Re: Economics and the GPL ... etc

On Thu, 20 Jun 2002, Roger Fujii wrote:

> "Deven T. Corzine" <[email protected]> wrote:
> > Antitrust laws exist to make the marketplace fair to the CUSTOMER (the 
> > public), not the vendor...
> 
> well, not quite that simple.  The govt will act on behalf of the public, 
> but vendors can collect damages from it (witness the suits against M$ 
> after the trial)

True enough, but that's a matter of trying to redress that inequities of 
what has happened to the "damaged" vendor as a result of the monopolist 
breaking the law.  Civil cases are often trying to redress damages; the 
antitrust aspect is just a supporting fact for proving that the damages 
suffered never should have occurred, but for the illegal behavior of the 
monopolist, but it still boils down to "he hurt me and should pay for it".

> Actually, I don't necessarily equate your examples here of being unfair 
> as the restrictions on everyone.  Antitrust should really be handled 
> separately, as it only applies to relatively large companies (and not 
> people) and addresses corporate misbehavior.  

It might be possible for a person to be sued for an antitrust violation, 
although is seems nearly inevitable that it will be corporate "persons" who 
are going to the ones who achieve (and abuse) a monopoly position...

> > Competition often spurs greater efforts.  (Consider KDE and GNOME.)  
> > Maybe the GPL helps eliminate competition, but is that necessarily a 
> > good thing?  Once the competition is gone, so is the competitive 
> > impetus to improve the code, which can lead to stagnation.  Does the 
> > GPL shoot itself in the foot by eliminating competition?
> 
> Competition is almost always a good thing.  One only has to look to find 
> examples of great progress caused by competition: space race.  kde/gnome.  
> the genome project.  netscape/IE (if it weren't for the anticompetitive 
> M$ behavior).  Public (as in non-governmental) items that generally 
> discourages competition is fundamentally flawed IMHO.

I agree, competition is generally a good thing.  Hence my question; does 
the GPL's anti-competitive nature constitute a fundamental flaw in the GPL?

> > The NCSA httpd server was originally funded by public money.  Later, 
> > Apache developed as an underground activity of webmasters looking to 
> > improve on the NCSA httpd server -- by the time management generally 
> > became aware of it, the Apache project was well-established and already 
> > becoming dominant.
> 
> well, it was a little more complicated than that, as NCSA was changing 
> their license for the webserver at the time, if I remember correctly.

The point was that Apache didn't need the GPL's protections to succeed, and 
it was developed initially by puclic money, and later largely incidentally.

> The effect was probably the same as the BSD licensing problem that 
> spawned linux.

What "BSD licensing problem that spawned Linux" are you talking about?

Linus Torvalds started Linux as a replacement for Minix, a proprietary 
product he was unwilling and/or unable to pay for.  What does that have to 
do with any BSD licensing problem?

> > Do you really think that most companies would have been willing to 
> > allow NCSA httpd improvements to be released if they thought of that 
> > work as strategic?  
> 
> uh, bad example as apache is not a copyleft license. 

Indeed.  And as it turns out, it didn't really need to be.

> > More likely we would have seen that work incorporated into proprietary 
> > NCSA httpd derivatives.  However, once Apache was established, it was 
> > too late -- the public project was moving too fast to compete with.
> 
> regardless, the licensing here (apache) didn't seem to slow it down 
> noticably.

That's the point -- the protection of copyleft isn't necessary required to 
ensure that most of the code remains available to the users...

> > But I never mentioned Gangplank in this conversation; what made you go
> > searching for it?  (After all, it's certainly not a well-known project!)
> 
> based from an RPI package, eh?   ooooo that brings back chilly memories ;)

It imitates the user interface CONNECT used back at RPI, more or less.  But 
it's not based on the CONNECT source code at all; that code was written in 
Pascal on MTS with its own funky IPC/networking system and would be hellish 
to port to C under Unix using Berkeley sockets.  Trust me, people looked at 
porting it, and ran screaming away from the idea every time. :-)

Did you used to be at RPI?  Have I talked to you on CONNECT, Clover or lily 
and not even made the connection? :-)

Deven