[License-review] Request for Approval of Universal Permissive License (UPL)

[Henrik Ingo]

Richard, I think you make a worthwhile observation.

It is maybe worth pointing out that I'm aware of a similar arrangement
already happening in practice since 2008. MySQL has a CLA for inbound
contributions. However Google, Facebook and other web companies'
lawyers decided they didn't want to sign the MySQL CLA. So those
companies publish their MySQL contributions under BSD. Sun/Oracle have
then integrated those BSD licensed patches into the GPLv2 licensed
MySQL. Reluctantly, I'm sure, but this has been going on for years now
so that it's fair to say this is the primary way of contributing to
MySQL.

All this is just to say that I don't see this as such a special case
after all. There is precendent of a similar arrangement and it's just
another normal use for a permissive license. The UPL would be just yet
another license that could be used that way - in fact I'd read it in
favor of this license that there at least is some intended use for it
known by the time of submission. That's not always the case.

Still, the problems with the restrictive/exclusive dynamic around the
LARGER_WORKS.txt file remain. In my opinion they disqualify UPL as an
open source license. (It is true that it would still be a useable CLA,
as CLA's may indeed be used to restrict license grants to a specific
party or scope, creating asymmetric benefits not available to the
community at large.)

henrik

On Mon, Apr 14, 2014 at 5:05 PM, Richard Fontana
<fontana at sharpeleven.org> wrote:
> On Sun, 13 Apr 2014 10:41:11 -0700
> Jim Wright <jim.wright at oracle.com> wrote:
>
>> The intent here is to allow the authors to grant a license, including
>> patent rights, in a particular, listed larger work to which they're
>> contributing.  I wrote it with various uses in mind, but the most
>> immediate use case is in the Java Community Process for the
>> development of JSR reference implementations in other forges like
>> Eclipse.  Folks have been complaining for a long time that everyone
>> needed to sign an assignment to contribute, which won't work on other
>> forges (who don't want to require or inquire about Oracle assignment
>> agreements), and this way people will be able to contribute without
>> that.
>>
>> I also see a lot of other people's contribution agreements in my role
>> at Oracle (I have to approve all of them), and am hoping that a
>> standard form could help harmonize practices in this area.
>
> This is key for the OSI and license-review community to understand,
> and (based on the comments already made) what I'm not sure is going to
> be well understood unless it is spelled out really clearly here.
>
> Jim is asking OSI to approve, as an Open Source license, something that
> Oracle intends primarily to serve as a CLA, or rather a
> CLA/copyright-assignment substitute. Unlike typical CLAs, the UPL is
> written in a way that makes it entirely possible to use as a normal,
> outbound open source license. But that is AFAICT not how Oracle intends
> to use it and for all we know it may never be used in that way by
> anyone.
>
> So think of this as somewhat like the ASF submitting the Apache ICLA for
> approval as an Open Source License. Or Project Harmony submitting its
> contributor agreement suite for approval as an Open Source license. The
> only real difference is that there's a *chance* that the UPL might see
> some uptake as a *normal* open source license -- a chance that, say,
> the Ruby community or the JavaScript or Python communities might start
> using the UPL in place of their current tendency to use the MIT
> license (the license this resembles most closely).
>
> I would rate the chances of that happening as very low. I assume Oracle
> is not going to try to evangelize use of the UPL as an open source
> project license and even if it did it would be about as successful as,
> I dunno, Red Hat trying to evangelize some new open source license I
> might draft, which is to say probably not at all. :) The likelihood of
> open source communities or individual developers discovering this
> license on the OSI website (were it approved) and deciding to use it
> because they like the idea of an MIT-like minimalist permissive license
> that treats patents explicitly seems very low.
>
> None of that necessarily goes to the merits of the UPL or whether it
> should be approved by the OSI. But it's important to understand the
> unusual and unprecedented nature of this submission: Oracle wants the
> OSI to approve a license that it primarily intends to serve as a CLA for
> codebases that will *not* use this license on the outbound side. Jim
> has been very transparent about that -- he specifically noted the use
> case of commercial developers of copyleft and commercial (proprietary)
> software.
>
> This is, I believe, the first time that the OSI has ever been asked to
> approve a new license that is primarily intended by its license
> steward to serve as a CLA (loosely defined), an inbound contributor
> license, not as an outbound open source project license. In all past
> license submissions, to my knowledge, the license was contemplated as an
> *outbound* license. That is what is (nonsubstantively) new here.
>
> (Jim if you think I am wrong about any of the foregoing, please
> comment. :)
>
>  - RF
>
>
> _______________________________________________
> License-review mailing list
> License-review at opensource.org
> http://projects.opensource.org/cgi-bin/mailman/listinfo/license-review



-- 
henrik.ingo at avoinelama.fi
+358-40-5697354        skype: henrik.ingo            irc: hingo
www.openlife.cc

My LinkedIn profile: http://fi.linkedin.com/pub/henrik-ingo/3/232/8a7