[conspire] (forw) Re: Can OSI specify that public domain is open source?

Thu Sep 8 14:23:31 PDT 2011

Chad is a partisan of 'copyfree' licensing, aka ultri-permissive 
licensing for non-software content (yet another obscure neologism;
yippee).  He posts to OSI's licence-discuss mailing list, and somehow
suckered me into a side-discussion in e-mail about his beef with people
like me (who dislike licence proliferation, especially of pointless and
badly written permissive licences) and my beef with people like him (who 
write such licences, such as Chad's Very Own, which is called Open Works
License).

----- Forwarded message from Rick Moen <rick at linuxmafia.com> -----

Date: Thu, 8 Sep 2011 13:52:10 -0700
From: Rick Moen <rick at linuxmafia.com>
To: Chad Perrin <perrin at apotheon.com>
Subject: Re: Can OSI specify that public domain is open source?
Organization: If you lived here, you'd be $HOME already.

Quoting Chad Perrin (perrin at apotheon.com):

> What you said previously seemed to draw some kind of equivalency between
> the public domain and your complaint about the warranty disclaimer in the
> OWL.

Apologies if I was unclear.  As noted, I was trying to fit
correspondence into the middle of a busy day.

> I find it a bit annoying too -- but far less so for permissive licenses
> than restrictive licenses, because with the former at least the law
> doesn't tend to prohibit combining content from two works.

A valid point.  I can't think of any permissive licences that are
immiscible.

> If a license does something good that no other license does, that seems
> important.

Granted.  It really needs, to get any traction in adoption, to be enough
better / distinctive to outweigh the inherent burden any new licence
needs to bear (various things including just reading and getting to know
another darned licence).  

And -- personal opinion, but widely held -- there just aren't very many
significant free software _or_ free non-software licences, with
'significance' judged by your choice of metric, any of which will
necessarily involve some handwaving but, again, the consensus is widely
held.  Write a new one, and you have a large presumption of
insignificance to overcome, and so you have to do everything right and
don't flub anything -- and it's useful to represent a large project
like, say, OpenOffice.org back in the Sun days.

People will quibble about the edges, but most would go for a list like
this:

1.  Software.

A.  Strong copyleft.

 - GPL [23] with or without Affero mods

B.  Weak copyleft

 - MPL / CDDL
 - LGPL [23]

C.  Permissive
 - 3-clause BSD
 - MIT, 2-clause BSD (FreeBSD variant) and ISC Licence -- 
   which are all functionally identical
 - Apache Licence 1.1 and 2.0

2.  Non-software

 - GFDL (though it's not much liked, and is gradually losing everywhere
   but FSF)
 - the four of six CC licences that are free licences
 - Open Database Licence

That's about it.  Artistic Licence is dumb and probably gradually going
away, despite having its more gross defects fixed with the 2.0 revision.
Design Science License, Open Content Licence, and Open Publication
Language are obscure (and the latter two are deprecated even by their
author in favour of CC licensing). Open Publication License is obsolete
and forgotten everywhere except _Linux Gazette_.  PHP Licence and OpenSSL
Licence are fossils.  APSL is a Cupertino ghetto.  CPL is an IBM ghetto.
Eclipse Public Licence is a Java ghetto.  Python Software Foundation
License is a harmless Python ghetto that's functionally the same as
3-clause BSD.  And everything else really doesn't matter at all, has no
traction, and probably never will have any.

I actually feel bad about the doomed obscurity of two really excellent
software licences crafted by founding OSI lead counsel Lawrence Rosen:

 - Academic Free Licence (AFL)
 - Open Software Licence (OSL)

These are what, respectively, a _well drafted_ copyleft and permissive
licence for software should be, dealing intelligently with patents,
warranties, and contractual assent.  They are doomed because they
arrived too late, and less well drafted licences are ubiquitous, known
quantities, and have predictable legal effects.  

AFL is _not_ software-specific.  You should have a look at it, if only
to learn from its structure.

It should also be noted that GPL is not software-specific, either.
https://www.gnu.org/philosophy/nonsoftware-copyleft.html
(However, whenever I apply GPLv2 to a non-software work, I include an
additional permission absolving recipients of the duty to include with
the work a copy of the licence text.  You should note that, if you find
one or two requirements a licence imposes on recipients, you can waive
them via a statement right under your copyright notice.)

> CC-BY is also, as I mentioned, restricted in terms of technological
> implementation and/or delivery.

Just because 4a requires, in some usage scenarios, inclusion of the URI
for CC BY (or inclusion of the full licence text)?  Just because 4b
requires, in some usage scenarios, the upstream work's URI _if_ the 
Licensor supplies one?  I don't see that as 'restricted' in any
meaningful sense of the term.

> It is, in some respects, not a permissive license at all.

The signature complaint of the permissive-licence fan is that the other
guy's permissive licence isn't permissive enough.  ;->  For them, I 
can heartily recommend my one-line or two-line licences, with the choice
depending on whether you care about being socked with warranty liability
or not.

Copyright 2011 (C) George Coder.  Do whatever you want with this work,
provided you agree that there is no warranty of any kind.

> . . . and still farther from "perfect".

You may care about perfect.  The rest of us are tired of licence
proliferation and licences that variously (1) have easily avoidable
screwups like flubbing warranty disclaimers, and (2) don't achieve
anything new that merits the time spent on them.

Also, a lot of us have noticed that laymen's notion of 'perfect' often
is completely oblivious to legal realities.  (I am not an attorney, but
I've consorted with them on software licensing long enough to learn
quite a bit.  Also, I had to study business law pretty intensively in
order to pass the CPA exam.)

> In fact, if anything, I think Creative Commons is by far the worst
> offender when it comes to license proliferation.

I have several complaints about Creative Commons, but that is not among
them.  They differ from prior content licences in being _competent_.

[About what look in even well-drafted licences like pointless
legalisms:]

> The thing that most annoys me about it is the fact that it's so difficult
> to find an explanation for *why*.

I can actually recommend, on that matter, Lawrence Rosen's book _Open
Source Licensing: Software Freedom and Intellectual Property Law_, which
is available printed in stores and also, free of charge, as a set of
PDFs (published under Rosen's permissive licence Academic Free License
version 3.0).  See:  http://www.rosenlaw.com/rosen.htm

If I recall correctly, Larry walks readers through the merits, taxonomy,
and problems of common licences, and then in one particular chapter lays
out an analysis of AFL and OSL, which I recommend to you.

----- End forwarded message -----