Re: tlug: Last night's RMS discussion

["Frank Bennett (=?iso-2022-jp?B?GyRCJVUlaSVzJS8kWSVNJUMlSBsoQg==?= )" <bennett@example.com>]

On Tue, Dec 21, 1999 at 08:57:46AM -0800, Scott M. Stone wrote:
> On Tue, 21 Dec 1999, Stephen J. Turnbull wrote:
> 
> >     Scott> I'll start calling it GNU when Stallman writes his own
> >     Scott> kernel
> > 
> > He has, and Debian is about to start distribution GNU/Hurd Debian.  At 
> > least the policy manual is prepared for that.
> > 
> > rms is not my favorite person, but he has done a lot of good and you
> > should oppose him where he's wrong, not where he's arguably correct.
> 
> "arguably" is the key word here -- Linux isn't all GNU, therefore it
> shouldn't be called GNU.  If RMS has his own kernel and it doesn't have
> X11 or any other non-GNU components, then he can call it GNU all he wants
> and I'll support that.  I won't buy/download/install/use it, but I'll
> support his right to call it GNU.

This has been an interesting discussion.  Here's my two bits.  It seems to
me that there are three balls in motion here, and that it is important to
keep the eye on each.  One is the *label* that should be applied to a
composite product.  Another is the *license* under which a product is
released.  And finally there is the collective environment that emerges
from conflicts over these two. 

Frankly, I don't see the point of all the fuss over labelling.  "Linux" is
associated with the person of Linus Torvalds.  "GNU"  is associated with
the person of Richard Stallman.  In the world of proprietary products,
each would be keen to prevent the other from using their label without
paying for the privilege and meeting certain standards.  The idea is that
the label itself has value, and that value must be both honored and
preserved. 

This is not such a dispute.  Like much else in the open source world, the
logic of proprietary negotiations gets turned on its head; Richard stands
accused by some of trying to freeride on the back of Linux' prominence in
the national press, by *demanding* that his own label be applied to
products that contain material that was *not* written by him and his
organization.  In the absence of any proprietary assertion in respect of
the "external" material, he and his organization are welcome to call their
product anything that they like (so long as it does not conflict with an
existing trademark --- calling his OS "GNU Microsoft Windows" would get
him into hot water :) 

Richard's insistence that GNU be recognized in the labelling of products
that incorporate his organization's material represents a further
flip-flop of proprietary logic.  A proprietary interest would drive the
holder to demand the *withdrawal* of the label, in the absence of a
specific licensing arrangement.  And this is where the underpinning of the
discussion becomes unclear for me.  If, as some seem to feel, the demand
is driven by a disfunctional lust for self-promotion, you might as well
just ignore it. 

But the second ball comes into play here.  The GNU label does not, in
Richard's view or in that of anyone else, just stand for a big archive of
code, or for a bundle of licensing agreements in the FSF filing cabinets. 
It also stands for the GPL concept itself.  Promoting the working model
behind the GPL is what the GNU label is all about.  Viewed in this way,
Richard's strident and aggressive insistence on the wide use of the GNU
label is --- even if driven by a lust for self-promotion --- in the public
interest, to the extent that the GPL is in the public interest.  We should
probably be listening to what he says, even if we don't like the way he
says it. 

I wonder, though, whether the GNU label is the best way to accomplish the
purpose of calling public attention to issues that affect the future of
the community.  To me, it looks as though the critical element in open
source software is the third ball --- the way in which licensing schemes
promote or inhibit what Steve Turnbull has referred to as the recycling of
code.  The GPL has been a demonstrably successful vehicle for this
purpose.  Why not put this at the center, by adopting conventions for
labelling that refer to the license that covers the core of the OS (kernel
and utilities)?  There is always going to be enough unity here to keep
things simple;  open source projects with a diversity of licenses at their
core are not likely to survive for long anyway. 

My two bits, then, is "Linux (GPL mixed)" or "GNU Hurd (GPL FSF)", or
"Harvey's Homebrew Code Mix (GPL mixed)", the license name followed by the
principal holder, as, say, "mixed", "FSF", or whatever. 

I don't know what the people at the FSF would make of this suggestion, but
it seems to me a stronger model for promoting cooperation and assuring the
survival of their legacy.  In the open source world, the principal value
of labels, it seems to me, is in the sense of community that they engender
among those who work in the projects to which they apply.  Asserted beyond
those boundaries, they undermine cooperation to no particular purpose.

For what it's worth.  If anyone wants to pass this on to people in the
FSF, be my guest; I would be interested to know what their reaction is. 
It's their call entirely, but I'm ... curious. 

Cheers,
-- 
-x80
Frank G Bennett, Jr         @@
Faculty of Law, Nagoya Univ () email: bennett@example.com
Tel: +81[(0)52]789-2239     () WWW:   http://rumple.soas.ac.uk/~bennett/

-------------------------------------------------------------------
Next Technical Meeting: January 14 (Fri) 19:00
* Topic: "glibc - current status and future developments"
* Guest Speaker: Ulrich Drepper (Cygnus Solutions)
* Place: Oracle Japan HQ 12F Seminar Room (New Otani Garden Court)
-------------------------------------------------------------------
more info: http://www.tlug.gr.jp        Sponsor: Global Online Japan