Mailing list for all users of the OCaml language and system.
 help / color / mirror / Atom feed
From: Sven Luther <sven.luther@wanadoo.fr>
To: Jacques Garrigue <garrigue@math.nagoya-u.ac.jp>
Cc: sven.luther@wanadoo.fr, caml-list@inria.fr
Subject: Re: [Caml-list] Ocaml license - why not GPL?
Date: Mon, 31 Jan 2005 08:35:45 +0100	[thread overview]
Message-ID: <20050131073544.GB19902@pegasos> (raw)
In-Reply-To: <20050131.095711.27629180.garrigue@math.nagoya-u.ac.jp>

On Mon, Jan 31, 2005 at 09:57:11AM +0900, Jacques Garrigue wrote:
> From: Sven Luther <sven.luther@wanadoo.fr>
> 
> > > The QPL is an official open-source license.
> > 
> > Well, it seems that this same fact is highly disputed, and trolltech did in
> > fact dual licence Qt under the GPL too. The current ocaml licence was modified
> > from plain QPL though recently, after 2-4 week of intense flamewar on
> > debian-legal, and there are some clause yet in it which where subject to
> > discussion.
> 
> Well, I don't know who disputed this, at least it is an OSI recognized
> open-source license, which I believe to be the definition of
> "officially open-source".

There is a 200+ thread on debian-legal i can point you too if you like. If i
remember well, the problematic clauses where :

  1) the choice of venue clause is probably illegal in most countries, since
  you have to sue in the court where the defendant lives or does business, and
  imposing Paris is not accepatble.

  2) the clause 6c, which causes some burdensome constraint on developer of
  third party product linking with the code. The clause is also confuse and
  difficult to intepret, meaning that any third party coder has to keep the
  code version around forever waiting for the whim of original authors. It
  also forbids private releases of third party code, which altough not
  mentioned in the DFSG is considered as one of the elemental freedoms. Notice
  that matters are additionally confused by the fact that clause 6 speaks
  about linking, which applies to works using the library (well, the compiler
  in the ocaml case, but Qt was designed with libraries in mind), and not
  modified works, altough the differenciation between both is not really the
  same.

These two problematic clauses where dropped by Xavier and the ocaml team with
3.08.2, so there is no problem. The other problematic clause, which was under
discussion, but not considered a show-stopper, or rather, i used the preffered
debian-legal tactic of lengthening the thread until everyone got bored, and
only a vocal minority remained to not-be-happy about this once the two other
points were solved is :

  3) Clause 3b, which imposes a dissymetry of rights between the ocaml
  developers and other developers reusing code. The reused code is made
  available for re-usage in ocaml, under whatever licence the ocaml team
  choses in addition to the QPL, but this is not true the other way around.
  Notice that a strict interpretation, where one argues that if work B is
  linked with work A, then both B is a derivative work of work A, but also you
  can consider work A as a derivative of work B, once work A include code from
  work B. This is especially dubious once the amount of code in work B is
  greater than the amount of code in work A. This kind of interpretation would
  allow anyone whose code has been integrated in ocaml to relicence ocaml
  under whatever licence, altough i believe this will lead to a lengthy court
  case, and the judge in the end will consider the relative importance of both
  code bases, and maybe some chronology of the code bases. IANAL though.

Furthermore, the QPL is also non-free, which means aside from the choice of
venue/law, it cannot be modified, which means the ocaml patch has been forced
to do some patching of the licence to solve the above issues.

As you see, the matter is far from clean, and i believe that the confrontation
was only dropped because Trolltech dual-licenced Qt, and the matter mostly
became moot. There is as said, a controversy on the QPL, and many person
consider the decision to admit it as free a mistake that may be returned upon
if there is enough incentive. There is a pointer to that discussion in the
debian-legal thread, but i don't feel like digging it out.

> To the best of my knowledge, Trolltech decided to release a GPL'ed
> version of Qt because the QPL is not compatible with the GPL, which is
> no surprise at all: by definition the GPL is only compatible with
> itself, or strictly weaker licenses, like BSD or LGPL.

Like said above, the issue is not all that clear, and was only dropped because
of the dual-licencing issue. Also the incompatibilities is a bit more subtle
than plain GPL subsumption of weaker licences. The above analysis is fully
orthogonal to the GPL compatibility issue.

> This is less of a problem with ocaml, because the part covered by the
> QPL is the compiler, not the library, so this should only concern
> persons hacking the compiler itself.

Yes, and no, there are some part of the compiler, like the ocamlmktop
generated stuff, or whatever is used by some parts of camlp4 or the cameleon
stuff, or i don't remember exactly, but those that build-depend on the
ocaml-source package in debian, which have a vocation of being reusable and
are covered by the QPL.

Hope this clarifies things a bit, and sorry for the long post, legal issues
are usually bothersome to everyone involved and i was only dragged into those
because of the threat of ocaml having to be moved in non-free by debian.

Friendly,

Sven Luther


  parent reply	other threads:[~2005-01-31  7:38 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-01-28 16:47 Jozef Kosoru
2005-01-28 17:08 ` [Caml-list] " David Brown
2005-01-28 19:09   ` Jon Harrop
2005-01-29 23:46     ` Stefan Monnier
2005-01-28 17:14 ` [Caml-list] " Luc Maranget
2005-01-29  6:05 ` Jacques Garrigue
2005-01-30  6:22   ` Sven Luther
2005-01-31  0:57     ` Jacques Garrigue
2005-01-31  7:03       ` Alex Baretta
2005-01-31  7:38         ` Sven Luther
2005-01-31  7:47           ` Alex Baretta
2005-01-31  7:59             ` Sven Luther
2005-01-31  9:09               ` skaller
2005-01-31  9:23                 ` Sven Luther
2005-01-31  9:00             ` Xavier Leroy
2005-01-31  9:15               ` Sven Luther
2005-01-31  9:47               ` Nicolas Cannasse
2005-01-31 10:29                 ` Xavier Leroy
2005-01-31 10:41                   ` Nicolas Cannasse
2005-01-31 11:08                 ` Alex Baretta
2005-01-31 11:49                   ` Alex Baretta
2005-01-31  7:35       ` Sven Luther [this message]
2005-01-30  6:18 ` Sven Luther

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20050131073544.GB19902@pegasos \
    --to=sven.luther@wanadoo.fr \
    --cc=caml-list@inria.fr \
    --cc=garrigue@math.nagoya-u.ac.jp \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox