Subject: Returns to service professionals (was Re: New ESR paper: The Magic Cauldron)
From: (Frank Hecker)
Date: Sun, 27 Jun 1999 16:58:35 +0000

Ian Lance Taylor wrote:
> ESR has a new paper out: The Magic Cauldron:
> It's about the economics of free software, so it is directly relevant
> to FSB.

(And I'll try to stay relevant as well.)

One point early in section 14 ("Getting There from Here") caught my eye:
"This conceptual revolution will have some cost in foregone profits for
people investing in the sale-value 5% of the industry; historically,
service businesses are not as lucrative as manufacturing businesses
(though as any doctor or lawyer could tell you, the return to the actual
practitioners is often higher)."  I found it odd that ESR used doctors
and lawyers as examples of service workers making higher returns than
one would expect in a service business.  A major reason doctors and
lawyers make those higher returns is because those professions have
artificially high barriers to entry, including state-mandated licensing.

Whether ESR meant it this way or not, I read his parenthetical comment
about doctors and lawyers as implicitly reassuring programmers about
their prospects should software development switch to a 100% service
business: "See, you'd still have a chance to make serious money, just
like doctors and lawyers."  However unless he's advocating mandatory
programmer licensing and other artificial barriers to entry (which I'm
sure he isn't) then I think the implied analogy between programmers and
doctors and lawyers is false and misleading.

I don't write this just to be picking a nit.  One of the major areas of
FUD around FSBs is whether a shift to libre software will cause
developers to be worse off, both in the sense that developers would have
lower returns on average relative to today (because they'd supposedly be
competing in a service industry with minimal barriers to entry) and also
in the sense that the maximum returns possibly achievable by any
particular developer would be lower (because they'd supposedly be
forgoing the possibility of Microsoft-like gross margins from software
licensing fees).  For example, I see many of the comments made by Ed
Muth and others at Microsoft as being in this vein, as in essence
saying, "free software is a bad (economic) deal for programmers, don't
do it."

As a side note, I suspect that the occasional spectacular returns
realized by programmers in the proprietary software business
(programmers at Microsoft being the classic example) are in large part
why programmer licensing, unionization, and similar schemes have never
been embraced within the profession.  Even though programmers or
would-be programmers would not necessarily realize those high returns on
average, nevertheless the possibility of them has made programmers
resistant to arrangements that would decrease maximum possible returns
by imposing collectively-binding compensation arrangements (e.g., union
contracts) and/or greatly increasing the up-front investment necessary
to enter the profession (e.g., licensing requiring expensive education a
la doctors and lawyers).  I think similar thinking is behind much "folk
opposition" to libre software licenses and the continuing attempts to
create "quasi-libre" licenses with some proprietary component: people
are focusing on preserving the potential for high returns and to some
extent ignoring the likelihood of their actually achieving such
returns.  Thus although the "sale-value" software business may represent
only 5% or so of total software development activities (as ESR claims),
IMO it nevertheless has had a major influence on programmer attitudes
and behavior across the other 95%.

Getting back to Ed Muth and friends, what's a reasonable response to the
criticism that libre software will decrease both programmers' average
returns from programming and their maximum possible returns?  I'll leave
aside the question of average returns in favor of the question of
maximum possible returns, partly because I think it's more important
(for reasons noted above).  One possible response (made by Nicholas
Petreley, among others) is that in the existing proprietary software
industry the likelihood of achieving high returns is decreasing for most
everyone except Microsoft (and Microsoft employees), so that moving to
libre software is arguably no worse than the alternative of continuing
on the present path, and could well be better. A complementary response
(and the one ESR makes) is that increased use of libre software is
(directly or indirectly) opening up lucrative new business opportunities
that did not previously exist; that then leads to the question of which
new business opportunities will actually be most lucrative in practice.

ESR gives the example of a venture-funded startup providing 24/7 Linux
support (presumably Linuxcare); that particular example is open to the
criticism that a pure support business would not be that lucrative
because low barriers to entry could and would produce many able
competitors, driving prices and profits down.  An alternative argument
(again, made by Nicholas Petreley) is that libre software opens up
opportunities for the "sale-value" software industry because it provides
a new platform on which proprietary software can be sold, a platform not
controlled by Microsoft and hence potentially more profitable for
companies competing with Microsoft (e.g., Oracle, IBM, Corel).  Another
alternative argument is that lucrative opportunities exist for companies
not in the business of providing or supporting libre software but rather
using it as a major component in providing net-based services (e.g.,
Yahoo). (Now that I'm officially an AOL employee, this alternative hits
closest to home for me :-)

I think it's presumptuous to claim that any of these is _the_ best
alternative. However I do believe that (especially for the typical
programmer-as-employee) FSBs focused primarily on providing,
distributing, and supporting libre software will be potentially less
maximally lucrative than the other alternatives, unless the FSB has some
reasonable strategy that allows it to command higher-than-average
prices, e.g., specializing in a market niche with significant barriers
to entry (a la Cygnus in the embedded market), and/or incur
lower-than-average costs, e.g., using automation and economies of scale
to service more customers with fewer people.

Frank Hecker            Work:     Home: