[License-discuss] Any Free License, an open source license

Fri Nov 13 22:06:07 UTC 2015

Quoting Christopher Allan Webber (cwebber at dustycloud.org):

Congratulations on coming up with a novel licensing concept.  
Pretty wild, but definitely novel.  That is a rare accomplishment.

>   # <PROGRAM NAME> -- (C) <YEAR> <AUTHOR NAME>
>   # Released under the "Any Free License 2015-11-05", whose terms
>   # are the following:
>   #   This code is released under any of the free software licenses listed on
>   #     https://www.gnu.org/licenses/license-list.html
>   #   which for archival purposes is
>   #     https://web.archive.org/web/20151105070140/http://www.gnu.org/licenses/license-list.html

I note without special objection that this means the licence terms are
indeterminate, or, to put it a different way.

> You may notice that this license has a convenient upgrade procedure.

;->

(Actually, you probably mean, more specifically, that licensee may
select among any of the licences that _were_ on the FSF list as of the
date of the cited archive.org archive link.  But your draft text is just
a bit unclear on that.)

> I suggest rolling releases of the license over time, updating to the
> latest link available from:
> 
>   https://web.archive.org/web/*/https://www.gnu.org/licenses/license-list.html

Whereas _this_ version of the suggestion definitely expands to
indeterminate licensing terms.

> If the OSI is not comfortable trusting the FSF license list at a
> particular date, one possible procedure in the rolling release system is
> that periodically, the OSI can review the FSF list, and make new
> releases based on verified and approved copies of the FSF license list.

Yes, quite.  I'll return to that point in a moment.

I don't speak for OSI, but suspect no one here would question FSF's
commitment to free software or would have more than _some_ worries about its
ability to assess licences (examples to follow, below), but OSI would be
remiss in just outsourcing OSI Certified assessments to FSF or anyone
else.

> I would like to submit this license for consideration and inclusion on
> the OSI license review list, but am seeking feedback first.

OK, first question, why bother?  Cui bono?

Your meta-licence offers a choice among licences that are clearly (par
excellence) free-software ones, and also are either explicitly OSI
Certified or most probably would pass certification if submitted
(e.g., GNU All-Permissive License).  But I spot several odd entries on
a quick review:

o 'public domain', is problematic is part for reasons FSF
  cites and also for reasons OSI has FAQed.   And it's not a licence.
o  Unlicence: incompetently drafted and doesn't have the intended
  effect; see below.
o  WTFL:  incompetently drafted and doesn't have the intended 
  effect; see below.
o  Informal license:  This is a vague entry about a concept, and
  not a licence at all. 

CC0 is _likely_ to pass certification if it were re-submitted (having
been withdrawn by submitter during prior consideration that raised
concerns over its patent-grant-explicitly-not-implied language.  My
opinion, anyway; views may differ.

My prior comments about Unlicense:

  Its first sentence professes to put the covered work into the public
  domain.  However, then the second sentence professes to grant reserved
  rights under copyright law.  However, who is granting those rights, the
  erstwhile copyright holder who, one sentence earlier, professed to
  destroy his or her own title?

  By contrast, CC0 states explicitly that the current copyright holder 
  is attempting (I paraphrase) to the extent permitted by local law to
  disavow in perpetuity and on behalf of all successors all reserved
  rights, and _if that is locally unsuccessful_ grants a permissive
  licence under his/her powers as copyright owner.

  I realise there are a whole lot of software engineers out there who'd
  like to handwave copyright law out of their lives (including you), but
  it'd be really nice if they'd occasionally bother to consult suitable
  legal help before shooting themselves and others in the foot.  
  [...]

  Paragraph (and sentence) #1 professes to put the covered work into the
  public domain.  As mentioned, paragraph 2 professes to be a grant of
  rights normally reserved by default to a copyright owner, which makes no
  sense given that the preceding sentence professed to eradicate the
  work's quality of being ownable.  _However_ (upon reflection), in itself
  that would be harmless if redundant and pointless:  One can interpret
  paragraph 2 as an elaboration of the consequences of the first
  paragraph.

  Paragraph 3 is mostly further explanation of the concept of public
  domain, and therefore harmless if not useful.  Its middle sentence 
  elaborates that the erstwhile author aims to bind heirs and successors,
  too (which is a logical inclusion, irrespective of whether it works).

  Paragraph 4, though, is the one that would be amusing if it weren't
  tragically broken:  It's the warranty disclaimer.  People accepting the
  covered work are obliged to accept the condition of no warranty,
  otherwise there is no licence.  Except, oh, wait:  Paragraph 1 professed
  to put the work in the public domain, so the erstwhile owner has sawed
  off and evaporated in paragraph 1 all power to require the condition in
  paragraph 4.

  And, my larger point is:  This sort of low comedy happens just about
  every freakin' time software people attempt to engineer a way to nullify
  copyright law by copying and pasting from existing licences,
  non-licences, and semi-licences.  Bendiken, for example, says that he
  copied and pasted SQLite's copyright waiver and MIT/X11's warranty
  disclaimer, and voila!  Unlicense.

  Text salad from popular projects:  What could possibly go wrong?  

  Bruce calls these crayon licences.  I tend to use the phrase 
  'cargo-cult law'.   
  (https://secure.wikimedia.org/wikipedia/en/wiki/Cargo_cult#Metaphorical_uses_of_the_term)

My prior comments about WTFPL are similar:

  WTFPL v. 2 (latest) is so badly written it grants rights only to the
  _licence_ itself, and not to any ostensibly covered work.  (Read it.)

  Noting that it leaves warranty liability intact (probably accidentally)
  seems beside the point, in comparison.  

  It's an object lesson in why coders should not attempt to draft what are
  often on this mailing list termed 'crayon licences'.

  A broader point:  The quest for the shortest possible licence (of
  whatever category) strikes me as solving the wrong problem.  If your
  problem is that you're dealing with people having difficulty contending
  with the reality of a worldwide copyright regime and trying to wish it
  out of their lives, maybe overcoming that lack of reality orientation
  ought to be your task.  (My opinion, yours for a small fee and waiver of
  reverse-engineering rights.)

When I say 'why bother', I mean, yes, I get that it might be convenient
for someone to say 'What the heck, pick any from this smorgasbord.'  But
that's asking OSI's licence reviewers to do a heck of a lot of work to
permit this licensor one-stop-shopping convenience -- because the truth
is that your submission would _not_ be asking certification of just your 
language, but also certification of each and every one of the 90 free
software licences, two incompetently written and ineffective licence
(WTFPL and Unlicense), and two non-licences ('public domain' and
'Informal license').

Speaking for myself, basically:  Hell no.

But, getting back to the 'why bother', what I specifically mean is that
if you want a metalicence that OSI smiles upon, why reference _FSF's_ 
list?  Why not say...? 

  # <PROGRAM NAME> -- (C) <YEAR> <AUTHOR NAME>
  # Released under the "Any OSI Certified 2015-11-05", whose terms
  # are the following:
  #   This code is released under any of the OSI Certified[sm] licenses 
  # http://opensource.org/licenses/alphabetical ,
  #   which for archival purposes is
  # https://web.archive.org/web/20150905080406/http://opensource.org/licenses/alphabetical .

(Comma before 'which', and terminal period at the end of your sentence,
added to propitiate the Grammar Police.)

Then, instead of asking the reviewer community to pass judgement on 90
arguably reasonable free software licences, two incompetent crayon
licenses that FSF screwed up in approving, and two non-licences, you
would be referencing a list of 70 licences OSI has already certified?
(Some of those licences, IIRC, are generally considered to have been
blunders on OSI reviewers' part, but typically are very obscure and
little-used entries.)

If you make that revision, there would be really nothing to review.
But then, also equally, I would then see little or no point in OSI 
reviewing _that_ metalicence.  A choice among OSI Certified licences
obviously produces as its result an OSI Certified licence, and there's
no point in reviewing that, as there's nothing to review that hasn't
already been reviewed and approved.

-- 
Cheers,                                  "If you see a snake, just kill it. 
Rick Moen                                Don't appoint a committee on snakes."
rick at linuxmafia.com                                         -- H. Ross Perot
McQ! (4x80)