(fwd) article on open source

Mike Bilow mikebw at colossus.bilow.com
Tue Mar 28 07:10:33 EST 2000


First, of all, unless things have changed radically, we need to remember
that Boardwatch is run by a lunatic who, as I recall, thinks cigarette
smoking is a good thing.  One can, of course, defend it as a personal
freedom to kill oneself in the most gruesome way possible, but it seems
odd as a philosophy.  A large part of the publication used to be taken up
by the publisher attacking and ridiculing his readers who sent letters for
publication, but I have not read Boardwatch in years and I am not sure if 
this rather idiosyncratic tradition has continued.

The idea of compensating programmers for "free" software on a royalty
basis is really crazy.  As Richard M. Stallman said, "Think 'free speech,'
not 'free beer.'"  The goal of "free" software is one of openness and
equal access to information, not saving or making money.

Milton Friedman, of all people, used to give an example of how difficult
it is to get people to cooperate on even the simplest things, and of why
"planned economies" are therefore doomed to fail.  He would hold up a
pencil and talk about all of the different material components -- wood,
graphite, paint, metal, plastic -- and all of the different labor
contributions -- lumberjacking, sawmilling, mining, sanding, painting,
finishing, transporting, inventorying, storing, selling -- which are
necessary for getting a pencil to the consumer.  Now, pencils are not
exactly free, but they are pretty close to it: no one thinks about the
cost of a pencil, yet it quite literally requires the active cooperation
of thousands or perhaps tens of thousands of people across several
continents to make one.  (Freidman's example of a pencil was not
arbitrary.  One of the ways the industrialist Armand Hammer, who was
closely associated with the Soviet government, made his fortune was by
obtaining the pencil importing monopoly concession from Stalin.  The
Soviets could have made their own pencils, but they would have been
prohibitively expensive.)

People choose to participate in the production of "free software"
precisely because they feel they are adequately compensated, monetarily or
otherwise, for doing so.  Imposing some sort of regime where lines of code
are counted as an absolute measure of productivity is somewhere between
deeply misguided and utterly insane, as it would be exactly the sort of
draconian mechanism which is anathema to those who choose to particpate.
Many people are monetarily compensated for contributing to free software,
some directly through being employed by Red Hat or other such companies,
and some indirectly through using free software in their jobs.  There are
also academics who actively contribute -- FSF itself is heavily connected
with MIT and has been hosted on their network since inception -- and it is
very hard to measure the value or productivity of academics monetarily.

This connection with academia is not a coincidence, as the economic models
are somewhat similar.  How much is a research paper worth?  Should we pay
academics on the basis of how many papers they publish, or by the page?
Everyone expects academics to be paid for their work, and we acknowledge
that it is often extremely valuable, but this does not imply that we have
any real way to measure or estimate that value.  If you try to apply
traditional "pure capitalist" methods to the assignment of value, you end
up with ludicrous proposals on the order of auctioning off publication
rights for research papers to the highest bidder.

The key link between free software and academia is the critical importance
of what economists call the "network effect."  This is a way of explaining
how value can arise from the actions of others relative to you.  For
example, your telephone has value because you can use it to call other
people and because they can use it to call you.  A telephone is worth
something to each subscriber only because other subscribers pay the
telephone company to be connected, too.  It is this collective action of a
large number of people, each acting out of individal self-interest, which
makes the telephone network valuable.  If we tried to implement a closely
tied fee-for-service compensation model for the telephone network, it
would become useless.  If everyone had a 900 number-like billing
arrangement where they were compensated for their time talking to whomever
called, the value of the network as a whole would plummet to zero.

This is why "digital cash" or "microcash" proposals have all met with
failure: the obstacles are intrinsic and economic, not technical.  If
someone really does implement one of these nutty schemes where we pay
three cents to look something up in a search engine or eight cents to read
a news article, their competitors who remain free will eat their lunch.
With pencils, there is a real non-zero minimum cost for which pencils can
be made and shipped and stored, but with information the unit cost really
is effectively zero.  This is what is really meant by the famous (and
often misunderstood) maxim "Information wants to be free."

Regardless of whether "lines of code" is a legitimate metric for value --
and I certainly do not think that it is -- the need to pay for software at
such rates eventually degenerates into a nuisance fee.  Most importantly,
introducing monetary compensation would restrict the free and unlimited
distribution of this software, destroying the whole point of the effort.
Companies who now turn a blind eye to employees contributing code to free
software projects would insist on a piece of the action, creating a
disincentive to contribute.

Even if this could be done, it would not be fair.  Free software depends
upon the efforts of many essential people who do not write code, including
those who carefully debug code and those who write documentation.  These
are also much less prestigious roles in the free software community,
making them harder to fill and therefore more in demand.  It can take days
of debugging to find the one mistyped semicolon; how much do you pay for
that?  If you pay programmers for "lines of code," do they have to pay for
debugging out of their own pockets?

Finally, there are many important free software contributors whose main
contribution is so intangible as to render attempts to quantify their
influence in economic terms as absurd.  People such as Richard M. Stallman
or Linus Torvalds have contributed primarily through force of personality.
Both may be decent programmers, but their essential contributions have
been providing leadership.  Nearly any competent programmer could write
Linux, given enough time, but only Torvalds has been able to convince a
few hundred people to work on it together within an entirely informal
organizational structure that would horrify those who are comfortable in
the corporate environment.  I would venture to say, with confidence, that
never in the history of the development of Linux has any developer ever
been asked by Linus to write a status report.

-- Mike


> ------- Forwarded Message
> 
> From: TonStanco at aol.com
> Message-ID: <3b.2a83ebf.260f5397 at aol.com>
> Date: Sun, 26 Mar 2000 06:50:47 EST
> Subject: article on open source
> 
>  These are follow-up questions to the one I posed to the open source community
>  for my proposed article in Internet.com's Boardwatch magazine on a royalty
>  system to pay developers in open source. I'd like to get community feedback
>  on these, too.



-
Subcription/unsubscription/info requests: send e-mail with
"subscribe", "unsubscribe", or "info" on the first line of the
message body to discuss-request at blu.org (Subject line is ignored).



More information about the Discuss mailing list