He makes an offhand comment implying he doesn't feel they're incompatible, but not in any official stance as an Oracle lawyer. He then qualifies that by saying, more or less, Oracle has not come after anyone over OpenZFS, so everyone should be taking that as an indication that it's OK with Oracle.
It is till it isn't. And some lawyer, even Oracles lawyer, who says in an offhand comment something is fine means nothing if it ever goes to court and "but I saw it in a video" won't be a defense.
He's not the only person that feels CDDL and GPL licenses are compatible. Their alleged incompatibility[1] seems to stem from a lot of hair splitting over technicalities that don't align with the spirit of either license. [2][3][4]
It's particularly funny that the community (except Canonical and a handful of individuals) largely settled on the CDDL-GPL issue--deciding they were incompatible and giving up on the possibility of upstreaming ZFS--but is still so highly split on the question of RHEL's GPL compatibility w.r.t. calling source code sharing a violation of their customer agreement.
The RHEL issue stems from the GPL literally saying,
> You may not impose any further restrictions on the recipients' exercise of the rights granted herein.
Conspicuously, you may not impose restrictions on the EXERCISE of these rights, not just further restrictions to the rights themselves. There's a pretty clear argument to be made that RHEL's actions in ending business relationships with any customers would be a restriction stemming from their having exercised their right. There's basically no hair splitting here. This should be done and dusted in the collective open source community conscious.
But the CDDL-GPL issue? There is no smoking gun in either license. The arguments against are much more strenuous.
But the community settled one debate much more decisively than the other.
If ZFS was as desirable to the Linux community as a fully open RHEL, I think there would've be much more enthusiasm in interpreting legal theory around it.
I think the community ought to be coming down much harder around the theory of RHEL incompatibility. It's really not a stretch from the most straightforward reading of the license.
It's also pretty conspicuous that Oracle has done precisely nothing to stop OpenZFS's inertia. If they felt it was incompatible and were going to act, they had their chance.
Agreed. It's especially ironic because Red Hat itself was so precious about reading in an incompatibility re the ZFS-Linux combination, and now it's parsing the words of the GPL so closely.
> There's a pretty clear argument to be made that RHEL's actions in ending business relationships with any customers would be a restriction stemming from their having exercised their right.
Disagree, here, though. An additional restriction on the exercise of rights, might be, "This/your license terminates should you distribute the source packages distributed to you." Future updates and support are not guaranteed in the world of the GPL, are not even mentioned, and, your rights are not restricted, when I refuse to do business with you. You still have a license to the software you received. And you can go do business with SUSE or someone else!
> your rights are not restricted, when I refuse to do business with you
This is the part that I don't think it's so safe to assume. A case can and should be made that my rights are restricted and/or I have experienced genuine harm/damage as a result. You would have to look at the GPL violation and the contract violation separately, each in a vacuum from one another, to argue otherwise.
Courts may want to look at the whole picture. This looks very much like a "strings attached" situation on the license of the software. Courts generally do respect the authors' intent when issues of copyright/license infringement are not clear-cut, and the GPL is extremely wordy about what the intent was -- freedom to share and deal in the software, out in the open.
Red Hat's own corporate materials on open source acknowledge as much, so that's ammunition for any court battle. They can't claim ignorance of the spirit of open source (or the spirit of the licenses in question). They have been including verbiage about the benefits of openness and public access/development of open source software when selling the very product in question.
This seems to be anything but a slam dunk case if really tested in the courts. I think there's a genuinely good chance that somebody with the pockets to challenge this wins and sets a significant precedence for open source.
8 comments
[ 3.4 ms ] story [ 29.7 ms ] threadIt is till it isn't. And some lawyer, even Oracles lawyer, who says in an offhand comment something is fine means nothing if it ever goes to court and "but I saw it in a video" won't be a defense.
It's particularly funny that the community (except Canonical and a handful of individuals) largely settled on the CDDL-GPL issue--deciding they were incompatible and giving up on the possibility of upstreaming ZFS--but is still so highly split on the question of RHEL's GPL compatibility w.r.t. calling source code sharing a violation of their customer agreement.
The RHEL issue stems from the GPL literally saying,
> You may not impose any further restrictions on the recipients' exercise of the rights granted herein.
Conspicuously, you may not impose restrictions on the EXERCISE of these rights, not just further restrictions to the rights themselves. There's a pretty clear argument to be made that RHEL's actions in ending business relationships with any customers would be a restriction stemming from their having exercised their right. There's basically no hair splitting here. This should be done and dusted in the collective open source community conscious.
But the CDDL-GPL issue? There is no smoking gun in either license. The arguments against are much more strenuous.
But the community settled one debate much more decisively than the other.
If ZFS was as desirable to the Linux community as a fully open RHEL, I think there would've be much more enthusiasm in interpreting legal theory around it.
I think the community ought to be coming down much harder around the theory of RHEL incompatibility. It's really not a stretch from the most straightforward reading of the license.
It's also pretty conspicuous that Oracle has done precisely nothing to stop OpenZFS's inertia. If they felt it was incompatible and were going to act, they had their chance.
[1] https://www.fsf.org/licensing/zfs-and-linux
[2] https://blog.hansenpartnership.com/are-gplv2-and-cddl-incomp...
[3] https://opensource.stackexchange.com/a/2106
[4] https://softwarefreedom.org/resources/2016/linux-kernel-cddl...
Agreed. It's especially ironic because Red Hat itself was so precious about reading in an incompatibility re the ZFS-Linux combination, and now it's parsing the words of the GPL so closely.
> There's a pretty clear argument to be made that RHEL's actions in ending business relationships with any customers would be a restriction stemming from their having exercised their right.
Disagree, here, though. An additional restriction on the exercise of rights, might be, "This/your license terminates should you distribute the source packages distributed to you." Future updates and support are not guaranteed in the world of the GPL, are not even mentioned, and, your rights are not restricted, when I refuse to do business with you. You still have a license to the software you received. And you can go do business with SUSE or someone else!
This is the part that I don't think it's so safe to assume. A case can and should be made that my rights are restricted and/or I have experienced genuine harm/damage as a result. You would have to look at the GPL violation and the contract violation separately, each in a vacuum from one another, to argue otherwise.
Courts may want to look at the whole picture. This looks very much like a "strings attached" situation on the license of the software. Courts generally do respect the authors' intent when issues of copyright/license infringement are not clear-cut, and the GPL is extremely wordy about what the intent was -- freedom to share and deal in the software, out in the open.
Red Hat's own corporate materials on open source acknowledge as much, so that's ammunition for any court battle. They can't claim ignorance of the spirit of open source (or the spirit of the licenses in question). They have been including verbiage about the benefits of openness and public access/development of open source software when selling the very product in question.
This seems to be anything but a slam dunk case if really tested in the courts. I think there's a genuinely good chance that somebody with the pockets to challenge this wins and sets a significant precedence for open source.
English is funny. Both "strained" and "tenuous" work here, so why not "strenuous"?
A dictionary definition says something like "characterized by great effort" so yeah. I wouldn't win that argument. But it sounds wrong to my ear.
https://xkcd.com/739/