Re: Economics and the GPL

On Tue, 11 Jun 2002, Matt Seitz wrote:

> I don't see why the GPL is unfair.

I didn't say that it _is_ unfair, but that many people consider it unfair.  
Meanwhile, many others consider it eminently fair.  It's all a matter of 
perspective.  I don't think you could ever get unanimous agreement here.

> If all you want is to use a portion that was written by 1 person,  all 
> you need is permission from that 1 person.  The GPL does not compel 
> exclusive licensing:  if I release something under GPL, I am free to 
> release it under other licenses, too.

The GPL does not compel exclusive licensing because it cannot.  It remains 
the perogative of the copyright holder to offer alternative licenses at any 
time.  However, this is circumventing the GPL, not working with it.  Yes, 
the GPL can be circumvented.  The reason it would need to be, is that it's 
not truly a free license.

When was the last time you saw someone to offer an alternative license on 
(2-clause) BSD code?  You're not likely to ever see that, because anyone 
can just use it under the BSD license -- it's a much more  free license.

> Again, I think we agree:  if the GPL codebase were a monopoly, then it 
> could be as harmful as Microsoft's dumping of IE.  Until then, I don't 
> see it as "unfair".

It's all a matter of degree.  The more you constrain your focus, the more 
likely it is that in a niche market, the GPL may already be unfair, even if 
on the broadest scale it isn't (yet).  But it could get there unnoticed...

> True, but is there a better alternative?

Exchanges of similar value (however that's determined) is an alternative.  
Whether or not it's better is a matter of perspective, again.

> I would say "leaves little" instead of "leaves nothing", but I essentially
> agree.  That is a problem with the GPL, but I don't think that automatically
> disqualifies it from consideration as a licensing option.

No, but it sure makes it harder to fund GPL development from scratch!

> > But dig deeper than this prima facie inequity and you come face-to-face 
> > with the conundrum -- without permitting this apparent inequity, the 
> > BSD community would have received no benefit at all, instead of the 
> > partial benefit they did receive.
> 
> I think this is an interesting point.  This reminds me of the apparent
> contradiction of Adam Smith's invisible hand:  that by allowing everyone to
> pursue their own interest, society gets more benefits than by trying to force
> everyone to serve society.  The same may be true of GPL vs. BSD.  GPL, by
> trying to force people to give away code, may result in less code being
> available.  BSD, by not forcing people to give away code, may result in more
> code being available.

I'm not sure, but I have to concede that this could be the case.  On the 
surface, copyleft seems like a better approach to ensure the availability 
of more code, yet in practice BSD has been surprisingly effective.  This is 
counterintuitive, I think -- but that doesn't make it wrong.  I'm still on 
the fence on this one.  Is the BSD license is ultimately more effective?

It's possible that the BSD license is only preferable for older, mature 
code and/or for vital, active projects.  I'm not sure.  It's so easy to 
"steal" that it seems like the project might need to be well-established to 
truly benefit from the BSD license.  (But that's just a guess.)

> I misspoke.  The point of the GPL is to promote allowing individuals the
> freedom to redistribute code.  Requring that money be paid to 
> redistribute code is the most common way that freedom is impeded.  So, 
> one of the primary goals of the GPL is to allow code to be redistributed 
> without requiring a cash payment.

And people want to claim that Stallman isn't a Communist?  I've never heard 
him say that he is, but he doesn't deny it, and he sure acts like one.

I don't have a problem with Stallman wanting to make changes to a program 
and send his changes to another user of the program, so that other user can 
benefit from his changes.  I do have a problem with him wanting to send the 
program itself, without regard for the author's wishes, to someone that's 
not already a user.  THIS is the thing that makes funding so difficult.

> It is not my experience that prices are always negotiable.  In fact, my
> experience is the opposite:  usually a price is a take it or leave it
> propositition.

That's true.  Of course, some people argue that the choice to walk away 
constitutes the "negotiation", but that seems a bit disingenuous to me.

> In that sense, I don't see the GPL as being any different than placing a 
> fixed price tag on any good or service.

If it were a fixed price tag, it might not be such a big deal.  The GPL, 
however, is an unlimited commitment.  You might start small, making a few 
changes, then more, and eventually find you've written a substantial amount 
of code.  If you had known from the start that you'd write so much code for 
that project, you might have decided to avoid the GPL.  The GPL can hook 
you while your commitment is small, but you're still on the hook no matter 
how large your commitment grows.  This is an insidious thing.

> And the GPL doesn't shut the door to that possibility.  You are always 
> free to negotiate with the rights owners for different terms.

The GPL certainly shuts the door on this possibility.  Which is exactly why 
you have to circumvent the GPL, going behind its back to the copyright 
holder who has more power than the GPL has.

> Short of negotiating seperate terms for every exchange, and having 
> perfect knowldege of the true value of both sides offers, there is no way 
> to guarantee that every exchange will be exactly fair and equitable, 
> especially when one is using a barter system.

Does it matter if it's objectively fair, as long as both parties are happy?

> There also comes a point of diminishing returns, where the effort 
> involved in trying to negotiate an equal exchange outweighs the benefit. 

Yes, this is undeniable.

> I agree, this is a weakness of GPL, and I think there is room for 
> improvement.  But I don't see the GPL terms as so unreasonable, so 
> outrageous, that they should be dismissed and never used.  I also don't 
> see the BSD's unconditional giving away of value without any guarantee of 
> return as a perfect solution, either.  

I think the BSD approach is more noble and truly free.  It's selfless, and 
they're more likely to be taken advantage of for the selfish benefit of 
another.  Yet, this must not be too objectionable to BSD developers, if 
they continue to use the license knowing this can happen.

I don't recall saying that the GPL should never be used, only that I see 
potential danger if it grows too powerful.  Okay, maybe that implies that 
it should not be used, to keep it from getting too powerful, but I think it 
works well in the underdog role.  Perhaps the logical solution is to use 
the GPL or another copyleft license early in development, and switch to the 
BSD license once the project is well-established.  Of course, this applies 
to volunteers not seeking a return on their investment.  The GPL remains a 
difficult problem for companies seeking ROI.

> > The GPL will inherently tend to take advantage of the more prolific coder 
> > in an exchange.  
> 
> Again, you are correct.  That is a weakness of the GPL system.  But again:
> -I don't see it as a fatal weakness

Maybe not, but it accounts for the widespread perception of unfairness.

> -I don't see a simple way to easily balance the value of each person's
> contribution.

I'm not sure either.  This is a tough question.  But the GPL doesn't even 
try at all.  Surely something better could be done, even if it's just done 
by voting for which person is most deserving, or something.

> -I don't see how BSD/X11 is any better at solving this problem.

I'm not sure that it is.  I think the BSD/X11 approach makes the most sense 
if you feel the software is already "paid" for (in whatever fashion, maybe 
just in satisfaction for a volunteer), but I don't see BSD/X11 working well 
at all if you are still seeking a return on the investment you've made...

> In a co-op, it is easier to come up with more objective criteria for how 
> much each person has contributed: x dollars, or y hours of work, or z 
> acres of land, etc.  It is much more difficult with source code; see how 
> many metrics there are for measuring the "value" of code (lines of code, 
> function points, etc.).  Unless someone can come up with some neutral, 
> objective way of measuring the "value" of code, how can we avoid this 
> problem in any open source project?

It may always be subjective.  Moreover, a large component of the value is 
what it's worth to the users.  If a large program with many lines of code 
and function points is efficient, effective, beautiful and elegant, but 
only useful to a very small niche of users, is it really more valuable than 
another program which is "objectively" not as good, but useful to millions?

A simple scheme would be to collect equal "dues" (monetary contributions) 
from everyone in the community, then allow the community to vote on how to 
split up that money among the projects they value.  Unfortunately, that 
devolves into a popularity contest, but it could work.  But if anyone can 
just give away the code, how can you collect "dues" for membership?

> True, but the GPL as it stands today does not require buying entirely 
> into the FSF philosophy.  It doesn't require one to give up all one's 
> claims to all copyrights, past, present and future.  Of course, I think 
> that's a good reason to avoid the "or any later version" clause when 
> using the GPL.  

Think of the GPL as a commune where the gate is open -- people can come and 
go at will, at least for now.  I'm sure that Stallman will try to make it 
harder and harder for people to walk out that gate in the future, though, 
as he gets more and more power.  (He really is like a cult leader...)

> > On thing's for sure.  The BSD folks have a much stronger claim to the term 
> > "free software" than the GPL ever had.  If anything, Stallman's use of the 
> > term "free software" amounts to Orwellian Newspeak...
> 
> I agree.  I tend to think of GPL software more as "public" software, and that
> it is like a public park.  Like a public park, it is available at little or no
> cost to use.  It is open to everyone, but there are rules restricting what one
> can do there.  The rules are designed to keep the park safe, enjoyable, and
> accessible for everyone, and one is probably not allowed to run a business or
> post advertising there.  Like public parks, I think public software has its
> place and can be beneficial.  

I think I'd consider BSD-type software (plus public domain of course) to be 
the true "public" software.  An appropriate designation for GPL software 
would be "communal" software.  Of course, that would be an even harder name 
to sell than "free" software!

Deven