[Advisors] Thoughts on a license change.

Greg DeKoenigsberg greg.dekoenigsberg at eucalyptus.com
Mon Aug 12 15:30:57 UTC 2013


No sense in making a move to permissive licensing unless there's a
plan to go with it.

If you can identify big industry partners who would work with Gluster
but do not because of the GPL, then I think it would be time to build
a plan to move away from it.  Can you identify any such partners?

You don't want to go through the hassle of a move until you have
identified concrete benefits, I don't think.  And you obviously have a
big issue to sort re: QEMU integration regardless.

But generally, there's nothing inherently superior about GPL, and many
drawbacks, and all things being equal, I think permissive is probably
superior.  Even the dual-licensing tactic is suspect; you can get the
same ability to create differentiated product offerings with
permissive licenses, without the headaches of copyright
assignment/licensing. The one unique benefit you *do* get from
dual-licensing is proprietary redistribution lock-in -- and this may
be a useful tactic down the road. You need ubiquity first. This was so
successful for MySQL because of the sheer number of people using
MySQL, and the fundamental suitability of a slightly-modified database
as an embedded piece of other products.  I don't think Gluster has
those same advantages.

--g

On Mon, Aug 12, 2013 at 11:20 AM, John Mark Walker <johnmark at gluster.org> wrote:
>
> ________________________________
>
> Yes, ideally. How ever, linking to older GPLv2 code may be problematic with
> Apache License. QEMU is GPLv2.
>
>
> Of the reasons against, this is the one that gives me pause. Our signature
> feature that went into 3.4 was the QEMU integration. We cannot, under any
> circumstance, do anything to jeopardize that collaboration. I'm not entirely
> sure what the consequences are of linking Apache-licensed code to GPLv2
> libraries. I can ask for a report on that and present the results here.
>
> Otherwise, I'm generally in favor of the move. I used to be a huge fan of
> copyleft, but over the years I've seen that community governance and
> development process seem to mean a whole lot more than which license a
> project chooses.
>
> I'd be very curious if anyone here strongly disagrees with a move to the
> Apache License 2.0, the QEMU integration issue notwithstanding.
>
> -JM
>
>
>
> On Sat, Aug 10, 2013 at 11:07 AM, David Nalley <david at gnsa.us> wrote:
>>
>> Personally, I see dual-licensing as being better, but not ideal. It
>> still leaves a lot of confusion on the part of the non-FOSS-zealots.
>> I'd personally advocate for a single permissive license.
>>
>> --David
>>
>> On Sat, Aug 10, 2013 at 1:54 PM, Anand Babu Periasamy
>> <abperiasamy at gmail.com> wrote:
>> > I cannot agree more. I have been advocating to dual-license or
>> > re-license
>> > GlusterFS under Apache Software License v2 for the last two years. GPL
>> > defends freedom strongly, but hurts adoption. Folks at Red Hat care
>> > about
>> > software freedom more than adoption. Lets push it once again.
>> > -ab
>> >
>> >
>> > On Sat, Aug 10, 2013 at 8:40 AM, John Mark Walker
>> > <johnmark at johnmark.org>
>> > wrote:
>> >>
>> >> Adding advisors list.
>> >>
>> >> On Aug 10, 2013 11:16 AM, "David Nalley" <david at gnsa.us> wrote:
>> >>>
>> >>> Hi folks,
>> >>>
>> >>> Just tossing out a question or two.
>> >>>
>> >>> So I'd like to propose that we consider changing to either the ASLv2,
>> >>> MIT, or BSD licenses.
>> >>>
>> >>> Why? So I personally strongly identify with copyleft principles, but
>> >>> my experience in the past few years are that the practical and
>> >>> irrational concerns around licensing hurt adoption and hurt
>> >>> contribution.
>> >>>
>> >>> Specifically, I found that several very large enterprises (~100k
>> >>> employees each) said that they never even considered CloudStack at the
>> >>> time because it was licensed under a GPL license. The dual-licensing
>> >>> bit muddies the waters a bit rather than helps. For the folks who are
>> >>> educated very well in open source, it's great. For folks who aren't as
>> >>> sophisiticated in OSS licensing it's merely confusing.
>> >>>
>> >>> Second - there's the potential damper on contribution. Despite how
>> >>> long GPL has been around, much FUD still remains around copyleft
>> >>> licensing; and that keeps people employed by large corporate users
>> >>> from contributing (at least that has been my experience) The more
>> >>> enlightened understand that it isn't going to virally apply to
>> >>> anything that they develop, but there is still a substantial number of
>> >>> companies that simply don't get it.
>> >>>
>> >>> Finally - I don't see a downside to becoming more permissively
>> >>> licensed aside from the work involved. Moving to a single, liberal
>> >>> open source license has the potential for us to increase our community
>> >>> size, both user and contributor. And from a weird marketing angle,
>> >>> it's also likely, as a one time event, to drive some interest in the
>> >>> project, as relicensing events tend to be geeky news that attracts
>> >>> attention.
>> >>>
>> >>> Having done it once, I know it's a ton of work to get all of the
>> >>> contributors to agree to relicense. That said, what are the collective
>> >>> thoughts on this?
>> >>>
>> >>> --David
>> >>> _______________________________________________
>> >>> Board mailing list
>> >>> Board at gluster.org
>> >>> http://supercolony.gluster.org/mailman/listinfo/board
>> >>
>> >>
>> >> _______________________________________________
>> >> Board mailing list
>> >> Board at gluster.org
>> >> http://supercolony.gluster.org/mailman/listinfo/board
>> >>
>> >
>> >
>> >
>> > --
>> > -ab
>> >
>> > Imagination is more important than knowledge --Albert Einstein
>
>
>
>
> --
> -ab
>
> Imagination is more important than knowledge --Albert Einstein
>
> _______________________________________________
> Advisors mailing list
> Advisors at gluster.org
> http://www.gluster.org/mailman/listinfo/advisors
>
>
>
> _______________________________________________
> Advisors mailing list
> Advisors at gluster.org
> http://www.gluster.org/mailman/listinfo/advisors
>



-- 
Greg DeKoenigsberg, Eucalyptus
Build your own AWS-compatible cloud in 30 minutes:
http://eucalyptus.com/faststart


More information about the Advisors mailing list