[License-review] Submission of the Upstream Compatibility License v1.0 (UCL-1.0) for approval

Josh berkus josh at postgresql.org
Tue Oct 25 17:29:55 UTC 2016


On 10/25/2016 09:56 AM, Nigel T wrote:
>  Then again, the copyright conditions are asymmetric, thus I
> cannot get rid of that blinking red light on my dashboard.

But as Nigel points out, this is really no different than copyleft plus
a CLA in its asymmetric effects.   While I'm hardly the biggest fan of
dual-licensing (or copyleft), it's a reality for a lot of the software
world right now.

So, let's take my suggestion for simplifying the UPL.  That is, instead
of its complicated contribute-back language, you have a license where:

1. the original core code from the named party is licensed UPL, which is
copyleft.

2. any additions to this core code must be licensed Apache 2.0, and
documented.

While hardly ideal, from a *developer* perspective this is a big
improvement over CLA+Dual, which is the common pattern.  I don't have to
fill out a CLA (something which blocks contributors from many countries
and employers).  I'm guaranteed that the collected work from public
contributors will be available under a mix of Apache + UPL, even if the
original company goes away.  And the company has more of an incentive to
release everything it can under the UPL because of the limited risk in
doing so.  Finally if I want to grab just one of the code additions
without using the core code, it's Apache.

There are basically two drawbacks I can see to the revised UPL as a
developer:

A. 15 years from now when the original company is gone or acquired and
the code base has changed beyond recognition we'll still have funky
composite licensing.

B. The asymmetry is proof that this project is never going to be a
community (which is also true of dual-licensing).

Now, I can't speak to what's *legally* allowable; can one license
require that additions to its copyrighted material be licensed under a
different license?  I'm not an attorney.  But I don't see an inherent
issue to a simplified version of this license.

(I wouldn't support the license as-is because the language of it is
baffling)

--Josh Berkus






More information about the License-review mailing list