Re: Economics and the GPL

On Wed, 12 Jun 2002, Matt Seitz wrote:

> I found myself agreeing with the vast majority of your last article.  You make
> some very cogent observations and suggestions.  I hope you're not too burned
> out yet to read yet another response to some of your points.

I'm getting close, but I'll try.

> Couldn't the FSF just add terms to the GPL that say something like "software
> offered under these terms may not be offered under any other terms"?

I suppose they could.  Somehow I don't think that would sit well, even with 
many of their supporters.  Also, I'm not sure it would be enforceable.  
Perhaps that would be a null statement, in a legal sense -- as the KDE 
developers were free to violate the GPL on their own code, wouldn't the 
copyright owner be perfectly free to ignore this self-constraint?

> I agree that the RMS would probably prefer that one use the GPL as an 
> exclusive license.  But from what I see on the FSF site, I don't see that 
> they view this as abusing or circumventing the GPL.

I didn't say that it constitutes abuse of the GPL, merely circumvention.  
While the FSF may not try to stop you from offering alternative licenses 
(since they know they can't), they should don't encourage it!

Of course, if they convinced you to transfer copyright to the FSF, then 
they're in complete control over your code -- it's no longer truly yours.

> "Also, for the record, I am not a Communist or anything similar." 
> --Richard Stallman (June 28, 1999), "RMS Responds", SLASHDOT
> (http://slashdot.org/articles/99/06/28/1311232.shtml)

Okay, then he has denied it.  Very well then, he may not accept that label.  
Nevertheless, his philosophies have a definite affinity with Communism, so 
whether or not he's "really" a Communist isn't important.  He quacks like a 
duck!  ;-)

> Now we're dealing not with the GPL, but with the FSF's core philosophy. 

It's more about Stallman's personal moral code, though the FSF may reflect 
that, of course.

> Personally, I think RMS makes some very good arguments about why some 
> limited copying by private individuals should be allowed even against the 
> author's wishes.  

That's more a question of public policy, I think.  As such, it would be a 
good topic for public debate.  I don't think Stallman should be seeking to 
decide this issue for the rest of us, whether or not he has a point.

> >THIS is the thing that makes funding so difficult.
> 
> Even RMS accepts that placing restrictions on unauthorized mass, commercial
> duplication may be necessary to make funding of works possible.

And does he suggest that can be done in a way compatible with 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. 
> 
> That is a valid concern.  But there are similar risks in other common
> transactions that have long-term contracts.  I sign up for a cel phone 
> plan, but then it turns out I don't use it as much as I thought I would.  
> I take out a loan to buy a car or home, and later my income decreases.  
> In each of these cases, I'm still on the hook for what I said I would pay.

At the same time, most of those potential regrets tend to be time-limited.  
The contract expires, or the loan is finally paid off.  If the GPL had the 
code "expire" and fall into the public domain or BSD-style free licensing 
after a fixed time period (e.g. 5-10 years), I wouldn't worry as much...

> > 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.
> 
> That is an excellent suggestion.

Of course, Wine has sort of done the reverse -- it was well-established 
under the BSD/X11 licenses, and switches to LGPL.  Alexandre belives it's 
in the best interests of the project, and maybe it is.  A BSD advocate 
might argue differently, counterintuitive though it seems.

The BSD camp makes an interesting argument, but I can't fathom how to 
construct a valid test to figure out if they're right.  There's too many 
other factors you wouldn't be able to control, so it's all anecdotal...

> > But the GPL doesn't even try at all.  
> 
> It seems to me that it is BSD/X11 that doesn't try at all.  The GPL 
> formula may be extremely simple, but it at least tries to establish some 
> balance.

No, BSD doesn't seem to try.  Yet somehow they sometimes appear to succeed.  
Maybe it's some sort of Zen thing.  Do you maintain a better grip by not 
squeezing as hard as possible?

> > I think I'd consider BSD-type software (plus public domain of course) 
> > to be the true "public" software.  
> 
> I'd say both are "public" in the sense that anyone can use them.  They do 
> have different costs and rules, but then again many "public" resources 
> have costs and restrictions on how they are used.

If anyone could use them, Microsoft would be.  It's a private club, and 
only members can benefit.  Now, it's true that membership is open to the 
public, but under conditions that are unacceptable to some.

> > An appropriate designation for GPL software would be "communal" 
> > software.  Of course, that would be an even harder name to sell than 
> > "free" software!
> 
> Agreed on both counts.  Perhaps "community" software?  And you have to be a
> member of the community to get all the benefits (just like you have to be a
> member of my local homeowner's association to use the association's pool)?

That's a better name.  It doesn't carry the '60s hippie connotations that 
"communal" does.  Of course, it might be argued that Stallman is a '60s 
hippie, even now!  He's got the hair for it...  ;-)

Maybe I'll use the term "community software" myself as a categorical term.
I think it applies to the GPL, but I know Stallman won't change his choice 
of terminology.  Still, it's a good, neutral description for software that 
is limited to a particular community, and it doesn't connote "free beer".

Like I said before, I've been considering trying to create a community that 
doesn't necessarily support "free beer" code, but otherwise could have the 
sorts of benefits the GPL community derives from the GPL.  Some (but maybe 
not all) code might be "free beer" within that community, but would never 
be allowed to be given for free to outsiders -- if they want it, they would 
have to join the community too.  But, like the GPL community, it should be 
open to all.  However, every member would have to pay some minimum "dues" 
as an ongoing requirement of membership -- for the purpose of funding new 
development or other benefits to the community...

So, do you think it could work?  People are willing to have all GPL code 
restricted to the GPL community, why not another community with a model for 
sustaining new development without depending strictly on volunteer effort?
It wouldn't fit the FSF's "free software" definition, the Debian FSG or the 
Open Source Definition, because it wouldn't offer free redistribution to 
the public at large.  (Some projects might allow redistribution within the 
community, if they choose.)  Some projects might charge extra above basic 
membership because they're high-value and high-cost, but the basic idea is 
to have all code in the community benefit the community, even if some of it 
might be "commercial" in nature...

How could I avoid backlash from free software/OSS ideologues who would just 
assume it must be bad just because it doesn't conform to the FSG/OSD?

Deven