Re: Economics and the GPL

Matt Seitz <[email protected]> wrote:
> "Roger Fujii" <[email protected]> wrote in message
> news:[email protected]...
> > Why does market share alter why it's unfair?  You can say the effects
> > are minimal UNTIL it has a sizeable market share, but I fail to see why that
> > makes it right.
> Now we are getting into the whole ethics of antitrust laws. 

you're confusing different issues.  GPL has been around way before M$ was a
monopoly, so using it as an anti-monopoly response isn't right.  Besides,
anti-trust laws are for anticompetitive behavior - it is not an anti-monopoly
remedy.

> And I think one can make a perfectly valid, reasonable argument that what is 
> "fair" should not depend on one's market share.  Personally, I see it as the
> difference between a boxing match between equals and one between Mike Tyson 
> and a child.  When one side has an overwhelming advantage, I think you do need
> to change the rules if you want to have a fight in which either side has an
> opportunity to win.

Why favor ANYONE by changing the rules (assuming the rules are fair)?  Just because
cisco has > 90% of the market, doesn't mean it's right to give its competitors a
subsidy.
While I can think of many other valid claims that can be made (for diversity,
national
security...), FAIRNESS is not one of them.

> Is handicapping a golfer's or bowler's score unfair?  I think it depends on what
> your goals are.

the golf/bowler analogy isn't appropriate.  You make handicaps because there would
be no point of the competition if you didn't.  A good general rule for "fairness"
is that you treat people equally.  How can you be fair by discriminating?  

> "Roger Fujii" <[email protected]> wrote in message news:[email protected]...
> > the *only* way you can have "unauthorized copying" of the GPL is to change it.
> Yes, the way the GPL is written and licensed today.  What I had suggested was
> that the FSF could, if it chose, change the licensing of the GPL itself to try
> to require that works distributed with the GPL be exclusively distributed with
> the GPL.

That would make the GPL incompatible with all other licenses (including X11).  In
addition, all GPLed projects would then have to weed out all the patches given to
them that were not exclusively GPLed (no X11 type patches).  That would kill
the GPL real quick.

> > > > However, [reducing copyright length] is an uphill battle against
> > > > powerful entrenched interests, so it's hard to win this one.
> > If that were their intent, they would have time-limited the GPL.
> RMS and the FSF have repeatedly stated that they support reducing the length of
> copyright*.  I imagine the reason they do not time-limit the GPL is that they
> do not want to unilaterally weaken the GPL and give an advantage to proprietary
> code.
> 
> *For example, see http://www.fsf.org/philosophy/eldred-amicus.html,
> http://www.fsf.org/philosophy/copyright-and-globalization.html

It's unclear that they *truely* would support such a move, as a relatively
short copyright would undermine the GPL (since they would be affect by this
too), releasing more code into public domain.  Given that they argue against
X11, it's not like they are consistant in their argument.  You can't be FOR
putting more stuff into PD and argue against putting stuff into PD.  


> --- Roland <[email protected]> wrote:
> > The example FreeBSD x Linux show that GPL products can have more success
> > than BSDL.
> 
> You are right, that is good evidence in favor of GPL over BSD (and rather
> obvious, I'm embarassed I didn't think of that comparison).  Of course, RMS
> argues that the Linux kernel isn't 100% pure GPL because of loadable kernel
> modules and binary firmware compatability modules.  So one could argue that
> Linux is proof that a GPL/BSD hybrid is more effective than either pure GPL
> (Hurd) or BSD (FreeBSD).  However, I still think this is strong evidence of GPL
> being a more effective license.

Say this in a few years after adding up the macosX numbers.  RedHat lost $140M
at year end Feb, 2002.  Their revenues were only $79M.  It's a little too early
to declare a winner....

-r