Jumping the licensing shark

Posted Mar 23, 2023 21:40 UTC (Thu) by bluca (subscriber, #118303)
In reply to: Jumping the licensing shark by Cyberax
Parent article: Jumping the licensing shark

> Yup, exactly. MPL2 is a relatively unknown license,

> it has some strange language about patents, and it doesn't provide that much protection anyway. So why bother?

What do you mean by "doesn't provide protection"? It's a copyleft license: if you ship my code in binary format, you also must ship the sources.

Jumping the licensing shark

Posted Mar 24, 2023 0:46 UTC (Fri) by Wol (subscriber, #4433) [Link]

Cries in LibreOffice too :-(

LibreOffice is MPL/GPL, and the GPL is very definitely the red-headed step-child. Okay, MPL2 explicitly permits re-licencing as GPL iirc, so there's no reason it shouldn't be MPL-only, but I think the LO crew are covering their backsides.

The original code was released GPL2, and it was Apache that did the audit and said "everything is now Apache 2.0". The LO crew haven't done their own audit, so they're playing safe and leaving the GPL in place, though in practice they might accept code that was solely MPL. They wouldn't touch solely GPL with a barge pole.

Cheers,
Wol

Jumping the licensing shark

Posted Mar 24, 2023 8:50 UTC (Fri) by joib (subscriber, #8541) [Link] (12 responses)

> > it has some strange language about patents, and it doesn't provide that much protection anyway. So why bother?

> What do you mean by "doesn't provide protection"?

Since it's a per-file weak copyleft, Evil Inc. can just put their proprietary extensions of the MPL2 code in a separate file, possibly helped by some trivial entrypoints to access internal functionality in the MPL2 licensed files which are of no interest to upstream. Of course this is largely possible with LGPL as well, just a few more minor hoops to jump through. Which is the crux of the matter: if it's this easy to circumvent, why bother in the first place? Either go "full copyleft" or then pick a simple permissive license.

The counter-argument, IMHO, as I mentioned in the grandparent comment, is that a weak copyleft communicates the intents and wishes of the author(s); "I don't care if you use this library proprietary code, but I'd like improvements to the library itself back, please", even if it's legally easy to circumvent.

Jumping the licensing shark

Posted Mar 24, 2023 8:55 UTC (Fri) by joib (subscriber, #8541) [Link]

s/if you use this library proprietary code/if you use this library in proprietary code/

Jumping the licensing shark

Posted Mar 24, 2023 12:18 UTC (Fri) by Wol (subscriber, #4433) [Link] (4 responses)

> if it's this easy to circumvent, why bother in the first place? Either go "full copyleft" or then pick a simple permissive license.

Because your motives are not the same as the people who choose that licence.

I mentioned ScarletDME, which, because it's a p-code engine, even the GPL cannot stop commercial actors "easily circumventing" it.

The point of the MPL is it provides a clear fence between what is Copyleft and what is commercial/proprietary. Something like MIT allows the whole gamut from "here's the complete source" through "here's the public source we started with" to "bugger off".

MPL at least draws the line at "here's the public source we started with PLUS any changes we made to that source".

Cheers,
Wol

Jumping the licensing shark

Posted Mar 24, 2023 20:41 UTC (Fri) by joib (subscriber, #8541) [Link] (3 responses)

> Because your motives are not the same as the people who choose that licence.

FWIW I have chosen MPL2 for some of my personal projects. I don't think it's a bad license, and I reflects how I wish said project to be used (that is, I don't care if you use it as part of some proprietary software, but I'd appreciate if any improvements to my project is contributed back). My argument here is merely that I don't harbor any particular illusion that the weak copyleft provisions in MPL2 would be a big roadblock for somebody that wishes to improve my code without giving anything back (e.g. by doing what I described in the grandparent comment). So from that perspective I should have just chosen a permissive license (for maximum convenience) or gone to a full copyleft license (to incentivize the creation of more FOSS software, and not help proprietary software to compete with FOSS).

As for ScarletDME, I've never heard of that project, but isn't it a bit like e.g. openJDK? So a compiler that generates some intermediate format bytecode (p-code like ye olde Pascal compilers?), and then a runtime that runs that code. And just like openJDK is GPL it doesn't mean that the GPL "infects" the code that is compiled and/or ran.

Jumping the licensing shark

Posted Mar 25, 2023 12:24 UTC (Sat) by Wol (subscriber, #4433) [Link] (2 responses)

ScarletDME - yes it's a compiler that creates byte-code, and an interpreter that runs said byte code. That all sits atop a powerful database.

A descendant of GIRLS and a clone of Pick - an OS/Database (along the lines of the IBM AS/400, I believe).

If you haven't seen my occasional database rants you must be new here :-) ScarletDME basically is an incarnation of "the database is the computer", although like all modern versions it's now an application that sits on top an OS.

It's MultiValue, not relational, so if you want to be converted to the light side, you can find my email easy enough. Find ScarletDME on github, or find me on linux raid.

Cheers,
Wol

Jumping the licensing shark

Posted Mar 27, 2023 12:38 UTC (Mon) by joib (subscriber, #8541) [Link] (1 responses)

> If you haven't seen my occasional database rants you must be new here

Well, my user id, while bigger than yours, is still relatively low, so I've been around on this site for quite a while.

As for your database rants, I've seen them, but it's not a topic particularly relevant to my interests so I tend to not comment on them.

Jumping the licensing shark

Posted Mar 27, 2023 14:28 UTC (Mon) by Wol (subscriber, #4433) [Link]

No I'd seen your id was low, hence the smiley :-) ...

The more I deal with SQL, the more I think it's a piece of ****, but it doesn't pay to off on one too often - it upsets people.

Cheers,
Wol

Jumping the licensing shark

Posted Mar 24, 2023 14:05 UTC (Fri) by bluca (subscriber, #118303) [Link] (5 responses)

That's a good thing, if what you are looking for is a license for a library that is usable by everybody everywhere, and you only care about changes to the library being shared back. As you said, this is already doable with LGPL and such other licenses, there are just silly hoops to jump through, which are largely a waste of time for everybody involved and little more.

Jumping the licensing shark

Posted Mar 24, 2023 20:30 UTC (Fri) by joib (subscriber, #8541) [Link] (4 responses)

> That's a good thing, if what you are looking for is a license for a library that is usable by everybody everywhere, and you only care about changes to the library being shared back.

I understand that's the motivation behind weak copyleft licenses in general, yes. My argument is that the protections offered by weak copyleft against entities that wish to break the spirit of the agreement while remaining compliant with the letter of the law, are not particularly strong. If one accepts that argument, one is thus faced with the choice of going either to a full copyleft license or to a permissive license, as the half-way point of weak copyleft offers neither the convenience of a permissive license nor the benefits offered by a full copyleft license.

Jumping the licensing shark

Posted Mar 27, 2023 12:06 UTC (Mon) by bluca (subscriber, #118303) [Link] (3 responses)

You keep characterizing in general terms as "weak", but that's only relative to the point of view. If you want your code to be used only in copyleft projects, then sure, it's "weak". If you don't care where it is used, but only that changes to it are shared back, then it's not "weak" at all, it's just as "strong" as any other license. That's the use case: you want to ensure that changes to your library are published without forcing a license change for all users. And it seems to fulfill that purpose just fine. You might disagree with the purpose, which is perfectly fine of course.

Jumping the licensing shark

Posted Mar 27, 2023 12:28 UTC (Mon) by joib (subscriber, #8541) [Link] (2 responses)

> You keep characterizing in general terms as "weak"

I use "weak" as in "weak copyleft", a commonly used term covering a bunch of licenses like LGPL, MPL, EPL and so forth. No moral judgement implied. See e.g. https://en.wikipedia.org/wiki/Copyleft#Strong_and_weak_co...

> You might disagree with the purpose

I'm not disagreeing with the purpose. I think the idea is perfectly valid, and suitable to a lot of software. My only argument, which I've been repeating in this entire subthread (and I'm beginning to wonder what I'm doing wrong as I'm clearly not making myself understood?), is that it's pretty easy to circumvent the protections offered by weak copyleft licenses. Thus, against an unscrupulous leecher that doesn't want to share their improvements to the library, they offer little more than a permissive license.

Jumping the licensing shark

Posted Mar 30, 2023 13:05 UTC (Thu) by bluca (subscriber, #118303) [Link] (1 responses)

The vast, vast majority of copyleft-licensed libraries use a "weak copyleft" license such as LGPL, because it's just unusable otherwise. What evidence do you have of "unscrupulous leechers" making improvements to said libraries and not sharing them due to the license?

Jumping the licensing shark

Posted Mar 30, 2023 20:08 UTC (Thu) by joib (subscriber, #8541) [Link]

> The vast, vast majority of copyleft-licensed libraries use a "weak copyleft" license such as LGPL, because it's just unusable otherwise.

I know, and I don't disagree. Though the FSF seems to think the world should create more GPL licensed libraries: https://www.gnu.org/licenses/why-not-lgpl.html

> What evidence do you have of "unscrupulous leechers" making improvements to said libraries and not sharing them due to the license?

Off the top of my head, I can't come up with any; OTOH it's not an issue that I've been tracking either. But if such cases are indeed very rare, as I suspect they are, that would imply that the entire issue of 'unscrupulous leechers' is mostly non-existent, and in reality users prefer to contribute back improvements to the libraries so they don't have to shoulder the burden of maintaining their own forks. From that it follows those libraries could just as well have been licensed under permissive licenses, as what protects the libraries from 'proprietarization' is mostly the desire to avoid maintaining their own fork rather than the weak copyleft license (which, as mentioned, is easy to circumvent in case anyone would want to go down that route).

Oh dear, this is starting to sound a lot like the open source side of ye olde 'open source vs. free software' debate.