[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Open-Source Software Arouses Researchers' Curiosity



Virgina Postrel addresses the economic incentives of the free source
movement in this recent NYT article.


April 20, 2000
Open-Source Software Arouses Researchers' Curiosity
By VIRGINIA POSTREL


WHEN technology stocks took their sharp tumble last week, many companies
appeared to lose one of their most important assets -- the ability to lure
talented employees with options. To attract and hold the best, you have to
offer the chance to strike it rich.

Or do you? What are we to think when the best of the best -- the elite
programmers that industry wisdom deems 100 times more productive than the
typical competent coder -- donate their precious time to develop software
anyone can use without charge? That is the puzzle the open-source movement,
most famous for the Linux operating system, presents to economists.

Unlike most commercial software, which comes as incomprehensible "ones" and
"zeroes," open-source programs always include the source code, written in a
programming language like C++. Any programmer can look at the code and
change it to fix bugs or add features. Programmers contribute such patches
to the user community by sending them to a central authority, recognized as
the project leader. That person (or small group) decides which changes are
good enough to become part of the recognized version of the software that is
available to everyone. For instance, Linus Torvalds, the creator of Linux,
screens patches for the core portions of that operating system.

There are thousands of open-source projects, of varying popularity and
complexity. Aside from Linux, some of the most prominent include the PERL
language, Apache server software, the Emacs text editor and the Sendmail
e-mail routing program. Companies like Red Hat make money not from the code
itself, which customers could get free, but from the products or services
bundled with it.

The great advantage of open-source development is that the process taps
thousands of independent minds, making it much easier to find and repair
software bugs. "Given enough eyeballs, all bugs are shallow," writes Eric S.
Raymond, a programmer and movement observer, in "The Cathedral and the
Bazaar." (The essay is available at
http://www.tuxedo.org/esr/writings/cathedral-bazaar/cathedral-bazaar.html.

Open-source programming thus emulates the marketplace -- Mr. Raymond's
bazaar -- as well as the scientific peer-review process. Improvements emerge
incrementally, from an open-ended, decentralized process rather than a
central design. The software becomes more useful and reliable over time.

While its development looks like a marketplace, open-source software itself
is a classic public good. You can use it without contributing to its
maintenance and without paying a cent to all those programmers who created
and improved it.

Hence the economic puzzle. As Josh Lerner of the Harvard Business School and
Jean Tirole of the Massachusetts Institute of Technology ask in a recent
paper: "Why should thousands of top-notch programmers contribute freely to
the provision of a public good?" (Titled "The Simple Economics of Open
Source," the paper is available at http://www.people.hbs.edu/jlerner/.)

Professor Lerner says he is skeptical of the argument that says "this new
technology is so fundamentally different that all the old rules are off."
Incentives, the economists believe, are bound to matter. But if that is the
case, the question has ramifications beyond software, since many public
goods depend on tapping the time and talents of volunteers.

Of course, open-source programmers often reap some direct advantages from
their improvements. But solving their own problems has spillover benefits
for other people, just as fixing up a rundown building for your own business
makes the neighborhood nicer for everyone.

And a lot of people work on open-source projects they do not need for their
jobs. One reason is that it is fun and interesting, as much recreation as
work. That is true of many other public-good endeavors, from planting trees
to running charity balls. They are not all selfless drudgery. Where virtue
must be entirely its own reward, volunteers will be hard to find.

But fun cannot explain everything. After all, many open-source hackers are
already paid to program, and they do not do that for free. Mr. Raymond
argues that open-source developers work for status among other programmers.
If you do good work, you become a big shot. Your name is forever associated
with the code you have contributed, and because the code is open, everyone
can see just what you have accomplished.

Hackerdom is a ruthless meritocracy, dedicated to winnowing out bad code and
spreading good. A lesson for other voluntary endeavors is that praising
people just for showing up will not attract the best talent. To be valuable,
reputation and recognition have to signal something significant.

Along these lines, Professors Lerner and Tirole add a pecuniary motive --
one that traditional charities might call "networking." Contributing to
open-source projects can enhance a programmer's career prospects. Especially
for someone living and working in an out-of-the-way place, open-source work
can attract the attention of potential employers or financial backers. The
most striking example is Mr. Torvalds, who was a graduate student in Finland
when he released Linux and is now a highly paid executive at a Silicon
Valley start-up.

Professors Lerner and Tirole's signaling story assumes that the programmer
already has a portfolio of skills and that open-source development simply
offers a means to display them. But volunteering to work on difficult
problems is also a way of honing new skills -- building your human capital,
as an economist would put it -- and the open-source community offers
feedback that encourages learning. This, too, is a form of career
enhancement.

Open-source programmers are an interesting bunch of volunteers because they
are among the most sought-after employees in the economy, and because their
volunteer work has produced lucrative commercial spinoffs. But software is h
ardly the first public good provided through private efforts. While it may
sound crass to consider them, every voluntary endeavor needs incentives for
excellence. The public good often depends on respecting private gains.

http://www.nytimes.com/library/financial/columns/042000econ-scene.html