Subject: Re: small worlds and better than ransom
From: "Stephen J. Turnbull" <stephen@xemacs.org>
Date: Tue, 09 Oct 2007 04:08:07 +0900

La Monte Henry Piggy Yarroll writes:

 > Stephen J. Turnbull wrote:
 > > La Monte Henry Piggy Yarroll writes:
 > >
 > >  > I've always seen the lagged release of open versions as close to 
 > >  > pessimal.

 > Please pardon my shorthand. I should have specified the quantities being 
 > pessimized.

I understood those.

 > Zealots such as myself are usually interested in optimizing something 
 > like "social good".

Economists like me, too.  No conflict of goals.

 > The Andrew model pessimizes all of these.

But you didn't say "Andrew model", which may have been horrible (I
don't know the history).  You said "lagged release of open versions".

There are lots of ways to do that, assuming by "open" you mean FLOSS,
and not merely "you can look but not touch the source" or something
like that.  You partially open the code, allowing redistribution under
conditions that do not compete with your main business.

For example, Peter Deutsch's original proposal to RMS was that
software distributed in embedded form (Postscript in ROM, for example)
should not be considered as licensed under the GPL.  The argument is
that even if you get the source, it's not useful for bug fixing.  RMS
shut that down, of course, but note that a Postscript text book would
be allowed to distribute Ghostscript on disk: a much weaker restriction
than the actual "no commercial redistribution in any form" restriction
in the Aladdin license.

 > So is the Aladdin model dual license with lagged release ("last year's 
 > version") or dual license with simultaneous release ("source ... was 
 > released to the public ... pretty much immediately.")?

It was that latter.  They had a reasonably aggressive legal department
which went after commercial publishers (who put Ghostscript on a disk
in books on Postscript programming) and printer and fax makers (who
put it in ROM) without licenses.  On the contrary, Debian stuffed them
into "non-free".

 > I've been using the model of "Find an employer who thinks what I do for 
 > the community is valuable enough to them that they want to feed me and 
 > my family." This worked well for three years at Motorola, but hasn't 
 > been so hot at two VC-funded startups. I'm rather tired of the "Oh yes, 
 > we recognize that we need to be more involved in the community." 
 > followed by vigorous resistance to efforts to promote such involvement.

Yup.  "Internal patronage".  Still requires that the patron be
well-heeled and have a long enough history of being so to feel
"noblesse oblige".  VC-funded shops sometimes qualify on the first,
but almost never on the latter.

 > The advantage of my model over Thomas Lord's is that I need only
 > find one "customer". The disadvantage of my model is that the first
 > customer changed their priorities and subsequent ones have not been
 > overly truthful about what they were hiring me to do. I still
 > haven't found the one customer I'm looking for.

Would you be interested and willing to discuss the parameters?  For
example, would an Aladdin-style phased release (current version source
is redistributable but only for non-commercial purpose, lagged version
is fully FLOSS) be acceptable?  Maybe even attractive?  Then something
like Ghostscript where there are drivers and optimizations that
individual widget makers would pay a lot for you to create for a
proprietary version they only use, with eventual release as FLOSS
might be a good target.  Alternatively, how about a web framework
company as their Apache specialist (or, since you mention the kernel,
khttpd specialist), with the secure server being proprietary but the
general efficiency stuff that you work on being FLOSS?  (NB I'd be
surprised if the examples I'm giving make total technical sense.)