Re: ANN: Fuzzy machine learning framework v1.2

[Ludovic Brenta <ludovic@ludovic-brenta.org>]

"Dmitry A. Kazakov" writes on comp.lang.ada:
> On Mon, 28 May 2012 21:23:09 +0200, Ludovic Brenta wrote:
>
>> "Dmitry A. Kazakov" writes on comp.lang.ada:
>>> On Mon, 28 May 2012 13:36:53 +0200, J-P. Rosen wrote:
>>>> Le 28/05/2012 12:08, Dmitry A. Kazakov a écrit :
>>>>> This release is packaged for Windows, Fedora (yum) and Debian
>>>>> (apt). The software is public domain (licensed under GM GPL).
>>>> Just to be picky: if it's licensed under GMGPL, it's free software,
>>>> but it's not public domain. Public domain means that you give up
>>>> all your intellectual rights (possibly allowing someone to take
>>>> your work and putting a restrictive license upon it).
>>>
>>> I thought GMGPL allowed derivative works to be restricted as anybody
>>> wanted?
>> 
>> The GMGPL or, since GCC 4.4, GPL3 with Runtime Library Exception,
>> allows people to distribute *binaries* including (aka linked with)
>> your libraries under whatever license they wish.  They cannot,
>> however, redistribute the *sources* of your library under whatever
>> license they wish.  If they choose to redistribute the sources, they
>> must do so under your original license terms and with your copyright.
>
> Why anybody wanted to distribute same stuff under a different license?

That's the whole point.  Some people want link your libraries into their
programs and sell their programs including your libraries for a profit
without paying you or acknowledging your contribution.  i.e. they want
to sell your libraries under a proprietary license of their choosing,
different from the license you chose when publishing your libraries.

> Licensing should encourage new work, public or proprietary, no
> matter. Or I again missed something?

Everyone in the free and open source software movement agrees with these
goals, the disagreement is only about the means to achieve them.

The FSF school of thought states that, to encourage sharing,
redistribution and new work based on free software, it is necessary that
free software remain free.  On other words, taking away the Four
Freedoms granted by free software must be forbidden.  Here, "sharing and
new work" is understood in an academic way: to further the state of the
art and make new breakthroughs.

The other school of thought states that permissive licenses such as BSD
and MIT encourage sharing and new work much better by allowing Big
Proprietary Corporations to use free software in their proprietary
products, even if said Corporations change the license along the way to
make the software proprietary.  Here, "sharing and new work" is
understood in an industrial way: to allow more software to be sold (not
necessarily by the author of the libraries, though).

The GPL with Runtime Library Exception is a middle ground: the original
author retains control on source redistributions and license thereof,
but allows binary redistribution, as part of larger works, under any
license outside the control of the original author.  Here again, you can
see the middle ground in two different ways.  One is to say: since Big
Proprietary Corporations are going to sell proprietary software anyway,
let's allow them to run this software on a Free Software operating
system and libraries, so we can displace proprietary operating systems
and libraries.  The other is to say: let's allow Big Proprietary
Corporations to sell even more proprietary software by giving them our
libraries for nothing.

You alone can decide on the license that best meets your particular
goals.

-- 
Ludovic Brenta.