Subject: Re: anti/Law (an attempted explaination)...
From: Russell Nelson <>
Date: 14 Apr 1999 13:34:06 -0000

Brought in from a big CC list, use of which I detest, and besides,
it's really an FSB issue.  Hopefully you can get enough of the context 
to make sense of it.

Stephen J. Turnbull writes:
 > >>>>> "rn" == Russell Nelson <> writes:
 > Stephen J. Turnbull writes:
 >     >> Presumably partly because they don't waste resources on
 >     >> computing metrics that are subsumed by prices.
 >     rn> Prices don't just compute these metrics, they distribute them
 >     rn> as well.
 > Be careful in using words like "these"; prices are a sufficient
 > statistic _for their purpose_, but they are lossy.  They don't contain
 > all the information that's available (in particular, you won't be able 
 > to deduce any of Rich's metrics from them).  Just _exactly_ enough.

I'm just pointing out that prices do more than represent value, they
also communicate value.  I've been thinking that a good public service
for a free software business would be to set up a book on demand
service, and print up the documentation that often comes with free
software.  Set a really high price on the books, so that it's quite
profitable.  That way, someone else will show up and compete with you, 
and figure out what is the minimum supportable price.

 >     rn> Prices paid for software are a good indication of what needs
 >     rn> to be written.  Free software has the problem that nobody's
 >     rn> likely to invest in its development because they can't get the
 >     rn> price paid for it.  That's why everybody's looking for an
 >     rn> appropriate non-open-source license which preserves the
 >     rn> benefits of Open Source with the benefits of the price system.
 > I'll waive my consulting fees.  Let's talk about this.  <grin>  But
 > the source needs to be published, no?  Open Source here is referring
 > to Open Source[tm]?

Right, that's what I mean -- Open Source is always free software, but
there's a lot of free software which isn't free software.  For
example, Larry McVoy's BitKeeper is definitely libre software by many
people's definition (arguably not RMS's), but it denies certain
freedoms that are required for Open Source certification.

 > There are techniques for analysis.  Technically, called "mechanism
 > design," "principal-agent theory," "auction design," "implementation
 > theory," and probably a bunch others I can't recall off-hand.  Very
 > much work-in-progress (since the early 1970s), but usable for many
 > purposes.
 > We _may_ be able to test proposals, too, in laboratory experiments.
 > I'm not real sure about that, though, because in my current
 > (half-baked) model of the free[1] software phenomenon the important
 > tension is between developer preference for free software and profit
 > from proprietary/secret software.

Are you talking about a developer preference to create free software,
or to take advantage of free software?

 > This is hard to capture with "induced preferences" as used in
 > laboratory experiments.  If we're just interested in a
 > profit-maximizing model, then those mechanisms _can_ be tested
 > (although we would have to abstract from the "I just like free
 > software" motive).

 > Footnotes:
 > [1]  In the GNU Manifesto sense, as the extreme case.  I would be
 > interested to hear from anybody who thinks free software in itself
 > might be undesirable for any reason, apart from the economic issues
 > (development disincentives and right to economic compensation) we are
 > familiar with.  The reason is that I may want to use "free software"
 > as a goal or preference in itself in my models, and would like to know
 > of any bad effects free software might have that would _not_ be
 > captured by loss of incentive to create on the part of other
 > developers.  (These would have to be added to the model as ad hoc
 > external effects, for the moment.)  I can't think of any, myself.  :-)

If you view support as insurance, then a proprietary support model
(which comes with proprietary software) creates a larger base of
rate-payers, and turns support into a cost center to be minimized.
This addresses Keith Bostic's classic concern that paying for free
software by selling support creates an incentive for buggy software
and inadequate documentation.  Presumably a free market incents
manufacturers to reduce the need for support rather than reducing the
available support.

-russ nelson <>
Crynwr supports Open Source(tm) Software| PGPok |   There is good evidence
521 Pleasant Valley Rd. | +1 315 268 1925 voice |   that freedom is the
Potsdam, NY 13676-3213  | +1 315 268 9201 FAX   |   cause of world peace.