Cathedra and the Bizarre: Why Free Stuff is Good
Tim May
timcmay at got.net
Sat Sep 6 19:32:10 PDT 2003
[Note: This is just a typical list article. I make no claim that this
is polished and is some article along the lines of Eric Raymond's
"Cathedral and the Bazaar." My punning title is not meant to imply
this. I've had a glass of wine, have something I want to say, and am
saying it. It may even overlap with Raymond's article...it's been years
since I've looked at it. Don't forward this article to other lists, OK?]
I'm not much interested in the machinations of the micropayment/digital
cash markets, for reasons which should be clear. But I've certainly
seen something very similar being played-out in the _language_ market
(what marketdroids would call "space").
All of the interesting languages now generating a lot of buzz, and
substantial communities, are essentially "free." Or non-profit, or open
source, or whatever one wishes to call them. Some examples:
* Python. I'm not a Python person, but a lot of my friends are. And the
community is substantial. Could this have come out of a corporate
division with a committee of corporate managers and some grunts to do
the implementation?
* Perl. An earlier success in the same tradition.
* E, the capability-oriented language. Again, free.
* The last four languages I have downloaded and spent time with are all
free:
- Squeak, a version of Smalltalk, available on nearly every platform
(generates bytecodes read by a VM, so writing the VM for a platform
means Squeak runs on it). Associated with the Xerox PARC group which
developed Smalltalk (including Alan Kay, Ted Kaehler, Dan Ingalls,
etc.), then by Apple, then Disney, now scattered. Nominally open-source.
- OCaml, the object-oriented version of Caml, standing for
"Categorically-abstract ML"). A variant of ML (Milner's "meta
language"), coming mostly from France. (One of the CPunks list members
who cracked a famous challenge, Damien D., was one of the principal
developers. Julian Assange, also a list member at one time, was also
involved.)
- Haskell, the category theory-oriented pure functional language.
Mostly associated with Yale University, Glasgow U., Oregon Graduate
institute, Chalmers in Sweden, and many hundreds of academic
researchers.
- Self, the derivative of Smalltalk developed by several
Stanford-affiliated people. Now mostly at Sun, with parts of the
technology in Java Hotspot, and in the free language Squeak, a version
of Smalltalk.
There are other examples. The obvious example is Unix, in its variants
including Linux, despite the attempt by SCO to collect $1000 per CPU or
whatever silly number they have floated in their lawsuits.
We free marketers, we believers in profits and property and contracts,
we must acknowledge that most of the really good stuff in computer
science, aside from the hardware, is often free. (Caveat: I use and
enjoy immensely Mac OS X, which is not free. The software itself is
very _close_ to being free, but the link to Mac hardware costs a bit
more. Still, I enjoy it enough to pay for it. Nearly everything else I
use on my Macs is free or very, very close to being free.)
But "free" arises for some reasons which are readily-understandable to
Hayekians and Randians and those interested in markets and capitalism:
* the creators are anticipating rewards _other_ than salaries from
employers, e.g.,
-- fame ("Yes, I am Guido")
-- job opportunities ("I wrote Digital Datawhaque, the leading open
source frobbolizer")
-- publish or perish
-- simple pleasure or some mission (applies to several Cypherpunks
projects...)
Others have written at great length about how Unix took over the world
over the past 25 years, about why Unix won out over VMS and a dozen or
two other proprietary OSes.
I'll look at just one language, one of the the four above: Haskell.
(For the curious, www.haskell.org.)
The developers of Haskell don't get paid through sales. They are
professors, associate professors, lecturers, their students, and anyone
trying to "make a mark" in computing. They are generally brainier than
mere engineers or programmers hired by BigCo Software Factory, Inc. to
produce software. They write papers, which are peer-reviewed and which
are mercilessly criticized if flaky. So the big names in Haskell are
the people who contributed really important ideas.
(The same applies to Python, OCaml, etc.)
Sometimes someone tries to do such a thing "for profit." In the "space"
of functional or OO languages I talk about above, two such examples
stand out: Erlang and Clean. Erlang is an in-house, proprietary system
at Ericsson, the cell phone company, and Clean is a product of some
German or Dutch company. As neither are readily downloadable, the
academic and student communities with each are virtually nil (pun
intended).
The large communities, and probable large adoptions by corporations
later, are in the free stuff areas. I don't even think the important
defining characteristic is that the thing be "open source." The
important thing is "free." Free as in no hassles, no licenses, play
around, copy it for your friends, write about it without fears of being
contacted by lawyers, and so on. Free. Unencumbered.
(Yeah, there are various kinds of licenses having to do with whether
products based on the freebie can be sold for profit. Another topic. To
first order, the important reason people play with the things is
because they are free. And to students and home users, this is what's
important.)
Does this have any relevance to Cypherpunks?
Just the obvious one: any digital money system needs to be free, or
open source, to be widely adopted by our kind of people.
Any money which is later made will likely come from one of the above
reasons (fame, job opportunities, publish or perish, or
pleasure/mission) or from being in a position to exploit the
technology, e.g., by operating a digital money system.
Attempting to hold onto the "intellectual property" (cough) and then
profit from carefully licensing it out to others is usually a lose. The
results with Digicash speak for themselves. Had David Chaum, a man I
respect a great deal, freely published and distributed his ideas, he
would likely today have a lot more fame and fortune.
--Tim May
More information about the Testlist
mailing list