Subject: Ghostscript, GPL, printers, ...
From: "L. Peter Deutsch" <ghost@aladdin.com>
Date: Sun, 30 Nov 97 21:25 PST

(I've added the full fsb list to the distribution, since I think this thread
is of interest to the entire list as a discussion of business models.)

> Peter Deutsch writes:
> 
> >Ghostscript is in the process of being licensed to some of the largest and
> >best-known manufacturers of hard-copy output devices in the world,
> >displacing software that is completely proprietary.  Those manufacturers
> >would never have considered using GPL'ed software, because it would have
> >required them to give away the source code to their entire printer firmware,
> >which is one of their key sources of differentiation in the market.  (Most
> >of these manufacturers use hardware from the same few Japanese companies, so
> >the firmware, and their internal cost of operating the business, is what
> >keeps them competitive.)  So which do you think is the most beneficial
> >approach in the long run:
> 
> >	1. Put Ghostscript under the GPL and wait until a manufacturer comes
> >along who is willing to GPL all their firmware.
> 
> >	2. Put Ghostscript under the GPL, but allow manufacturers to license
> >it for $$ under commercial terms.
> 
> >	3. Put Ghostscript under the LGPL.
> 
> 	First choice: option 2.  Second choice: LGPL + additional
> licensing option.
> 
> [...]
> 
> >#2 is pretty close to Aladdin's business model.
> >[...] in
> >any case, the differences between the GPL and AFPL are much smaller than the
> >issue of whether commercial entities who could not possibly afford to live
> >with the burden imposed by the GPL should be allowed to license GPL'ed
> >software for non-GPL'ed use.
> 
> 	The impact on Aladdin's immediate business opportunities may
> be small or non-existant,

Well, that's the difference of opinion that was part of the motivation for
the AFPL in the first place.

First, let me say that I believe the (labor-intensive) support business
brings in much less net revenue per unit, given customer expectations of
lower costs and a requirement for a much larger staff; but I don't believe
anyone has posted real numbers about this on this list.  I think it would
really help make this discussion more concrete if Yggdrasil, Cygnus, and
some other companies currently in the free software business could give us a
breakdown of revenues:
	- support
	- contract development
	- sales of software and documentation
	- other revenues (specify)
and of costs:
	- support staff
	- development staff
	- other staff
	- other costs (specify)
for their last complete fiscal year.

Now that a separate point brought up by Adam has led me to read the GPL more
closely, I think its "independent and separate" provision makes it possible
for a reasonably diligent implementor to build a product around GPL'ed code
without bringing any significant amount of the added code under the GPL.
(After all, it would be up to the copyright holder of the GPL'ed code to
persuade a court that the implementor's added work was not "independent and
separate".)  In other words, where large functional modules are involved, I
think the GPL is functionally almost equivalent to the LPGL.  That being the
case, I no longer believe that there is a useful distinction between
embedded and non-embedded products: I believe that any reasonably diligent
developer who would want to use a GPL'ed Ghostscript in a proprietary
product could find a way to do so without bringing the rest of the product
under the GPL.

Given the above, I think the list of choices I presented before is
irrelevant.  Aladdin would have to expect that all current commercial
licensees (most of whom now pay per-unit license fees, and some of whom pay
support fees in addition) would switch to the GPL (and would only pay
support fees, if any).  Given that I expect revenue would drop by about a
factor of 5, that would put us out of business.  (I know that the factor of
5 is approximately correct for our second largest customer; our largest
customer has a very special arrangement, and I don't know what would happen
if we tried to renegotiate the contract for support only.  Less than 5% of
Aladdin's income comes from development work done on contract.  And that's
assuming that our present commercial customers, many of whom have worked
with Ghostscript for a long time, would be willing to pay for support at
all.)  We currently have:

	- 1 full-time support engineer.

	- 1 full-time development engineer working on a proprietary
	project.

	- 1 full-time jack-of-all-non-technical trades (the president).

	- me, spending about half my time on development (nearly all of it
	product development, a little of it contracts) and about half my
	time dealing with problems and questions coming in from the
	Internet.

We would not be able to support the development engineer out of current
revenues, so we would have to lay him off.  The person I hired to run the
business would quit, since he wants to build up a highly profitable business
in a reasonable number of years, not work indefinitely at a relatively low
salary; we would have to hire someone less ambitious and capable.  I would
be unable to achieve any of my personal financial goals for the future.

> but the public benefit would be substantial.
> I am not referring to the public benefit of having Ghostscript based
> printers, which would be unaffected, but rather the benefit of having
> the contents of latest Ghostscript available for immediate use in
> other GPL'ed software projects.

The only significant one I know of is GNUstep, and I have already promised
to release the Display PostScript subset of Ghostscript under the GPL as
soon as it is ready.  What others do you know of?

> For Aladdin, such would improve bug
> detection, customer's confidence of in your product as a result of
> wider use by more people, and perhaps encourage your paying customers
> to upgrade to the latest version due to interest by more end users.

Actually, the overwhelming majority of Ghostscript users are end users of
the GPL'ed or AFPL'ed code available on the Internet, and they are quite
happy with the AFPL: switching to the GPL wouldn't increase this pool,
except insofar as FSF and libre software distributors have attempted to
prevent potential users from learning about the existence of the AFPL'ed
versions and have therefore led them to believe Ghostscript is less useful
to them than it actually is.  (This is not hyperbole, by the way: FSF will
not allow any mention of the AFPL'ed versions on the Ghostscript Web page
that they link to, and they would like me to remove any mention of
commercial licensing from the documentation distributed with the GPL'ed
versions.)  Therefore, if I've understood you correctly, it seems to me that
your argument turns on the belief that there is a large untapped pool of
potential Ghostscript-based GPL'ed software projects, which in turn would
draw in a large community of new users of Ghostscript.  I don't share this
belief.  Unfortunately, I don't know how to go about determining its
validity, other than to ask what such projects anyone knows of.

> 	I just hope that if a printer company does decide to go the
> GPL route, you'll have the good sense and flexibility in your business
> model to be able to graciously offer to sell them a service contract
> rather than denouncing them for "taking a free ride" or something like
> that.

Actually, we usually graciously decline to do business with them, because
the cost of support-only contracts is so high and the potential revenue so
low that we cannot afford to provide them at competitive rates with
high-quality people.  See above.

> 	By the way, regarding Adobe Postscript licensing costs, 3% of
> printer sales is not cheap; it, in all likelihood, translates into a
> large fraction of profits.  Perhaps your potential customers are
> minimizing this point with you to reduce their appearance of interest,
> just as a negotiating tactic.

Actually, neither of us has hard information on this subject, and as it
happens, I'm pretty sure that even the 3% figure I quoted before is out of
date (it's too high).  I have blind-copied this posting to a few friends in
the printer business, who I hope will respond with some real numbers.  (Note
to those people: please respond to me, not to the list, so I can sanitize
your responses a little before posting them.)

-- 

L. Peter Deutsch         |       Aladdin Enterprises :::: ghost@aladdin.com
203 Santa Margarita Ave. | tel. +1-650-322-0103 (AM only); fax +1-650-322-1734
Menlo Park, CA 94025     |        http://www.cs.wisc.edu/~ghost/index.html
*Oppose bulk-mail abuse of your mailbox and newsgroups: http://spam.abuse.net*