What is free software? [Re: Licenses and Open Source don't conflict.]

Chris Gonnerman chris.gonnerman at newcenturycomputers.net
Sat Apr 13 01:15:31 EDT 2002


----- Original Message -----
From: "François Pinard" <pinard at iro.umontreal.ca>


> > > Public domain software is [...] a
> > > special case of non-copylefted free software, [...]
>
> [Jeff Shannon]
>
> > Which *does*, in fact, mean that it is free software.  Unless perhaps
> > free software means something other than free software?
>
> There are many definitions of free software, and some people fight over
> which definition is the real one.  Some even want monopoly of freedom! :-)

There is a specific disjunction between "free software" as in the GNU def
above, and "Free Software" (with caps as shown) as used by Stallman.  He
uses the second form as a sort of synonym for GPL/LGPL programs... and he
says he's trying to avoid confusion :-)

> How I understand the GPL is that a software is not free, if it can be
> used to create a bound between parties, in such a way that one of the
> parties has lost part of its freedom on that software through that bound,
> in particular the freedom to modify the software and/or share it around.

This is essentially what Stallman finds offensive.

> A public domain software can be modified by anyone, and the result can be
> made available to someone else under a binding license.  So, that software
> was indeed free, but that quality of being "free" was totally fragile...

Exactly.  MIT (X11) and BSD licenses exhibit similar fragility.

> The GPL wants to help programs at being free in more permanent ways.

This always leads to a flamewar :-(  There are those (uniformly programmers)
who hate the GPL because it restricts their ability to "hide" the source if
they incorporate GPL'd code in their non-GPL programs.  Frankly I don't
understand the problem; if programmer Paul wants to create a proprietary
program, naturally he should avoid programmer George's GPL library.  Yelling
about it is useless, as George has just as much right to restrict those
using
his library as Paul has to hide his source.  In my opinion Paul is more
"wrong"
(or at least restrictive) than George.

Another reason often given (by the misinformed) is that GPL code can somehow
"infect" non-GPL free code (BSD let's say); it goes something like this:
Bob
writes an excellent library and releases it under the BSD license.  George
has
his previously mentioned GPL'd library, which is also most cool.  Fred comes
along, and writes a program which uses both Bob's and George's libraries,
and
for some reason Fred actually incorporates the library source in his project
rather than expecting you to separately install it.  Fred must release his
program under the GPL (he has no choice in this matter), but this has *no
effect* on the license of Bob's library.  Of course, as a GPL product Fred
is
required to release the code he got from Bob even though Bob didn't require
that.

The assertion is that Bob's library source has been "infected" by the GPL,
but
unless Bob agrees to this, Fred has no right to change the license.  Again,
the
GPL covering Fred's code *forces* him to reveal Bob's code, even though Bob
doesn't require him to do so.  Extracting just Bob's code may be easy or it
may
be hard, but if Bob has written his comments well, a third (fourth?) party
should be able to find Bob's source on the Web and avoid the middleman.

But as I said, this always leads to a flamewar... and here it comes, I'm
sure.

Chris Gonnerman -- chris.gonnerman at newcenturycomputers.net
http://newcenturycomputers.net






More information about the Python-list mailing list