java stuff out there is open source. And yet, I cannot find any _valid_ reason for Java to be closed source. I know Sun exectives quoted compatibility, but that is just so easy to achieve through tradmarks, that is completely unbelievable. Unfortunately there is no scientific way (yet) to measure what would happen to the platform if it were open sourced, but I think that mere fact that it would get included as a standard part of evey Linux distribution speaks for itself, not to mention all the JDK bugs that have been outstanding for months (in cases years) being fixed quickly. Why Sun doesn't want that to happen, I don't understand.

Open is Open.

Posted Nov 18, 2005 3:37 UTC (Fri) by robilad (guest, #27163) [Link]

As long as Sun makes money on licensing their proprietary Java implementation, and the whole technology collage around it, there is no business reason for them to release their implementation's source code under an open source license. It's a proprietary software vendor, it's not a charity. Pending huge management changes (and there would be quite a few changes necessary in the Java division afaict, including booting Graham Hamilton), it's not going to happen.

cheers,
dalibor topic

Open is Open.

Posted Nov 13, 2005 6:23 UTC (Sun) by bojan (subscriber, #14302) [Link]

> patents on gel-packs,

Just one more comment here, if I may. You seem to be reading too much of Jonathan's blog and therefore mocking all of those patents. The patents IBM pledged to OSS are here:

http://www.ibm.com/ibm/licensing/patents/pledgedpatents.pdf

To the majority of open source developers, if only one of those patents is something an OSS developer would ever use, IBM's pledge is just about infinitely better than whatever Sun promised in relation to OpenSolaris. That's because Sun's promise relates to (mosly likely unmodified) OpenSolaris code only.

Open is Open.

Posted Nov 9, 2005 22:15 UTC (Wed) by bojan (subscriber, #14302) [Link]

> The source to solaris is Open!

Well, except for the parts that aren't. The tarball is called:

opensolaris-closed-bins-DATE.PLATFORM.tar.bz2

Not so open...

Debian and Nexenta collide

Posted Nov 9, 2005 23:06 UTC (Wed) by bojan (subscriber, #14302) [Link] (4 responses)

> bizarro-world of half-open technology?

Probably the best description of Sun and their open source strategies I've seen in a while.

<rant>
On one hand, they contributed large chunks of LGPL code, like OOo. On the other, they have to do it "the Sun way" (or as one of the old Solaris admins that I work with would say: "In their infinite wisdom, Sun decided to <insert bizarro thing here>"). So, most of the stuff that exists in the FOSS world (e.g. font rendering) had to be duplicated in OOo, you have to install gazillion things to build the darn thing and then it doesn't build cleanly on x86_64 (as we learned from recent discussions here on LWN) etc. And then there is Java (no, not just the OOo pullution with it). The big lovers of open source, which have requests sitting in their bug system to produce x86_64 plugin for months, don't want to give the code to FOSS developers to do it for them. And, they are letting GCJ/classpath people waste their time on an alternative implementation, when there is no rational explanation for keeping Java proprietary. Then there is Solaris, or should I say "Open"Solaris. It's half open, released under "spanner in the works" licence in order to prevent things from jumping into their (now) long time enemy Linux.
</rant>

There is no doubt that Sun did some great things for open source, but if there ever was a company with a confusing (open source) strategy, it's Sun.

Debian and Nexenta collide

Posted Nov 10, 2005 8:07 UTC (Thu) by mjw (subscriber, #16740) [Link] (3 responses)

And then there is Java (no, not just the OOo pullution with it). The big lovers of open source, which have requests sitting in their bug system to produce x86_64 plugin for months, don't want to give the code to FOSS developers to do it for them. And, they are letting GCJ/classpath people waste their time on an alternative implementation, when there is no rational explanation for keeping Java proprietary.
I now this is part of your <rant> block, but rest assured that we are not wasting our time :) As you can see from our bi-monthly releases or from reading planet.classpath.org we are having a lot of fun and making a lot of progress providing an technical (and in our eyes better and more flexible) alternative for the core class libraries and java langauge compilers and tools then any proprietary offer. Through the integration of GNU Classpath/GCJ into GCC we have (indirectly) improved a lot of the fundementals of the GNU system, and through alternative systems like IKVM.NET we bring the java programming language and all programs written in it in realms that nobody could imagine before. Sure it would be nice if Sun would help out and released (parts of) of their proprietary implementation under some GPL-compatible terms. But we have such critical mass now that the projects stand on their own and provide people with real Free Software alternatives for the proprietary java platform if they choose to use such a language.

Classpath and GCJ

Posted Nov 10, 2005 16:30 UTC (Thu) by GreyWizard (guest, #1026) [Link]

Well done. I've been impressed again and again by the progress of Classpath and GCJ. When these projects are complete I suspect Sun will discover that even a paranoid proprietary license won't allow them to keep control of the Java community.

Tricky wording

Posted Nov 10, 2005 19:46 UTC (Thu) by bojan (subscriber, #14302) [Link] (1 responses)

> not wasting our time

Just to make myself clear, I think neither that GCJ/classpath is a waste of time, nor do I think you guys are wasting time. I think that Sun are being irresponsible by letting you guys spend preciuos time on this work, while they could have (and should have) released their JDK under the GPL long, long time ago. If they did that, Java would probably be at version 5.0 (no, _real_ version 5.0). And, every Linux distro would be able to ship a _fully_ _functional_ JDK, with absolutely no strings attached. How much more FOSS software would have been written in Java if that were the case? Truckloads. Simply because distributors would be able to _rely_ on the fact that it is always there.

Believe me, I admire what has been done with GCJ and classpath. But I hate to see a wasted effort in the FOSS world when there is no good reason for it. And, all because of the company that supposedly "invented" open source, but cannot understand the basic concepts around it.

Tricky wording

Posted Nov 18, 2005 3:53 UTC (Fri) by robilad (guest, #27163) [Link]

Sure, there could be a lot of great things Sun could have done in the past 10 years to advance the state of the art in that area. They've made it clear that they won't open up their implemetation, come what may, though, for whatever reasons they have. It's their code, so it's their problem. ;)

So the solution for the rest of us is to write a better implementation that removes the need to deal with the bizarre problems of the proprietary one. Worked for gcc, Linux, and a bunch of other things. It will work again, it just takes a bit of time to get done, like any good thing.

cheers,
dalibor topic

Debian and Nexenta collide

Posted Nov 9, 2005 18:39 UTC (Wed) by kirkengaard (guest, #15022) [Link] (2 responses)

Once they fully comply with the GPL, it will be a contribution to the free software world. And it will be compliant with the copyrights involved in GNU software. But for the community to serve notice to violators of its copyrights demanding that they comply with the licensing terms of the software used is not only right, but mandatory. The notice of violation is due diligence in maintaining copyright; the offer of help in achieving compliance is optional.

Issuing binary-only software from GPL sources in no way constitutes a free software contribution. It constitutes a breach of the duty of care on the part of the developers of Nexenta.

As long as they continue to remedy the lack of wisdom that caused them to start out in violation of the GPL, their work will become a free software contribution. I suspect that the reason that few people are complaining about the CDDL side is that this was not a Debian effort to integrate the Solaris kernel.

due diligence is only for trademarks

Posted Nov 9, 2005 22:21 UTC (Wed) by stevenj (guest, #421) [Link] (1 responses)

Um, I think that the requirement that you pursue violations is only for trademark law.

This is why orphaned works are such a problem in current copyright law.

due diligence is only for trademarks

Posted Nov 10, 2005 4:31 UTC (Thu) by kirkengaard (guest, #15022) [Link]

point, my bad.

Debian and Nexenta collide

Posted Nov 9, 2005 20:18 UTC (Wed) by Duncan (guest, #6647) [Link] (2 responses)

When I read the GPL complaint, the syslibs exemption was my first
reaction. The "shipping with" clause didn't occur to me. After I read a
bit further, however, I could see the legal argument, altho the irony of a
GPL restriction allowing allowing use on a closed system but not on a
system composed of Free software, certainly bites!

Then the split solution occurred to me, just before I got to that part of
the article. After reading the full article, I still don't see the
problem. Certainly, it shouldn't be difficult to script a solution in the
installer that downloads (if necessary) and unpacks dpkg and the like in
place, making them part of a core-package that provides the necessary
dependencies and pre-stuffs the dependency database as necessary. I just
don't quite see the problem, at least not as anything at a level
comparable to setting up an entire OS, something they're already taking
on. It's just one more thing an installer must manage, on top of the
others it already manages. If they're using the Debian installer and
that's GPLed, creating one's own installer isn't a problem others haven't
tackled already, either.

Duncan

Debian and Nexenta collide

Posted Nov 10, 2005 4:14 UTC (Thu) by AJWM (guest, #15888) [Link] (1 responses)

That's pretty much the way un-open Solaris has handled it. Any GNU or other GPL'd utilities are shipped on an entirely separate disc, so the OS components do not "accompan[y] the [GPLd] executable[s]". They are packaged such that Sun's own install mechanisms can be used to install them, avoiding the dpkg bootstrap issue.

Debian and Nexenta collide

Posted Nov 11, 2005 4:38 UTC (Fri) by kimoto (subscriber, #5244) [Link]

What about when I turned on some new hardware from Sun and found some incarnation of of Solaris 8 on it, including working versions of some GPL'd programs like gzip and bash?

Debian and Nexenta collide

Posted Nov 9, 2005 21:00 UTC (Wed) by paulj (subscriber, #341) [Link] (29 responses)

The Nexenta project uses, along with the Solaris kernel, a number of user-space libraries (including the core C library) from Solaris. These libraries, being licensed under the CDDL, are not compatible with GPL-licensed applications.

The argument is a bit strained from a copyright POV. You have pre-existing GPL binaries, making use of fairly standard interfaces (C99, POSIX, SUS), such that they can be built against and link to any library implementing those interfaces. Solaris libc happens to be one of those. Precedent in US copyright law at least does not allow interfaces to be copyrighted, it's one of the (many) reasons why the SCO lawsuit is quite bogus.

The GPL can only reach as far as copyright law allows. Copyright affects implementation (well, expression, but in software ~= implementation), nothing else. If an interface is well enough defined such that it has been implemented seperately across many systems (glibc, BSD libc, Solaris libc, etc.) and thousands of applications are portable between them, then it's possibly a boundary for the reach of a copyright, and hence the GPL.

From a practical perspective, that clause of the GPL is there to prevent people 'hiding' or 'closing' functionality out of GPL applications, putting them in shared libraries instead. (Without a clear, portable, stable interface it may not be enough of a boundary to prevent copyright reaching across though - though, you'd need a judge and jury to decide for the specifics of each case). No code of these GPL applications is being "hidden" here at all. Nor even of the underlying libraries.

From an ethical perspective, the Solaris libc is openly available and modifiable under the CDDL licence. The CDDL licence is OSI approved and even RMS has no problem with it (calling it "an ethical licence") other than that it's not GPL (or more to point, GPL compatible).

Could Sun relicence OpenSolaris libc under a GPL compatible licence? Well, theoretically yes. However, the GPL had been considered *before* and not adopted as the licence for OpenSolaris (GPL was a /very/ strong candidate). The reasons for Sun creating the CDDL, rather than adopting the GPL, remain in place. The software world today faces issues which were not a big concern when the GPLv2 was originally drafted, issues which Sun did not wish to leave unaddressed. The Free Software people too recognise that the GPLv2 needs updating for the "modern" world, hence why the GPLv3 is being drafted. (E.g. judging by reports, one of the things GPLv3 is to try address is the issue of software patents. Note that some of the key GPL incompatibilities of the CDDL relate to its software-pooling and patent MAD clauses. Provisions which would themselves make sense for the GPLv2).

Given the GPLv3 is still being drafted and would, on introduction, be easily applied to nearly all GPL software, it might be an idea to try make the GPLv3 compatible with the CDDL - problem solved.

As for Nexenta, well, they have been quick to correct mistakes (e.g. not offering package sources with the iso's.). There's no indication that they've done anything other than overlook some important requirements in their enthusiasm to get a GNU/Solaris built and distributed. Requirements they quickly fulfilled. (I help maintain a Free Software project, and I've done the same thing in the past: put out RPMs and completely forgot about SRPMs till someone asked).

Paul Jakma, Free Software associate, Sun employee.

Speaking for myself (Not that I have *any* authority to speak for either of those organisations ;) ).

Debian and Nexenta collide

Posted Nov 9, 2005 22:02 UTC (Wed) by kune (guest, #172) [Link]

Sun did choose a licence, which is not GPL compatible. Solaris kernel code can't be used in the Linux kernel and vice versa. This was the choice of Sun and not the choice of the community. Asking now the community to solve the issue that Sun created, is a little bit strange.Sometimes it appears, that it is quite difficult for an organisation named Sun to understand, that they are not the center of the world.

Debian and Nexenta collide

Posted Nov 9, 2005 22:26 UTC (Wed) by cventers (guest, #31465) [Link] (10 responses)

Richard Stallman called the CDDL "basically ethical":

http://www.cuddletech.com/blog/pivot/entry.php?id=306

Further up in this conversation I mention my objections to Sun's language
designed to hold their patents close to home in the CDDL, and Richard
Stallman sees this too:

http://trends.newsforge.com/article.pl?sid=05/01/31/13102...

Debian and Nexenta collide

Posted Nov 10, 2005 0:47 UTC (Thu) by paulj (subscriber, #341) [Link] (9 responses)

Actually, he did more than call it "ethical", he said the CDDL "is a free software licence" (with the reservations i mentioned regarding the GPL incompatibilities). RMS is not one to call something "free" lightly.

Regarding patents, yes, they're confined to the CDDL code implementing the patent. Quite obviously the CDDL can not give liberal patent grants if it wishes to be able to specify patent MAD and a patent commons amongst stakeholders (users/developers) in a CDDL project. Note that the patent grants implied by the CDDL *are* available to GPL code, indeed code under any licence. It is the *GPL* which restricts itself, simply because it is quite old now and doesn't deal with the question in any way except to say essentially the patents must be licenced unconditionally to everyone, royalty-free. Note that this is obviously *not* how the GPL licences software copyright (you must follow the restrictions of the GPL), and it's quite probably *not* how the GPLv3 will approach software patents. In other words, watch out: don't today condemn the CDDL for trying to use patents as a lever in the interests of Free Software, as the GPLv2 already does with copyright, and then tomorrow applaud the GPLv3 when it may well try very similar tricks[1].

A really simple way forward would be for the GPLv3 to recognise the validity of the CDDL patent-pooling/MAD clauses and allow GPLv3 applications to avail of any patent grants in CDDL code, by allowing GPLv3 code to link to such CDDL code. There is *nothing* in the CDDL to stop the FSF, or anyone else, doing so (as long you follow the CDDL wrt to the CDDL /files/, and any derivatives thereof).

I work for Sun, so I guess I'm tainted and part of the Sun conspiracy, so don't believe me - believe RMS.

1. This is pure speculation on my part, but it has been reported that tackling software patents is on the agenda for the GPLv3 and I personally think the CDDL patent clauses would be quite suitable for the next revision of the GPL.

Debian and Nexenta collide

Posted Nov 10, 2005 1:11 UTC (Thu) by cventers (guest, #31465) [Link] (8 responses)

Did you read the article Stallman wrote? The one called "Sun's no-op
announcement" that referred specifically to the patent grants in
question? Unless I'm drunk, or he's totally wrong, then Sun's CDDL patent
grants extend to nothing except Sun's own CDDL-released software.

>> So what has really happened here? Reading the announcement clearly, I
>> think that it doesn't announce anything at all. It simply describes,
>> in a different and grandiose way, the previously announced release of
>> the Solaris source code as free software under Sun's idiosyncratic
>> license, the CDDL. Outside Solaris, few or no free software packages
>> use that license--and Sun has not said it won't sue us for
>> implementing the same techniques in our own free software.

Patent grants implied by the CDDL are *certainly* not available to GPL
code. Read OpenSolaris.org's comments:

>> The CDDL provides an explicit patent license for code released under
>> the license. This means that you can use, modify, and redistribute
>> code released under CDDL without worrying about any patents that the
>> contributors of the code (including Sun) might have on the contributed
>> technology. The license also includes a provision to discourage patent
>> litigation against developers, by revoking the rights to the code for
>> anyone initiating a patent claim against a developer regarding code
>> they have contributed.

Found on ZDNet:

>> Perhaps the most interesting among the list of commentators is IBM's
>> vice president of standards, Bob Sutor, who in his blog entry,
>> commended Sun for embracing the spirit of open source, but then added
>> "As far as I understand it, Sun is not pledging these patents for use
>> in any open source project as [IBM] did."

So how can you say:

>> Note that the patent grants implied by the CDDL *are* available to GPL
>> code, indeed code under any licence. It is the *GPL* which restricts
>> itself, simply because it is quite old now and doesn't deal with the
>> question in any way except to say essentially the patents must be
>> licenced unconditionally to everyone, royalty-free.

Please clarify! Better yet, please tell your bosses at Sun to clarify in
public!

Debian and Nexenta collide

Posted Nov 10, 2005 2:29 UTC (Thu) by paulj (subscriber, #341) [Link] (7 responses)

Did you read the article Stallman wrote? The one called "Sun's no-op announcement" that referred specifically to the patent grants in question? Unless I'm drunk, or he's totally wrong, then Sun's CDDL patent grants extend to nothing except Suns' own CDDL-released software.

I did, and yes the grants extend only to the CDDL-released software concerned. However, my point was that the CDDL allows any software to make use of it - that includes GPL software. The only thing the GPL software author has to do is grant herself an exception to link her own code to the CDDL code implementing the patent, and hey presto, she's got a patent licence (by using the CDDL code).

RMS had a concern that Suns' grant was not royalty-free, non-exclusive (AIUI) and hence does not extend to GPL software. The thing though is that Sun can not have the CDDL setup a Free Software patent pool if it also grants such liberal rights (you could avoid the pooling conditions of the CDDL by simply availing of the RFNE grant then). That the GPLv2 is old, needs updating and doesn't provide for patent-pooling is hardly a drawback of the CDDL ;).

Finally, there's a simple solution, the GPLv3 can just avail itself of that pool, by allowing linkage or inclusion of CDDL code (ie, all the GPLv3 really has to do is to just recognise the CDDL as compatible for purposes of GPL - no more).

As both an FSF supporter and a Sun bod, I very much hope the GPLv3 will do that.

Debian and Nexenta collide

Posted Nov 10, 2005 3:18 UTC (Thu) by cventers (guest, #31465) [Link] (6 responses)

As of yet I've resisted letting your affiliation with Sun have any impact
on my replies. What I will say is that your attitude...

>> That the GPLv2 is old, needs updating and doesn't provide for
>> patent-pooling is hardly a drawback of the CDDL ;).

...is typical of Sun Microsystems and is the reason why I suspect you
won't ever be real open-source competition for Linux. Sun isn't the god
that it thinks it is.

Your assertion that GPL software could link CDDL software and enjoy its
patent is specifically identified in the CDDL as false:

>> (d) Notwithstanding Section 2.1(b) above, no patent license is
>> granted: (1) for code that You delete from the Original Software, or
>> (2) for infringements caused by: (i) the modification of the Original
>> Software, or (ii) the combination of the Original Software with other
>> software or devices.

I'll take the GPL over the CDDL any day. Thankfully, I suspect that the
silent majority agrees with me.

Debian and Nexenta collide

Posted Nov 10, 2005 4:41 UTC (Thu) by paulj (subscriber, #341) [Link] (5 responses)

>> That the GPLv2 is old, needs updating and doesn't provide for
>> patent-pooling is hardly a drawback of the CDDL ;).

...is typical of Sun Microsystems and is the reason why I suspect you won't ever be real open-source competition for Linux. Sun isn't the god that it thinks it is.

Odd, the above is a view held by the FSF - nowt to do with Sun. One of the greatest challenges today is what to do about the climate of patent-trolls, particularly in software. The CDDL attempts to do something about it, the GPLv3 almost certainly will attempt to introduce wording to tackle patent-risks too. As an FSF supporter, I certainly hope so anyway. The above was not intended a comment as a Sun employee, more with my Joe Bloggs (FSF supporter) hat on if anything. Sorry if came across as snide.

Your assertion that GPL software could link CDDL software and enjoy its patent is specifically identified in the CDDL as false:

"(d) Notwithstanding Section 2.1(b) above, no patent license is granted: (1) for code that You delete from the Original Software, or (2) for infringements caused by: (i) the modification of the Original Software, or (ii) the combination of the Original Software with other software or devices."

I don't actually know if my assertion is false or not. I *think* is true, but my assertion carries no weight - hope that's obvious. However, with respect to this 2.1d clause, note the "Notwithstanding", hence clearly 2.1d does not trump 2.1b. 2.1d therefore, by my layman's interpretation, is there to state the obvious:

"You can't infringe on Suns' patents in your own code and claim you have a patent licence through 2b simply by linking to, or wedging your infringing code in, with Suns' CDDL code".

In the latter case, you'd have to CDDL your code anyway, or breach the copyright. Note that even if you infringe on a patent of Sun, that still does not mean Sun are going to sue you, it is entirely at the patent holders discretion whether to sue or not. There are lots of things that can be done to resolve infringement other than suing people, e.g. you may be contributing to OpenSolaris, it's quite unlikely Sun would then sue you. Sun does not have a record of running around suing people, further if it's quite inconsequential Sun can easily just ignore it, at its discretion. Patent law is not like trademark law, your claims are not affected by ignoring infringement (AIUI).

(Any large corporation which does R&D almost certainly already has lots of patents in its portfolio which are being infringed here and there. By and large, corporations don't care as long as they're not sued themselves - only then does the portfolio come out ;). That is the current climate, for better or worse.)

I'll take the GPL over the CDDL any day. Thankfully, I suspect that the silent majority agrees with me.

You can use what you want. Sun uses the GPL too. Most of my work to date at Sun has been on GPLed software. Ditto for quite a few other Sun engineers. As for the CDDL, its fundamentals are fine - RMS has stated this too.

Whether it would have been possible to find the wording to allow the patent grants to follow modified works, I don't know. From a reply of yours in another thread, you obviously think that wording should have been put there. I'd agree in principle, but have doubts about whether it could be done without resulting in a vague and unclear licence.

Anyway, it's late, plus I still have to read the weekly edition! Night.

Notwithstanding

Posted Nov 10, 2005 6:36 UTC (Thu) by ncm (guest, #165) [Link] (4 responses)

"Notwithstanding" doesn't mean what you seem to think it means. Better look it up.

Notwithstanding

Posted Nov 10, 2005 8:22 UTC (Thu) by paulj (subscriber, #341) [Link] (3 responses)

Strangely enough, I did look it up to make sure it meant what I've always thought it meant and (AFAICT) it does. Care to expand?

Notwithstanding

Posted Nov 10, 2005 16:44 UTC (Thu) by vmole (guest, #111) [Link] (1 responses)

It means that 2.1(d) completely overrides (or perhaps "limits") subsection 2.1(b). That means that the patent grants do NOT apply to any modified code, or code that incorporates CDDL code in new ways. That may not be the intent, but that's what it says.

Notwithstanding

Posted Nov 10, 2005 23:03 UTC (Thu) by paulj (subscriber, #341) [Link]

Interesting, that goes against completely against the sense of "Notwithstanding" that I've understood. The "notwithstanding" refers to some obstacle that may limit or make more difficult the effect/action being referred to. I had thought till now at least.

Notwithstanding

Posted Nov 10, 2005 16:50 UTC (Thu) by rfunk (subscriber, #4054) [Link]

A simpler version: No matter (and despite) what 2.1b says, 2.1d applies.

This is NOT about interfaces!

Posted Nov 9, 2005 23:39 UTC (Wed) by khim (subscriber, #9252) [Link] (16 responses)

The argument is a bit strained from a copyright POV. You have pre-existing GPL binaries, making use of fairly standard interfaces (C99, POSIX, SUS), such that they can be built against and link to any library implementing those interfaces. Solaris libc happens to be one of those. Precedent in US copyright law at least does not allow interfaces to be copyrighted, it's one of the (many) reasons why the SCO lawsuit is quite bogus.

And this all is relevant... exactly how ? I think you are confusing things.

Let's forget about interfaces, GPL and CDDL for a bit. Let's talk about two writers: Justina and Max. Both created some great stories and now publisher want to publish single book with stories by Justina and Max. Justina granted right to publish stories in any form to everyone but Max does only grant this to anyone who will not publish Justina's creations in the same book as his storis. Can our publisher then publish this book ? Of course not: you need permissions from all authors! Note: stories are totally unrelated - the only connection is single bookcover (and empty piece of paper is hardly copyrightable).

Now back to Nexenta, CDDL and GPL. As we've seen from sample above it does not matter of libraries are using the same interface - if they are published as single product (and Nexenta presumably wants to do exactly this) then licenses must be compatible. In fact this was reason for GPLv2 ! Note this part of it: In addition, mere aggregation of another work not based on the Program with the Program (or with a work based on the Program) on a volume of a storage or distribution medium does not bring the other work under the scope of this License. Why such exception was ever needed ? Easy: without this exception SUSE (with GPLed bash and proprietary Macromedia Flash) will be illegal! GPLv1 had no such provisions and this was a problem GPLv2 solved. May be GPLv3 will solve this problem, may be not - but right now Nexenta is in quite real trouble.

You are confusing issues: NVIDIA (who never shipped linux) can claim that "interface is not copyrightable" - since NVIDIA does not ship linux kernel and NVIDIA's binary driver on the same CD. Nexenta does not have such luxiry: the mere fact that both GPLed dpkg and non-gpl'ed OpenSolaris's libc are on the same CD "published" by single publisher is enough for GPL compatibility to be involved.

P.S. Actually if you'll think about it all distributions who actually ship NVIDIA's binary driver on the same CD as linux kernel are in trouble: they are violating GPL - for exactly the same reason. Nobody tried to sue them yet (AFAIK, anyway) but it's clear copyright violation.

This is NOT about interfaces!

Posted Nov 10, 2005 1:56 UTC (Thu) by paulj (subscriber, #341) [Link] (6 responses)

And this all is relevant... exactly how ?

Well because that clause may or may not be enforceable. If there's copyright "boundary" (there likely is) then it might be harder for the GPL to reach across it. But that'd be something you'd have to argue before a judge and jury.

Let's forget about interfaces, GPL and CDDL for a bit. Let's talk about two writers: Justina and Max. Both created some great stories and now publisher want to publish single book with stories by Justina and Max. Justina granted right to publish stories in any form to everyone but Max does only grant this to anyone who will not publish Justina's creations in the same book as his storis. Can our publisher then publish this book ? Of course not: you need permissions from all authors! Note: stories are totally unrelated - the only connection is single bookcover (and empty piece of paper is hardly copyrightable).

The problem is that that's likely not the some thing. In the case you describe, the publisher almost certainly has contracts with these author. If an author has a clause in their contract that makes their copyright licence conditional on not publishing with the work of some other author, then it is *CONTRACT* law which gives that clause weight, not copyright - contract law is far more flexible.

A licence under copyright alone is more difficult to have mandate restrictions not relating to copyright - you can *grant* rights with a unilateral licence, but you possibly can not take away rights over which you have no claim through copyright, for otherwise copyright licences could make all sorts of silly claims. You can simply disregard such clauses, for they have no effect on copyright (you can't break a copyright licence if copyright is not involved). Least, it used to be that way, until stuff like UCITA came along to start giving shrink-wrap copyright licences more contract-like powers, at least in the USA.

As a thought experiment, imagine if the GPL had a clause stating that you could not distribute GPL works if you also used MS Windows, or if you had a dog. I.e. think of some silly clause which would be legal under contract law (not breaching discrimination laws, etc.) yet has nothing to do with act of copying the work concerned. It's not too hard to think such clauses would be unenforceable, given the right judge ;). In the case of this clause, even more so when the licence *does* allow the act of distributing a GPL work without the OS components it depends on, the contentious components here.

It poses problems both way really. If its not enforceable, that's a problem for any non-copyright related /restrictive/ clauses in licences like the CDDL and GPL. There aren't many such clauses though, but this clause is one. If it's enforceable, then obviously yes there's a theoretical legal and moral problem for Nexenta GNU/OpenSolaris.

However, that's all quite speculative and there's also a decent chance my reasoning is flawed. ;) So lets just assume the clause is quite valid from now on. Where do things stand then? And isn't this a storm in a tea-cup?

Nexenta have a licence to distribute OpenSolaris (the CDDL). They have, other than this complication, a licence to distribute GPL packages. Further, the GPL *does* allow GPL software to depend on non-GPL OS components, even where they are *completely* proprietary, (OpenSolaris is "Free", if not GPL). It also allows GPL software to be distributed in aggregate with non-GPL works, even completely closed proprietary works.

So we're having a big argument because one Free application is being distributed with a Free OS, because the licence of the Free application has an unfortunate oversight in it which neither the licences author nor the author of the application probably ever intended to constrain such distribution.

"May be GPLv3 will solve this problem, may be not - but right now Nexenta is in quite real trouble."

How exactly? They're not in trouble if either:

a) (ideally) The dpkg copyright holder(s) grant a CDDL exception for dpkg. Or even just clearly state "hey, its perfectly ok by us" as an official position - even that would be sufficient legally under the British legal systems in US, UK and elsewhere.

b) (far less good, if only from a moral perspective) The dpkg copyright holder(s) don't sue.

And *no one* but the dpkg copyright holders can sue btw. Their opinion is all that matters. (I not sure what, if any, status those posts on those debian lists have wrt expressing the opinion of the dpkg copyright holders)

If a sensible solution can not found then that means a Free OS and Free applications essentially can not be distributed together, which I think most sides would agree is not a desirable outcome.

This is NOT about interfaces!

Posted Nov 10, 2005 4:34 UTC (Thu) by kimoto (subscriber, #5244) [Link]

And *no one* but the dpkg copyright holders can sue btw. Their opinion is all that matters. (I not sure what, if any, status those posts on those debian lists have wrt expressing the opinion of the dpkg copyright holders)

At least one dpkg author is hostile to linking dpkg with a CDDL libc.

$ zcat /usr/share/doc/dpkg/changelog{.Debian,}.gz | grep -F doogie@debian.org | wc -l

345

This is NOT about interfaces!

Posted Nov 10, 2005 4:34 UTC (Thu) by AJWM (guest, #15888) [Link] (3 responses)

You were so close ... and then you missed it.

As a thought experiment, imagine if the GPL had a clause stating that you could not distribute GPL works if you also used MS Windows, or if you had a dog. I.e. think of some silly clause which would be legal under contract law (not breaching discrimination laws, etc.) yet has nothing to do with act of copying the work concerned. It's not too hard to think such clauses would be unenforceable, given the right judge ;).

But, the license is the only thing that gives you permission under copyright law to distribute copies. A clause stating that "you may only distribute if you do not own a dog" is no less enforceable than a clause saying "you may only distribute if you pay the author one thousand dollars". Giving somebody money has "nothing to do with act of copying the work concerned", directly, but it is certainly the basis of most licenses to copy.

A copyright holder may be as capricious as he chooses (bar certain antidiscrimination laws) about what he wants in return for granting permission to copy his work -- you have no right otherwise to copy it.

Of course any judge might make an arbitrary, erroneous decision -- which is why there are appeals courts. Unless the clause on which a license grant was conditioned were actually illegal, there's no way that a decision which effectively vacates the copyright holder's rights and wishes would be upheld.

(ob: IANMALTAEPH - I am no more a lawyer than anyone else posting here)

This is NOT about interfaces!

Posted Nov 10, 2005 8:34 UTC (Thu) by khim (subscriber, #9252) [Link]

Hey, he said: given the right judge. At this point it does not matter what GPL says, what CDDL says - given the right judge you can get any resolution writtable in English.

This is NOT about interfaces!

Posted Nov 10, 2005 9:09 UTC (Thu) by paulj (subscriber, #341) [Link] (1 responses)

A copyright holder may be as capricious as he chooses

Hmm, well with further research it does indeed appear that I'm completely wrong in thinking that a licence can't mandate arbitrary conditions (as long as they are legal). Cheers ;).

This is NOT about interfaces!

Posted Nov 11, 2005 6:31 UTC (Fri) by hppnq (guest, #14462) [Link]

One of the nice things about the GPL is that it is actually quite easy to grasp. The license itself explains why it exists, and how it can be applied.

Which, unfortunately, does not mean that people are any less confused by it. Some parts are of course a bit hazy, but I suppose it's got more to do with the fact that people tend to cherish their own set of definitions and notions, even if they don't realize it.

Consider the extremely simple concepts of copyright and copyleft and how both are abused by both sides in discussions like these: lots of knee-jerking. What's all the more surprising, is that these are definitions, it's not like we've discovered a new law of nature that we can't fully understand yet.

PS: paulj, of course I'm not trying to ridicule you by suggesting how simple all this is, I'd much rather salute you for admitting you were, errr, not completely right -- and flexible enough to adjust your point of view.

This is NOT about interfaces!

Posted Nov 18, 2005 3:22 UTC (Fri) by robilad (guest, #27163) [Link]

I'd hope that the morality of a deed trumps the question wheather someone can be successfully sued. One might get away with doing something, but that does not make it a nice thing to do.

cheers,
dalibor topic

This is NOT about interfaces!

Posted Nov 10, 2005 9:20 UTC (Thu) by Duncan (guest, #6647) [Link]

> [A]ll distributions who actually ship
> NVIDIA's binary driver on the same CD
> as linux kernel are in trouble: they
> are violating GPL - for exactly the
> same reason. Nobody tried to sue them
> yet (AFAIK, anyway) but it's clear
> copyright violation.

They haven't been sued, but they are under notice from some of the kernel
hackers to cease and desist or face legal action.

See the LWN kernel page quote of the week for the Oct. 26 edition (quoting
Greg K-H). There's quite a discussion there as well, including a couple
rather long "expositions" on the legal details of the GPL and how they
relate to binary kernel modules, by yours truly. =8^)

http://lwn.net/Articles/157225/

Duncan

Sounds like co-distribution is perfectly legal

Posted Nov 10, 2005 15:39 UTC (Thu) by hazelsct (guest, #3659) [Link] (7 responses)

Now back to Nexenta, CDDL and GPL. As we've seen from sample above it does not matter of libraries are using the same interface - if they are published as single product (and Nexenta presumably wants to do exactly this) then licenses must be compatible. In fact this was reason for GPLv2 ! Note this part of it: In addition, mere aggregation of another work not based on the Program with the Program (or with a work based on the Program) on a volume of a storage or distribution medium does not bring the other work under the scope of this License. Why such exception was ever needed ? Easy: without this exception SUSE (with GPLed bash and proprietary Macromedia Flash) will be illegal! GPLv1 had no such provisions and this was a problem GPLv2 solved. May be GPLv3 will solve this problem, may be not - but right now Nexenta is in quite real trouble.

But aren't you disproving your own point here? Since the GPL expressly allows distributing non-GPL software in the same product, it would seem to allow distribution of GPL software with CDDL libc.

Furthermore, Sun and other proprietary unices have been distributing GPL gcc with proprietary libc implementations for literally decades without infringing, or at least without infringement suits. Why the sudden storm in a teacup over distribution with free software?

(Incidentally, this is not the same as the old KDE/Qt flap, which was about distributing GPL apps linked against non-free non-system libraries with non-standard interfaces, and nearly sank the original Corel Linux distribution -- or at least its apt/KDE software installation tool.)

As a Debian developer myself, I have to agree with paulj that the complaint of these DDs is a counterproductive restriction on free software, let alone childish infighting between free software camps which will more than likely damage the reputation of the Debian project and the community as a whole. To one day say, "SCO is wrong, interfaces are not copyrightable", and the next day turn around and say, "You'd better not link my free software to your free system software with a standard interface!" is truly pathetic.

Sounds like co-distribution is perfectly legal

Posted Nov 10, 2005 16:53 UTC (Thu) by vmole (guest, #111) [Link]

You've missed the distinction between "mere aggregation" and "derived work". The claim is that when you distribute (say) dpkg with the Solaris libc, you are distributing a derived work combining the dpkg code and the libc code, because (this particular binary) of dpkg will not work without (this particular binary) libc. Remember API!=ABI, and by including the header files provided by the Solaris libc, you're including Solaris libc specific code.

Sounds like co-distribution is perfectly legal

Posted Nov 10, 2005 17:12 UTC (Thu) by khim (subscriber, #9252) [Link] (5 responses)

But aren't you disproving your own point here? Since the GPL expressly allows distributing non-GPL software in the same product, it would seem to allow distribution of GPL software with CDDL libc.

Not really. GPL allows "mere aggregation" of two programs on the same medium regardless of their license if they are not related (relevant part os quoted in previous post).

Furthermore, Sun and other proprietary unices have been distributing GPL gcc with proprietary libc implementations for literally decades without infringing, or at least without infringement suits. Why the sudden storm in a teacup over distribution with free software?

This is other question. Why FSF never sued Apple and/or Sun ? May be they had no money or they were not sure it's a good idea. Who knows. Copyright owner is not forced to sue anyone who violates copyright (that's not trademark).

I agree that all this problem is quite silly and looks like storm in a teacup but this does not make it less real.

Sounds like co-distribution is perfectly legal

Posted Nov 10, 2005 20:54 UTC (Thu) by vonbrand (guest, #4458) [Link] (4 responses)

Oh, come on now. Before there even was Linux (and thus the possibility to even build a fully GPL-compatible system), all GPLed software had to run on propietary systems, simply because there was nothing else to be had. Before Linux, glibc was a sad joke, and everything used the propietary libraries that came with the OS. Standard joke was that the first task with a new Sun was GNU > /usr/local.

So now there is Linux. And glibc works. And what I did before to our Suns will not be tolerated anymore, just because?! I just can't see the difference between shipping gcc compiled for Solaris, and doing the same with dpkg. AFAICS (IANAL!) this kind of stuff is explicitly allowed by GPLv2.

Sounds like co-distribution is perfectly legal

Posted Nov 11, 2005 4:46 UTC (Fri) by kimoto (subscriber, #5244) [Link]

One rationale from Thomas Bushnell for why the funny wording is in the GPL2 (which was designed in the age of proprietary systems):
The special exception allows you to ship, for example, emacs binaries linked against the proprietary HPUX libraries, provided HP distributes those libraries along with the major components of HPUX (that is, they cannot have unbundled them), and provided you are not shipping those libraries yourself. This is specifically designed to prevent HP from including an emacs binary which is linked against their libraries, shipping the whole thing as part of HPUX, and not providing the source for their libraries.

Sounds like co-distribution is perfectly legal

Posted Nov 11, 2005 11:09 UTC (Fri) by khim (subscriber, #9252) [Link] (2 responses)

Oh, come on now. Before there even was Linux (and thus the possibility to even build a fully GPL-compatible system), all GPLed software had to run on propietary systems, simply because there was nothing else to be had.

Yes, but was GPLed software included with proprietary systems ? I know of one sample from that era: NEXTStep (later Mac OS X). All others shipped without any GPL components. If end-user does install GPL software on OS - this is one story (there are special exception in GPL), but if vendor ship GPLed component as part of base system... this is different story.

Sounds like co-distribution is perfectly legal

Posted Nov 11, 2005 13:12 UTC (Fri) by vonbrand (guest, #4458) [Link] (1 responses)

There sure were. DG-UX (Data General) used gcc as their stock compiler. The propietary (I believe source wasn't available) Pro toolset from Cygnus was built around gcc, gdb, binutils, etc. Not just some "shipping with" there, without those tools the rest made no sense at all.

Besides, this is more than a bit strange, all around: DEC provided a tape with "extraofficial software" with their systems, Sun later shipped a CD with such stuff. I'm sure other Unix vendors did likewise. There are vendor sites giving GPLed stuff compiled and packed for their systems (including whatever parts of the libraries and such this requires). And now, when the same thing is being done with an OS shipped in source form, under a somewhat free license (much freer that original Solaris or Ultrix, in any case),it is bad?

Sounds like co-distribution is perfectly legal

Posted Nov 13, 2005 20:28 UTC (Sun) by Ross (guest, #4065) [Link]

The same thing with NextStep. GCC was the system compiler. They modified it to support ObjectiveC, and were forced to release the source code to the ObjectiveC front end, but not for the whole operating system.

Debian and Nexenta collide

Posted Nov 10, 2005 6:18 UTC (Thu) by zlynx (guest, #2285) [Link] (1 responses)

Nexenta should have gone with Gentoo instead of Debian. If they were building packages from source code there would be no basis for CDDL/GPL copyright complaints.

Debian and Nexenta collide

Posted Nov 10, 2005 8:51 UTC (Thu) by Duncan (guest, #6647) [Link]

Interestingly enough, there /is/ a Gentoo Solaris project going. I don't
know how far it has gotten, but it's under the Gentoo alt umbrella, with
Gentoo FreeBSD and Gentoo OSX.

Duncan

Debian and Nexenta collide

Posted Nov 10, 2005 11:07 UTC (Thu) by kleptog (subscriber, #1183) [Link]

I hope I'm not going to ask something stupid here but...

The discussion is whether the distribution of GPLd program linked against a CDDL libc is compatable. The clause says "complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the executable". Whether the libc in this case falls under this is the point for debate.

My question is that clause 3 only says that this source code "must be distributed under the terms of Sections 1 and 2 above" *not* that it must be distributed under the GPL or compatable licence.

Maybe clauses 1 and 2 are the only ones that cause a problem for the CDDL but it would be helpful if someone could clarify this for me.

Debian and Nexenta collide

Posted Jan 17, 2006 2:22 UTC (Tue) by erast (guest, #35247) [Link]

Basically, Debian and Nexenta problem got resolved:
The end of story is here:
http://blogs.sun.com/roller/page/webmink?entry=gpl_v3_rel...


Copyright © 2005, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds