[License-review] Approval: Server Side Public License, Version 1 (SSPL v1)

Kyle Mitchell kyle at kemitchell.com
Sun Nov 11 07:05:59 UTC 2018

On 2018-11-10 01:11, Rick Moen wrote:
> Quoting Kyle Mitchell (kyle at kemitchell.com):
> > Apart from its position as license steward, of what particular
> > relevance is MongoDB, Inc., among licensors?  Was AGPL approved
> > because reviewers approved of Affero?  Because they approved of
> > Funambol?
> Counselor, you would perhaps have a point if the witness had alleged
> that the merit of a licence should be judged according ot the merits of
> the proposer, but the witness did not say that, so you can park the
> rhetoric.

Bradley criticized the submission by alleging that MongoDB
intends to "abuse the copyleft system" with the license, not
itself being subject to its terms, and so on.  You quoted
and seconded.  That's why I responded as I did.

Maybe we're in violent agreement?  I wouldn't stop anyone
responding to the submitter's comments on intent, or even
what we know about them, as a project or a company.  I have.
But those points aren't relevant to OSD conformance or
advisability of the license on their own.

> Mr. Mitchell, I have very limited time and patience for refuting your
> ankle-biting on small irrelevant points while sidestep my large ones
> like SSPL's enormous OSD #9 problem.  I believe I made my point clearly
> enough; I stand by it.

I don't mean to antagonize.  I hope that you don't, either.

I have no doubt that you read OSD 9 as you say.  But your
reading, your reasons, and others' disagreement goes to
exactly the still larger point I made up-thread:  OSD 9, and
especially what _isn't_ written in it, is not a helpful
frame for discussion when we don't agree.  Which we don't.
And not just you and I.


The "oldest trick in the book" you describe is using
copyleft to exclude competitors who won't release their own
code as free/open.  That isn't a trick.  It's the
functionality of copyleft.

RMS' writings describe copyleft as a means of excluding
nonfree developers from free work, both as an enticement to
work free, and to give free developers competitive advantage
over nonfree.  In other words, strong copyleft terms tell
nonfree developers to go away.  As Bradley and I have
discussed, use of FSF-drafted copyleft licenses for
commercial dual licensing predates OSI.  As of 2010, RMS did
not oppose "selling exceptions".


In other words, it's acceptable to use copyleft terms to
tell nonfree developers they can buy proprietary licenses,
instead of just going away.

OSI approved Sleepycat, QFPL, and RPL specifically for
commercial dual-licensing use.  Funambol, which submitted
AGPLv3, originally wrote its own network-copyleft GPL
variant, and published blog posts and a conference paper on
dual licensing.  Bradley and I agree that activist copyleft
has been dual-use in this way for a very long time.

Neither the potential for dual licensing under a submitted
license, nor obvious submitter intent to do so, has stopped
OSI approving stronger copyleft licenses.  I think we all
agree that where SSPLv1 applies copyleft, that copyleft is
stronger than what OSI has approved so far.  Why is that
strength problematic?  OSD 9 doesn't say, and we shouldn't
pretend that it does.

Kyle Mitchell, attorney // Oakland // (510) 712 - 0933

More information about the License-review mailing list