OVPL and open ownership

[Alex Bligh]

Chris,

--On 26 July 2005 17:44 +0200 Chris Zumbrunn <chris at czv.com> wrote:

> On Jul 26, 2005, at 5:10 PM, Alex Bligh wrote:
>
>> The OVPL (as amended) *would* allow them to be distributed under the
>> same
>> terms as the license as the original software. This is for two reasons:
>> a) The BSD-esque terms for contributions would form part of the OVPL
>> b) The contributors would be fully in their rights to distribute
>>   modifications under the OVPL even if this was not the case - it just
>>   says they must ALSO be distributed in BSD form.
>> IE it seems to me that the clause is there to prohibit RESTRICTIONS
>> on distribution of source, not to prohibit mandatory OPENING of source.
>
> If the terms would be the same then either both the initial contributor
> and all other contributors COULD NOT use others contributions in
> proprietary versions or both the initial contributor and all other
> contributors COULD use others contributions, including the initial
> contribution, in proprietary versions.

I don't think that is correct. If it was, this would effectively mandate
that all OSD approved licenses prohibited dual licensing by the ID
and by contributors.

As I understand OSD #3 is saying that the contributions MUST be available
under "the same terms as the license of the original software"
(that's the OVPL with the specified ID). It doesn't say they mustn't be
available on other terms as well (for instance more favourable terms).
Clearly, anything licensed on BSD terms is going to be available under
any OSI approved license.

>> Anyway, the above (slightly semantic) points aside, that's one reason
>> why
>> if we go with the BSD proposal, one option would be to do it by making
>> a
>> license back to the original developer (as presently) which is
>> conditional
>> on the ID making them available under a BSD-esque license - IE if the
>> ID
>> uses them in a proprietary version, the ID has to make them available
>> under
>> a BSD-esque license (as opposed to just under the under the OVPL). IE
>> if the
>> ID take special advantage of any code, so can anyone else.
>
> That would still not treat the initial contribution in the same way as
> contributions by others.

Correct. But that's not what OSD #3 says. Several existing licenses
deviate here to various extents.

>>> And if the initial contributor asks that modifications
>>> by others must be licensed to him for use in proprietary versions but
>>> other contributors can't do the same then that means they cannot make
>>> their modifications available under the same terms that the initial
>>> contributor did.
>>
>> I don't think so, because even under the current license the ID can
>> only
>> successefully demand (he can always ask :-) ) such modifications be
>> licensed to him if he makes them available under the terms of the OVPL.
>
> But you're giving special treatment to the initial contribution. And what
> about future contributions by the initial developer? Do they have to be
> BSD-esque as well?

No.

In the "make all contributions subject to the OVPL but make the ID
sublicense as BSD if he uses them in a proprietary version" model, the ID
only needs to make a BSD-esque grant if he makes use of the 3.3 grant. If
he does not require the 3.3 grant (because he has copyright, because he
received a separate out-of-band assignment, because he paid the contributor
to license him the work, whatever), then he does need to make a BSD
sublicensed version available.

In the "make all contributions BSD-esque licensed" model (i.e. change the
contributor grant), the contributor grant would apply to all
contributibutions other than those made by the ID, and the ID grant by the
ID. This might seem bizarre until you realize that even without a carve-out
for the ID as contributor, under 3.3 the ID could relicense the whole thing
as a "new" OVPL project with the same ID, with the "modifications" forming
part of the new Original Work (i.e. noone loses anything this way).

If your point is that the latter is ugly, and it's far harder to
see how it fits the OSD (though I think it does), I agree.

>> The question here is "does licensing under an open source license which
>> accords special privileges to a defined person X constitute a license
>> 'under the same terms' in respect of X and Y?". If the answer to that
>> is
>> "no", the same applies to (say) the MPL (let alone the QPL) which
>> gives the
>> ID distinct privileges (but the BSD mod above fixes it to a great
>> extent
>> for the OVPL). If the answer is "yes", then the existing OVPL is not
>> problematic in respect of OSD #3 anyway.
>
> Based on the above argumentation, the QPL does not comply with OSD terms
> 3 and 5. I don't think the MPL qualifies as precedence in this regard.
> But certainly, either the approval of the QPL was a mistake or it is
> precedence for approval of the OVPL. It does come back to that, of course.

The MPL and (I think) the CDDL (I haven't looked for others) materially
differentiate between the ID and subsequent contributors - see past
messages. Granted, not so much as the QPL or the OVPL, but nevertheless
they do.

Alex