By David Warsh, Globe Columnist, 12/13/98
He had been mowing lawns in suburban Newton after getting his degree in
classics from Harvard, according to company legend. Then a friend drafted him
to write an instruction manual for the Unix computer operating system. Unix
had been written by the telephone company, which was prohibited by an
antitrust agreement from selling it. Its authors happily gave the code away to
anyone who asked, but nobody at Ma Bell got paid to explain how it worked.
He set up a picnic table at a trade show and went home with a wad of bills
the size of a roll of toilet paper. That was Tim O'Reilly's initiation, 20
years ago, to the mysteries of open-source software.
Today, he is sole proprietor of a $35 million-a-year publishing empire with
headquarters in the seaside town of Sebastapol, Calif., and offices in Paris,
Koln, Germany, Tokyo, and Cambridge. O'Reilly and Associates published far
and away the most influential titles among how-to books about the Web -- so
well-branded that they rate a special kiosk in many stores.
More to the point, O'Reilly is the proprietor of the view du jour of the
Next Big Thing. It may even be the correct one. In a 31-page essay last month
he turned his spotlight on the greatest threat to Microsoft's dominion of the
market for personal computer and Internet software.
The really, truly Next Big Thing has not much to do with the merger of
America Online and Netscape Communications, and their alliance with Sun
Microsystems. Instead, he says, it's a social movement known as
open-sourcing, meaning the widespread and uncoordinated collaboration of
programmers using freely distributed source code and the communications
facilities of the Net.
However obscurely named, open-sourcing possesses an economic logic so
powerful that O'Reilly thinks it could transform the software industry so
completely that, in the end, it would bear a stronger resemblance to the
social organization of the scientific community than to, say, the industrial
organization of the oil business.
After all, it was the ethos of science -- ``no kings, no priests, just a
rough consensus and running code'' in the phrase of one pioneer -- that
governed the Internet Engineering Task Force. That obscure peer-selected
planning group oversaw the creation of the Internet. It had a natural
affinity for ``the rough-and-tumble search for optimal solutions'' that is the
essence of the spirit of openness. Its ultimate expression -- so far -- is the
ultrademocratic collaborative medium known as the World Wide Web.
If open-sourcing continues to spread -- and O'Reilly shows that the
movement already is well under way -- then you can relax in some deep sense
about Bill Gates being the greatest threat to the planet after global warming.
The transition from the age of software to the age of infoware will be at
hand.
The issue surfaced four-square last month in an article O'Reilly wrote
about ``The Open-Sourcing Revolution,'' which filled an issue of Esther
Dyson's celebrated newsletter, Release 1.0.
This in itself was an example of open sourcing, for ordinarily Dyson's
letter is written by Dyson and her hirelings; O'Reilly was doing a guest shot.
No money would have changed hands; in a way, that is precisely O'Reilly's
point. Such is the ruthless efficiency of the Internet that you, too, can read
the piece at www.edventure.com.
release1/1198.html -- for free!
The essence of O'Reilly's argument is that the relentless commodification
process at work in computing for 30 years is about to enter a new stage. What
yesterday was proprietary and expensive is cheap and widely available today.
Thus in the 1950s and 1960s, IBM dominated through hardware. Software was
written by hardware suppliers, or by captive vendors. Barriers to entry were
high.
Then the personal computer took the computer out of glass houses and put it
on desktops. Suddenly barriers to entry were low. Software entrepreneurs like
Gates and Mitch Kapor made fortunes by creating brand names that came to
dominate their market applications.
Before long, however, a few software products had become hugely complex,
requiring the efforts of thousand of programmers writing for years.
Microsoft's Windows NT operating system is a good example. No longer was it
possible for a solo programmer to have much influence; he or she needed a
venture capitalist and a team.
Enter the code-sharers. The source codes of these huge commercial programs
-- their tightly written underlying mainsprings -- are their most jealously
guarded secrets, the very heart of their commercial advantage. But the secrecy
made it impossible to modify these programs, which were inelastic and often
error-prone.
To be sure, the thoroughly open -- and therefore modifiable -- Unix program
had been developed at huge expense and then given away by AT&T, starting in
the 1970s. But who today could afford to write -- and then simply give away --
operating systems that might challenge the dominance of commercial products?
A Finnish graduate student named Linus Thorvald, for one. In the early
1990s, Thorvald stepped forward with a home-cooked operating system kernel
that he distributed freely over the Internet. Within a few years, the Linux
system, extensively improved and expanded by hundreds of computer PhDs, was
employed by an estimated 7 million users -- most of whom affirm its
wide-ranging superiority to Windows NT.
O'Reilly writes, ``The open-source paradigm once again lowers the barriers
to entry. You can take somebody else's product and build your own custom
version of it. It lets you scratch your own itch . . . If somebody doesn't
like a feature, they can add to it, subtract from it, or re-implement it. If
they give their fix back to the community, it can be adopted widely very
quickly.''
And now Red Hat Software has gone into business distributing a basic Linux
package for the bargain sum of $49.95. Bob Young, the chief executive, says
his goal is ``to shrink the dollar value of the operating systems market''
from Microsoft's billions to $500 million -- of which he hopes to command the
lion's share.
There are other products, O'Reilly notes. The Free Software Foundation's
GNU project distributes a set of programmer's tools for open source systems.
The Perl language, created by Larry Wall (now an O'Reilly employee) is widely
used; so are tcl and Python.
The Apache Group, a dozen core developers and their surrounding user group,
has achieved a dominant share among Web servers, despite having no more
corporate organization than a shared Web site. Sendmail -- backbone of the
Internet's e-mail infrastructure -- is an open system. So is Samba, a
``stealth'' technology that permits administrators to work Linux applications
into their Microsoft systems.
None of this is lost on Microsoft, which, after all, has made a series of
dramatic and so far successful adjustments to the sudden rise of the Internet.
An internal Microsoft memo analyzing the threat posed by Linux and open-source
projects in general -- acknowledged authentic by the company -- has turned up
on the Web (www.opensource.org/
halloween1.html), extensively annotated by open-source enthusiast Eric
Raymond.
Will the open-source argument really overtake the commercial software
industry? The deeper beauty of O'Reilly's argument is the parallel he draws
between the emergence of open science in the Renaissance and among the
community of computer users today. The imperative to share knowledge in order
that it be tested is the same in each. ``Where scientists talk of replication,
open-source programmers talk of debugging,'' he says. The point is to get it
right and move on. He quotes Grace Hopper, inventor of the compiler: ``To me
programming is more than an important practical art. It is also a gigantic
undertaking in the foundations of knowledge.''
Most of us spend our lives and never note that the great delicate
apparatus for the production and distribution of knowledge that is science
ultimately has nothing to do with the market -- perhaps because so relatively
few people are involved. Indeed, the organization of science in certain ways
is just the opposite of the marketplace -- it is designed to free some people
to pursue their curiosity, wherever it leads. Yet science has made us rich.
If the passion of the large community of geeks, nerds, wonks, and other
computer programmers who are interested in solving problems of their world
succeeds in freeing themselves from Microsoft's business model, we will
benefit similarly -- and so much the worse for Bill Gates.
