OSI-approved license that assigns contributor copyright to me

Alex Bligh alex at alex.org.uk
Wed Jul 13 13:31:01 UTC 2005


David,

> b) It's really frickin' long and complicated.
>
> (b) is the bigger hangup of mine.  I feel like I'm reading an obfuscated
> C contest entry and my eyes glaze over after like the sixth page.
> Granted, I think the GPL is way too long too, but at least the GPL has so
> much momentum that it's easier to say "well, even if I don't understand
> it, people who I trust do".  But the OVPL has no such momentum, so it's a
> bit scary to release my source code under the license.  I mean, how can I
> explain to my users what it means if I personally don't even know?

Well, remember most of it is the CDDL, which has had a fair amount
of review, and that itself is based on the MPL, which has had a lot
of review. However, to address your criticisms:

> Specifically, I *think* the portion of the license relevant to this
> discussion is the first sentence of section 3.3.  However, that sentence
> is *30 lines long*, and even after several readings I don't see how it
> accomplishes anything at all.  At the very least, it doesn't appear to
> grant me any additional rights beyond what anybody else has.  At best it
> says I can "sublicense" the contribution, but this right is already
> granted to everyone under section 2.1.a.  Furthermore, I don't see the
> scope of the word "sublicense" defined anywhere, so it's not at all clear
> what prevents anyone else from "sublicensing" my code into a closed,
> proprietary fork.

2.1, and its mirror for contributions which is more relevant, i.e.
2.2 gives everyone (i.e. "You" in the context of them being a licensee),
rights to use (etc.) your software, but the important phrase at the
head is "Conditioned on Your compliance with Section 3 below". IE
the license given by 2.1 is only valid if you fulfil the obligations
of section 3. Amongst other things, under 3.1, you have to make
source code available "under the terms of this license". So if anyone
else relies on a section 2 grant, they MUST make source code for the
derived work available under the terms of the license. This is how
the CDDL works.

Section 3.3 gives an additional separate license grant to the initial
developer. It is specifically ADDITIONAL to the section 2 grant, and
thus not conditioned upon the ID's compliance. However, it is
conditioned upon the compliance with the second half of 3.3
(after the capitalized bit).

The easiest way to look at it is to understand the CDDL first, then
understand 3.3.

> Granted, I bet it does and I'm just not reading it right.  But even if
> you explain to me how it does work, my point remains -- if I'm not smart
> enough to see it after several readings, I can't expect my contributors
> to fare any better when they have much less incentive to understand.

Fair point

> For your and others' benefit, I've copied section 3.3 below:
>
> -------------------------------------------------------------------
> 3.3.      Additional License of Modifications to Initial Developer.

We could halve the size of this section by deleting all the stuff starting
"BUT SUCH GRANT SHALL BE AND SHALL REMAIN CONDITIONAL UPON" and the end of
that sentence (i.e. "if any"). It then reads exactly like the normal grant
clause. The remainder just talks about when the grant becomes effective.

I've thought about doing this several times. The idea of the second bit is
to give some protection to the rest of the community. What it says (in
English) is if the ID uses stuff in his proprietary version, he must make
the bits he uses publicly available in source form, for as long as he uses
them.

Personally, I don't think this actually gives a huge amount of additional
comfort, because the licensed bits are ALREADY licensed under the
terms of this license so the only difference is whether people can
find them. Also, the ID may *in any case* request a (separate) license-back
or assignment for large contributions, which removes the section 3.3
obligation in its entirety. However, we left it in because we thought
it was right in spirit. If it makes it that unreadable, I'd vote
for taking it out. That would knock 10 lines off.

> So I trust that we have the same goals, and I even trust that your OVPL
> accomplishes those goals (merely because I'm a trusting person and you
> seem like a nice guy).

Well, we've put what our goals are up in English on
  http://www.openvendor.org/
I'm not sure that yours are exactly the same as ours, but it looks like
what we're trying to do.

>  But my trust is insufficient; it's my
> contributor's trust I need.  And I don't see how this license will win my
> contributor's trust due to its long, verbose, and complicated structure.
> It might succeed in satisfying the lawyers, but I fear it does so at the
> expense of scaring away small-time contributors that havent't the
> resources nor interest to hire a laywer.

We recognize that danger - hence the web site and its attempts to
explain it, however...

> So with all that in mind, have you any interest in creating an OVPL-lite
> license, designed for the little guy?

...no, because I'd rather base it on a solid legal framework I understand
(i.e. CDDL) - HOWEVER I am very interested in changes to the OVPL
which reduce the number and size of modifications we've made to the
(relatively well understood CDDL). If people are happier with me removing
about a third of 3.3, that sounds like a good change to make to me.
It certainly makes it a lot more comprehensible.

Alex



More information about the License-discuss mailing list