[License-review] License drafting quality and process [was Re: Comment on MOSL and similar licenses]

Josh Berkus josh at postgresql.org
Wed Jun 12 17:56:17 UTC 2013

> This was not at all my idea, if you read me correctly: I suggested the
> rapid deliverance of an "OSD compliance certificate" (even if the license
> does not improve / revolutionate the FLOSS ecosystem). The certificate will
> be the evidence that the licence is open source (facilitating the adoption
> by the first followers).
> OSI may then take more time to classify eventually the license as
> "OSI-approved" once a representative level of projects will use it (as the
> case may be).

What problem are you trying to solve here?  It doesn't seem like a
problem we actually have.  I'll also point out that popularity is no
guarantee of validity.  For example, the "Beerware" license is extremely
popular, but is completely unenforceable.

First, let me agree with Rick and point out that the only licenses we've
approved in the last year were new versions of old approved licenses.
Every other proposed license we've seen has been rejected as either (a)
not OSD-compliant, or (b) very poorly conceived and written, or (c)
wholly duplicative of an existing license.  As far as I'm concerned,
that part of the process is working fine.

What's not working fine is how people *feel* about the rejection
process.  What happens currently in the worst case is:

(1) Crayon license submitter (CLS) emails their "clever" idea for a new
license to this list.

(2) We point out some of the shortcomings in the license and reject it.

(3) CLS becomes indignant and accuses us of personal enmity and/or

(4) CLS attempts to lobby and/or harass us into submission.  This
doesn't work.

(5) We are contemptuous of CLS.  CLS hates us.

(6) CLS decides to use their "license" anyway.

My thinking was that an explicit process where the submitter *had* to
consult an attorney before submitting ... and where the only reply we'd
give to one who didn't was to point to that requirement ... would cause
the CLS to talk to an attorney first who would at least point out that
the license they're sending is badly written and unenforceable.  Or,
more likely, get hung up on finding an attorney and never write their
crayon license in the first place.

However, I think enough people on here have pointed out that CLS's are
ignoring any kind of reasonable process anyway, so adding process
requirements won't work; they'll simply ignore them.

In a way, this is actually the process working as designed.  As the
number of possible new licenses which are OSD-compliant, original and
sensible approaches the vanishing point, increasingly the only
submissions we're going to see are going to be from the ill-informed and
impulsive.  So maybe there's no fixable problem here, at all.

--Josh Berkus

More information about the License-review mailing list