Re: [Fsfe-uk] Stalls at Marxism 2003

[Chris Croughton]

On Mon, Jun 30, 2003 at 01:28:57PM +0100, Alex Hudson wrote:

> On Mon, Jun 30, 2003 at 12:34:10PM +0100, Chris Croughton wrote:
> > > Tim's complied, but feels, IIRC, it's a bit hypocritical.
> > 
> > It is indeed.  But then that fits with RMS/FSF's view of 'free'
> > software, which they don't allow to be limited in distribution or
> > use (for instance by saying "not permitted for military use") but
> > they do allow to be limited by being incompatible with other free
> > licences (for instance if you have to send modifications back to the
> > creator of the software).
> 
> If you have to send mods back to the creator, it's not free software.

By the FSF/RMS's definition.  But by my definition of 'free' saying "you
can't link my library with yours unless you do what I say" isn't free
either, that's my point.  Both are arbitrary conditions that someone
else's code has to obey in order to get a licence to use the code, and
if you have to jump through hoops, any hoops, just to re-use software
then that software is not "free in the sense of freedom".  And we see
the same hypocrisy with the FSF documentation (and Debian for example
does not like the FSF's so-called "free documentation" licence, it's not
even as free as the GPL).

I use a modified BSD licence which says, basically, "you can do what you
want with your own code, just ensure that /mine/ stays free".  No
conditions on what you link with it, or even what you compile it into,
the only conditions are "say that it's mine, say if you changed it" (and
I regard distributing my source with your project to be enough of a
credit, I don't insist that you put anything in documentation of banner
text unlike the ).

> The RMS/FSF 'view' of that is quite clear I think. Look at the
> reasoning why the OPL is not considered a free software licence. Being
> 'incompatible' with non-free licences is the nature of the FSF, I'm
> afraid.

And it makes it "non-free", in the sense of freedom.  If I am not
allowed to link my code with both a GPL library and an OPL one, that is
limiting my freedom.  And since even looking at the source code and then
writing a version which does the same can be (and has been) regarded as
creating a "derivative work", this means that I can't legally even write
my own version without infringing the FSF's licence if I have used my
'freedom' to study their source code.

> IIRC, the problem RMS has with O'Reilly is that they would just be
> publishing the books without actually putting any revenue back the
> author's way. I don't see what is wrong by stating a preference for
> people buying from a publisher like GNU who will do that. 

It's supposedly free software, by its own licence it can be sold or
given away by anyone, and by RMS's own statements the only profit they
can expect to make from it is by selling their services.  If O'Reilly
(or anyone else) provides a better service than the FSF (and O'Reilly
books are a lot more available in the UK than FSF ones as far as I've
seen) then that's according to RMS's own stated principles.  He said
himself that "We can also encourage commercial publishers to sell free,
copylefted manuals instead of proprietary ones" (his essay "Free
Software and Free Manuals" on www.gnu.org).

> O'Reilly have their own cause at heart. Nothing wrong with that, but
> RMS also has his own views and I don't think there's anything wrong
> with that either.  I don't think there's anything wrong with stating
> your preference, either.

It's not a 'preference', at least as O'Reilly tells it, it was a direct
request to not publish.  Not quite an order or forbidding a licence, no,
but more than just a 'preference'.  And by limiting it at source RMS has
taken away the very choice that he claims to support, rather than saying
(as he did with Linux versus the Hurd) that he prefers users to use the
FSF version but supports the choice.

Chris C