Official License Anti-Proliferation policy?

Chuck Swiger chuck at
Tue Apr 12 05:02:40 UTC 2005

Evan Prodromou wrote:
> Wow. What a horrible mistake! I wish the board had read the very long,
> involved discussion here on license-discuss before taking this step,
> since there was much more opposition to anti-proliferation efforts than
> support for them. It seems to have been largely ignored.

Hmm.  I agree with you with regard to the "(a) non-duplicative" clause.

I am in favor of "(b) clear and understandable", and "(c) reusable"; although 
only the latter condition is significant.  It would help everybody if software 
licenses could be reused by lots of other people as-is, without having to be 
tweaked and modified for local jurisdiction case-law, org name, or whatnot.

Considering the alternative is what? being "opaque and incomprehensible", (b) 
seems to be wonderful criterion, if a little superfluous due to being blindly 
obvious.  C'mon, if nobody among the legal experts lurking here (or should 
that be: "a constellation of legal experts brightening the murky recesses of 
software copyright law"? :-) can figure out what a prospective license means, 
do we really need to have a rule saying that we shouldn't approve it until the 
license gets fixed enough to make sense...?

Of course, apparently "we" means "the OSI board", in this case, not "we" as in 
"everybody on this mailing list" or "everybody who writes and shares 
software", or even "everybody who _uses_ shared software".

My observation has been that few people start with "we" as in meaning 
"everybody, anywhere" (unfortunately), but considering human limitations, IMO 
the OSI board is still significantly above average.  One of the definitions of 
the political process is being willing to compromise certain principles for 
the sake of expedient compromise.  If the end result benefits the core 
principles, then perhaps the tradeoff is agreeable, after all.

> I think the whole misunderstanding is founded on this incorrect
> statement:
>         License proliferation has become a significant barrier to
>         open-source deployment.
> License _incompatibility_ is a barrier for deployment, but that's the
> case whether you have 4 or 40 approved licenses. There is absolutely no
> evidence that license proliferation is hurting open source deployment,
> nor even that we're seeing more license incompatibility problems, or
> problems on a different scale, in 2005 than in 1995.

I mostly agree with this.  The proliferation of simple, permissive licenses 
constitutes no significant burden upon anyone.  For example, I believe there 
are 36 different variants of the BSD license (and the CMU license, and the 
UMich license, and others) just in libm, not that anyone probably notices any 
difference in practice, unless they look.  Or unless somebody removes author's 
names, which has unfortunately happened from time to time....

Where I might disagree is-- to the extent that having 100 OSI-approved 
licenses (with maybe 20 is significant circulation), is a genuine problem 
compared with ~1000 commercial software licenses sitting in the CD cases on 
the wall behind me, or the 10K? licenses probably involved when I'm at work-- 
is: the proliferation of reciprocal licenses seems to be causing most of the 
"problem", simply because it is more complicated to combine such licenses.

It will be very interesting to see how the GPLv3 addresses this issue, if it 
does, and whether the net result will help make using GPL plus "another 
OSI-approved license" (tm) easier to do.  Obviously, GPL + AOSIAL[1] is likely 
to be a common case, although combinations of two other licenses are certainly 
possible, as well.

> I think this was a really, really bad smoke-filled room decision. I'm
> really sorry that the OSI board can't think of anything better to do
> with their time than this. Is this _really_ the number one problem
> facing Open Source in 2005? I think not.

You're begging the question, here?
Very well, what IYHO is "the #1 problem facing OS2005" (tm)...?


You exaggerate a bit: they are merely becoming political and bureaocratic.

> ~Evan Prodromou
> P.S. I guess at least we've saved Martin Fink.

There's nothing wrong with helping Martin Fink.

However, I'm not convinced this change is going to honestly make much 
difference-- it's not clear to me that the problem is being clearly defined, 
much less addressed in a useful fashion by the changes just made.


[1]: "another OSI-approved license", or perhaps "any OSI-approved license".

More information about the License-discuss mailing list