Kirkland: ZFS licensing and Linux

Posted Feb 18, 2016 23:07 UTC (Thu) by juliank (guest, #45896)
Parent article: Kirkland: ZFS licensing and Linux

I can only say that the Conservancy continues to "investigate the evolving situation."

https://twitter.com/conservancy/status/700094818035245057

Well, I could say more, but I really shouldn't.

Kirkland: ZFS licensing and Linux

Posted Feb 18, 2016 23:09 UTC (Thu) by juliank (guest, #45896) [Link] (19 responses)

Oh, with regards to nvidia.ko: I'm working on getting pre-compiled bits removed from Debian:

http://bugs.debian.org/815060

Kirkland: ZFS licensing and Linux

Posted Feb 19, 2016 16:33 UTC (Fri) by fandingo (guest, #67019) [Link] (18 responses)

> Several lawyers and people believe

As other commenters say, please produce these opinions, so we can evaluate them. How are you qualified to make legal decisions?

>Legal issues prevent me from saying more

Then table the proposal until you can discuss it freely. There's nothing requiring immediate action, so a deliberate path is best.

Kirkland: ZFS licensing and Linux

Posted Feb 19, 2016 16:40 UTC (Fri) by juliank (guest, #45896) [Link] (6 responses)

The first comment was obviously not entirely correct, as it missed a few points, but I assumed everyone knew the issue at hand: Combining GPL-incompatible works with GPLed work should be considered a copyright violation.

The second comment: I meant to say: I cannot say who says what, as that discussion was confidential. I can only say that people involved in kernel GPL compliance were notified in that process, whether by accident or planned.

Ben now wrote an email detailing his position, so this is already taken care of (others might comment too). Having Debian's kernel maintainer and major kernel contributor say that he considers it a violation of the kernel license should make it obvious that we need to take action on it.

Kirkland: ZFS licensing and Linux

Posted Feb 19, 2016 18:41 UTC (Fri) by fandingo (guest, #67019) [Link] (4 responses)

> I cannot say who says what, as that discussion was confidential.

Then, I assume you understand we can't give those sources any weight. We can't evaluate the opinions, and we can't evaluate the credibility of those making the opinions. They are spurious.

> Having Debian's kernel maintainer and major kernel contributor say that he considers it a violation of the kernel license should make it obvious that we need to take action on it.

Not really. Is he a lawyer? Does this person have some unique knowledge of international copyright law that qualifies him to reach a conclusion? This is a bunch of software developers who want to play lawyer on the Internet. The claims you're (and others in the thread are) making about violating licenses and stuff are unsubstantiated and uncredible without supporting materials.

You guys may well be right, but the argument and evidence presented don't formulate anything that can be used to reach a conclusion. Barring such convincing items, the proper action is to defer any decision -- letting the current behavior remain -- until a conclusion can be reached.

Kirkland: ZFS licensing and Linux

Posted Feb 19, 2016 19:08 UTC (Fri) by juliank (guest, #45896) [Link] (2 responses)

We don't need lawyers to make an internal decision to move out of a potential copyright violation, thank you very much. There's no reason for us to be involved in such shady business. This is a proactive action to avoid that. That avoids work and also saves money for more important GPL enforcement efforts like the VMWare case or the ZFS thing.

Technically, there also is no real benefit for everyone, given that both dkms and module-assistant packages are provided.

Kirkland: ZFS licensing and Linux

Posted Mar 22, 2016 1:03 UTC (Tue) by Rudd-O (guest, #61155) [Link] (1 responses)

Not commenting on the legal issues.

From a practical standpoint, use of DKMS effectively prevents distro installers from installing distros to ZFS file systems. It also HUGELY complicates ZFS upgrades, particularly if they need to make their way into the initial RAM disks necessary to support systems that require ZFS to boot.

If that was shipped as a kernel module and matched with the distribution's offering of kernel modules, all of that would be a thing of the past.

Of course, that would require the relevant bloodsuckers to agree instead of to fight.

Kirkland: ZFS licensing and Linux

Posted Mar 22, 2016 11:56 UTC (Tue) by tao (subscriber, #17563) [Link]

zfs-fuse for the installer (which isn't really performance critical), native module built using dkms for normal operations.

Also, what's "HUGELY" complicated about using dkms? dkms will automagically rebuild your kernel modules and install them to the initrd, no manual work necessary.

All that said I wouldn't touch ZFS with a ten-foot pole unless they change their license.

Kirkland: ZFS licensing and Linux

Posted Feb 20, 2016 8:49 UTC (Sat) by Del- (guest, #72641) [Link]

> You guys may well be right, but the argument and evidence presented don't formulate anything that can be used to reach a conclusion. Barring such convincing items, the proper action is to defer any decision -- letting the current behavior remain -- until a conclusion can be reached.

It seems you are playing it into the hands of those who want to abuse copyleft. For a company like Canonical, the intent of a license should suffice for them to honour it. With your statements your are supporting those who look for technical weaknesses in current law for their own profit, willingly breaking the intent of the law. There are many examples of it, usually driven by greed at the expense of society.

You need to ask yourself, did the copyright holder of ZFS specifically intend for ZFS not to be used in GNU/Linux? Then ask yourself, do the copyright holders of Linux specifically intend for kernel modules being GPL compatible? Then it should not be so damn hard to conclude. Canonical is going down a path I cannot follow, heck even Facebook puts their resources into btrfs rather than breaking license terms. Is it so damn hard to be a decent member of the open source community.

Kirkland: ZFS licensing and Linux

Posted Feb 28, 2016 17:37 UTC (Sun) by Wol (subscriber, #4433) [Link]

> The first comment was obviously not entirely correct, as it missed a few points, but I assumed everyone knew the issue at hand: Combining GPL-incompatible works with GPLed work should be considered a copyright violation.

Except that this is EXPLICITLY PERMITTED by the GPL.

DISTRIBUTING such a combined work, on the other hand ...

Cheers,
Wol

Kirkland: ZFS licensing and Linux

Posted Feb 19, 2016 17:42 UTC (Fri) by Wol (subscriber, #4433) [Link] (10 responses)

> Then table the proposal until you can discuss it freely

Isn't that what he just did - tabled it so we could discuss it?

:-) Hint - "table it" is not a very good phrase to use - I suspect it means the complete opposite to me than to you - it has a completely different meaning in English. (As opposed to American.)

Cheers,
Wol

Kirkland: ZFS licensing and Linux

Posted Feb 19, 2016 18:25 UTC (Fri) by fandingo (guest, #67019) [Link] (9 responses)

Table means to postpone consideration. It means that in both languages. The only discrepancy is the expectation on when consideration will resume. British English implies soon, but American English implies later or possibly never.

You could use either definition and not lose my meaning.

You seem to believe that things "on the table" are under active consideration. That's not true in either language.

Kirkland: ZFS licensing and Linux

Posted Feb 19, 2016 19:01 UTC (Fri) by Pc5Y9sbv (guest, #41328) [Link]

Actually, there are two different idioms understood consistently in all English dialects, except for certain members of either population who may be ignorant of one of the idioms. Tabling a discussion point is not the same idiom as putting something on the table.

A completely different metaphorical mapping is at play for "tabling" versus "putting/taking on/off the table". The former invokes a governmental procedure to change topics away from the item being tabled, while the latter invokes including/excluding a topic or item for negotiation or transaction in a business or gambling setting. The very specific syntax of table as a verb versus putting or taking are the idiomatic signals for these two different metaphors.

This post is about interoperable communication so I think is somehow on topic for LWN and license lawyering. :-)

Kirkland: ZFS licensing and Linux

Posted Feb 19, 2016 22:51 UTC (Fri) by pjtait (subscriber, #17475) [Link]

I think it is true in non-US English, according to OED: "(a) To present or submit formally for discussion or consideration." Seems to imply "active consideration", no mention of postponement....

Kirkland: ZFS licensing and Linux

Posted Feb 20, 2016 5:08 UTC (Sat) by csamuel (✭ supporter ✭, #2624) [Link]

fandingo writes:

> Table means to postpone consideration. It means that in both languages.

As someone from the UK who has spent more than a decade living in Australia I've never heard it used in that manner in either country (so far).

However, my 1990 copy (14th edition) of Brewers Dictionary of Phrase and Fable surprised me in that it gives both meanings as being used; both to postpone discussion and to put a matter up for discussion.

Language is a wonderfully bizzare thing. :-)

Kirkland: ZFS licensing and Linux

Posted Feb 21, 2016 10:04 UTC (Sun) by paulj (subscriber, #341) [Link] (5 responses)

I never knew that "to table a discussion" could mean "to postpone indefinitely" to an American speaker, if the timeframe of that discussion is left implicit.

The Irish & British use might not mean "under active consideration" right now, but it does imply some imminence to that discussion. At the very least, one would expect that the next discussion will include that topic, so if there is any discussion then that topic should be part of it.

Learn something new every day.

Kirkland: ZFS licensing and Linux

Posted Feb 21, 2016 12:37 UTC (Sun) by Wol (subscriber, #4433) [Link] (3 responses)

There's a great story about Churchill and Roosevelt, meeting in mid-Atlantic in 1942, where some advisor came up with a good idea.

"We need to table that" says Churchill.

"No no, we mustn't table it" says Roosevelt.

Cue some quiet amusement when they realised what had happened :-)

Cheers,
Wol

Kirkland: ZFS licensing and Linux

Posted Feb 21, 2016 16:48 UTC (Sun) by shmget (guest, #58347) [Link] (2 responses)

"The United States and Great Britain are two countries separated by a common language. " (uncertain attribution)

Kirkland: ZFS licensing and Linux

Posted Feb 25, 2016 19:48 UTC (Thu) by Wol (subscriber, #4433) [Link]

Just to add to that, as I understand "to table something" it doesn't necessarily mean to discuss immediately, but in *English* it very definitely gives the topic some urgency.

Used outside of a meeting, it would be understood as "this needs to be on the next agenda", and in a meeting we would typically say "this needs to be tabled for the next meeting".

Having watched one of my local council meetings, it was very definitely the case that if it wasn't on the agenda it wouldn't get discussed, so tabling something would make sure it got on the next one.

Cheers,
Wol

Kirkland: ZFS licensing and Linux

Posted Mar 22, 2016 1:04 UTC (Tue) by Rudd-O (guest, #61155) [Link]

Obviously by Oscar Wilde.

Kirkland: ZFS licensing and Linux

Posted Feb 21, 2016 17:16 UTC (Sun) by Pc5Y9sbv (guest, #41328) [Link]

It doesn't necessarily mean indefinite postponement to Americans. It really just means let's stop talking about this and move on to other items in an agenda. The ambiguity of intention may come from a cultural awareness of US congressional tactics, where tabling something may be a precursor to sending it to a smaller committee, where it may be mangled beyond recognition or left to die before it is ever brought before a full session again. So, it can have this connotation of terminating broad debate and handing off to power brokers.