Re: [Caml-list] License Conditions for OCaml

[Sven <luther@dpt-info.u-strasbg.fr>]

On Fri, Nov 09, 2001 at 04:39:13PM -0500, Patrick M Doane wrote:
> On Fri, 9 Nov 2001, Vitaly Lugovsky wrote:
> 
> > On Thu, 8 Nov 2001, Patrick M Doane wrote:
> >
> > >   "The LGPL puts no restrictions at all on programs linked with LGPL-ed
> > > libraries. Thus, users are free to distribute (or not) OCaml-generated
> > > binaries under whatever conditions they like."
> > >
> > > >From my reading of the LGPL, which seems to correspond with the opinions
> > > of others on the list, this just isn't true.
> >
> >  Please, read it again. Carefully.
> 
> I have read it again today, carefully, and still come to the same
> conclusions:
> 
>   1. Users are allowed to reverse engineer the application
> 
>   2. Source (or possibly object files) for my application must be included
>      in the distribution.
> 
> I could probably be convinced that object files are sufficient for
> purposes of satisfying the license, but that just exposes the next major
> problem (i.e. point #1).
> 

Yes, this is the real problem for you, why not say it from the begining and
explain a bit ?

> > > If I develop an application
> > > with OCaml, I must distribute that application with source code.
> >
> >  No. You must distribute a runtime source or just put a link how to get
> > it. Nothing more. There are a lot of commercial, closed source
> > applications linked with LGPL libraries - e.g. any Linux commercial
> > apps linked with GNU Libc.
> 
> The runtime source for OCaml must be included or a "written letter" must
> be provided. That's quite a bit different than a link.

Well, you have :

    d) If distribution of the work is made by offering access to copy
    from a designated place, offer equivalent access to copy the above
    specified materials from the same place.

That is, if you offer your app as downloadeable, you just need to offer the
object code from the same place.

In the other case, you probably ship a CD or something such, and probably
acompany it by a documentation, in this case, just add a page to the
documentation to comply with the LGPL.

> You also need to include either source (or again maybe object files) of
> the application that uses the library.

No, it is the source you need to provide for the LGPLed library.

but again, either a letter or a downloadable version would be ok.

> It's not sufficient to simply ship a static executable and include a
> notice where one can get the OCaml source code.

Well, not sure, the reasoning here is that you must agree to make the object
and library source available for at least 3 years. You cannot guarantee that
the ocaml team will be doing this for the next 3 years, altough it certainly
would. Now if you were a member of the consortium, then this may be
legailistically sufficient, but again, i am no lawyer.

That said, if you modify the ocaml compiler suite to build your app, then you
have to make these modifications available, maybe it is this that is required
?

> > > This isn't acceptable for commercial development
> >
> >  It IS acceptable. But here, I think, it's offtopic. Read slashdot, and so
> > on. Look at WineX, for example: open source, commercial binaries.
> 
> I agree that for open source commercial software, there are no problems.
> This hardly represents the majority of commercial development though.

But study have said that this represent a big share of those commercial
software that will still exist in the next few years, you have been warned
:)))

> If this isn't an appropriate place to discuss issues with the Caml
> license, where else is?

Don't know, is this of interrest to the other list participants ? If yes, we
can continue here, users bored with it can easily send it to /dev/null with a
proper mail filter. If not, then we could continue privately, or create a new
list ocaml-legalese, or something such ? What do the other mailing list member
say about this, and in particular the ocaml team ?

Friendly,

Sven Luther
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr