[kde-community] Request to join the Kde incubator for GCompris

[Laszlo Papp]

Hi Aaron,

On Thu, Feb 13, 2014 at 10:02 PM, Aaron J. Seigo <aseigo at kde.org> wrote:
> On Thursday, February 13, 2014 21:16:27 Laszlo Papp wrote:
>> On Thu, Feb 13, 2014 at 8:59 PM, Aaron J. Seigo <aseigo at kde.org> wrote:
>> > On Thursday, February 13, 2014 20:32:36 Laszlo Papp wrote:
>> * Before GPLv3, GPLv2 was a commercial friendly license. As far as I
>> know GPLv3 made it untenable.
>
> This is false. Samba (which you mention below) is used in many commercial
> products and is licensed under the GPLv3. So obviously it is not "commercially
> untenable".
>
> There are specific markets (which I already named) which have issues with GPLv3
> due to their business models. This is not a generic issue with the GPLv3, but
> one that comes into play in specific areas of the stack for specific types of
> companies. Is GPLv3 as commercially-friendly as GPLv2? No, it is not. Is it
> commercially untenable? No.

Yes, that is what I meant. There are cases where it is "untenable",
but it is not globally like that. I agree that it would have been
clearer to write "less commercial friendly".

> Moreover, I don't think it is our job to tell an application developer whether
> their product needs to be commercially viable or not. That is their decision
> to make as the creator of the work, imho.
>
>> * You  cannot go back to GPLv2 freely.
>
> This is true of other GPL versions as well, so nothing really new.

Sure, it is not new, but disadvantegous due to flexibility. Should one
decide one day to prefer code sharing, this limitation would block it;
at least in my current understanding.

>> * You already mentioned the sharing yourself. I will give you a
>> practical example: Linux and Samba cannot share code for implementing
>> two ends of the same protocol.
>
> Well, yes, we agree that GPLv3 code can not be shared with code released under
> licenses. It sucks when that happens, and I wouldn't encourage everyone to run
> to the GPLv3 due to that. The question is whether we can accept developers who
> choose to do so for their own code, however.

I respect maintainers for their decisions, but I am also trying to
consider from KDE point of view as a whole. Currently, if a maintainer
decides to go for GPLv3+, and then leaves the project, it is possible
that a rewrite would need to happen for the new people getting
involved to avoid the limitation of GPLv3 which is not present in
GPLv2. This would impose an additional overhead for the project.

>> * One could ask LGPLv3+ in the future (someone consistently after this
>
> Personally, I don't think we should adopt LGPLv3[+] as an acceptable library
> license: it creates too many issues that are not easily resolvable, as you
> note; however, this is not about libraries but applications. We already have
> different policies for libraries vs. applications.

I agree with you. My concern here is that many KDE applications start
off as a monolithic codebase without clear library and application
separation, hence the growing "library" code internally will manifest
as GPLv3 or LGPLv3, and then you have the issue later unless there is
a nice trick to get around it.

>> There are more among these lines. I am personally not happy with where
>> FSF went with this.
>
> I think we should try and keep the FSF and developers who pick that license
> separate in our minds.

Yes, I agree. I was just referring to that I oppose to the GPLv3.

I would also like to note that making such a significant change in the
base values of KDE is more than just including a new project appearing
to have this license already in place.

That being said, I am fine with exceptions for existing code to also
remain inclusive because I think that is also very important for KDE.

Cheers, L.

PS.: I would still go for asking the KDE/Qt foundation for their
advice as well, or some other group with experience around licenses,
as written in my original email, if that is the route the community
wishes to take.

PS2.: There are further drawbacks of the license itself. I only
mentioned a small portion of it. Bottomley and T'So have cogent
arguments against GPLv3, too. Google for it, or just check out this
for starter: https://lkml.org/lkml/2006/9/22/217