Subject: Re: Anyone know of a free software business list?
From: "Stephen J. Turnbull" <>
Date: Mon, 12 Jun 2006 13:43:42 +0900

Laurent GUERBY wrote:

    Guerby> Unless I miss something per device fee are a feature of
    Guerby> proprietary software offering, and developpers of those
    Guerby> devices turn to free software to avoid them (see the
    Guerby> poll).

>>>>> "Thomas" == Thomas Lord <> writes:

    Thomas> Unless *I* miss something, Gumby isn't talking about "per
    Thomas> device license fee" he's talking about "value based
    Thomas> pricing".

As you recognize yourself later, he *is* talking about a pro rata
license fee (although it might not be defined as "per device").

Namely, the one you can't charge in a market segment where passable
free-as-in-free-speech software exists, whether you are a proprietary
supplier of software or an open source supplier of software.

    Thomas> So, OEM X is going to sell a million units.  Each unit
    Thomas> sold is an increment to the profit of OEM X.  Gumby wants
    Thomas> to help write and support the OS and toolchain but he
    Thomas> hopes to get paid proportionately to the profit.  That
    Thomas> doesn't *have* to imply a per-unit license on the software
    Thomas> Gumby delivers.

In a world where you don't know how many units downstream is going to
sell, any value-based pricing scheme *must* involve pro-rata licensing.

    Thomas> For example, Gumby could implement a "brain drain" so that
    Thomas> only his firm is qualified to provide the free software OS
    Thomas> and toolchain (early Cygnus).  At a certain point in
    Thomas> history, that not only made sense but (afaict) worked.

Richard Gabriel says not; that's the central thesis of "Innovation
Happens Elsewhere".  Uli Drepper says not.  "Of course" he admits to
being far and away the best developer in the group working on
glibc<wink>, but he doesn't claim to be essential.  He also qualified
one of his co-worker's claims (at a talk at IBM's Yamato Lab in Tokyo)
that "we have the "best glibc at Red Hat" with "the best for the
purposes that people will pay money for".<wink again>

I'm willing to bet that Gumby will deny that was the model, too.  It
did happen, perhaps, but as a transient side effect of a different

    Thomas> Well, the pattern was good -- you just can't coast in one
    Thomas> area of expertise.  It's a single wave, not a way of life.

Then it's not profit; it's (rather uncertain) wages, or at best rent
to a certain talent.  Profit is not about the Red Queen's race.
Profit is about investing, so that in the future all labor you employ
is "high-powered".

This is definitional (for an economist, at least), but if you think
about it in terms of "empowering labor" (in a physio-economic sense,
even if socio-economically labor doesn't capture that power as wages),
you see why it is in a certain sense the "right" definition.  Profit
is the gain from empowering labor.[1]

    Thomas> The answer lies not just in a return to abstract "software
    Thomas> freedom" roots but to a new exploration of how to
    Thomas> *maximize the utility of possessing those freedoms*.

No, "maximizing utility" is the mistake that Laurent makes.
Monopolies do not maximize utility; they do make profit.  Why?  They
make profit because they collect revenue; they collect revenue because
they _decrease_ utility as a _necessary_ side effect of increasing
price.  Profits are partly about *maximizing surplus* (utility minus
cost), and partly about *extracting a share* of the surplus.  Monopoly
balances these two goals to achieve maximum profit.

Free software may be characterized as focusing on the former to the
exclusion of the latter (and thus to a first approximation,
guaranteeing zero profit).  To the extent that a FSB is profitable, it
has to be based on rents from something other than the software per
se.  Human capital in the form of programming talent or training and
experience with a particular body of software, etc.  Talent is a
natural monopoly (or oligopoly); there are no perverse incentives
there.  But training, experience, etc, do have incentives to hoard, or
at least disincentives to diffuse, built into them.

[1]  This is not to deny that in the real world employers often have
sufficient power to depress wages and gain additional profit from
that.  But in my experience the employers whose survival depended on
depressing my wages below what I thought reasonable didn't.  To
survive and prosper, one must make one's employees significantly more
productive than even a generous wage would indicate.

Graduate School of Systems and Information Engineering   University of Tsukuba        Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
        Economics of Information Communication and Computation Systems
          Experimental Economics, Microeconomic Theory, Game Theory