Replying to a question on
Nettime
on "open source and implementing this metaphor to a curatorial
practice":
"Curatorship" remains a problematic term not only in this context,
self-organization may be more appropriate, but extrapolates the systems
beliefs within Internet culture. ESR's text needs, on the one hand, to
be read its historical context of optimistic 1990s Internet cultural
visions of "crowd wisdom", "collective intelligence", "smart mobs" etc.;
cybernetic memes as much as variations of the liberal tropes of the
"invisible hand" (A. Smith) and "open society (Popper). In Raymond's
text - which has been overrated, but is nevertheless a historical
document -, the "bazaar" is first of all a systemic free market metaphor.
Linux, more recently, Wikipedia and other phenomena show that "critical
mass" theories are not completely off. The issues are, in essence, the
same as with all consensus-based projects - such as architectural
vision: Linux reimplemented Unix instead the Plan9 or Lisp Machine
kernel architectures simply because Unix kernel architecture is c.s.
textbook knowledge. Correspondingly, Wikipedia implements the most
clearly consensus-based form of writing, the general encyclopedia.
(Still, its value lies in the frequent eccentricity and obscurity of
phenomena it tracks, unless this is been stifled by angst-ridden
editorial self-control.)
That "open collaboration" is not a magic bullet, and "open curatorship"
is older than "Open Source", may best be studied in the Mail Art
network, beginning with Ray Johnson's New York Correspondance School in
the 1960s, and with the festivals and non-juried exhibitions of previous
avant-garde art movements as yet an older pretext. Bob Black said
everything that needs to be said about Mail Art when comparing them to
the Paralympics, i.e. a seemingly alternative but really just parallel
system to the established system [hard to avoid the term here] based on
its own - quantitative instead of qualitative - logic of reward and
punishment.
Obsessed with egalitarianism, the Mail Art network required
to never reject any contribution to an open-call project, despite the
known and often enough deplored "junk mail" phenomenon. It ultimately
renders "Mail Art" yet another cybernetic systems-obsessed art
paralleling the decline [or rather: continually present, but ultimately
dominant aspect) of Fluxus into the "intermedia" laboratory art
described in S. Youngblood's "Expanded Cinema" (1970) [H. Flynt's
criticism of Fluxus].
"Self-organizing systems" up your's.
Back to "The Cathedral and the Bazaar", it is, like Barthes' "The Death
of Author", a text that nobody has read yet everybody has an opinion
about. Contrary to popular belief and urban myths, it does not truly
pitch an Open Source "bazaar" model against a proprietary Microsoft-ish
"cathedral" model of software development, but analyzes the
decentralized development of one specific piece of software, the Linux
kernel supervised by L. Torvalds. The urban myth probably originates in
the fact that non-technical readers are unlikely to understand that it
is not about [what is commonly called] the "Linux operating system" as a
whole. In fact, the classical "cathedral" model of software development
in small, closed committees had been characteristic among others for GNU
software, the free BSDs and the X Window System, i.e. all the base
components of a typical "Linux distribution" except for the kernel
itself.
Ten years later, a clear-cut division of "bazaar"- and "cathedral no
longer exists in Free Software development: The development of the Linux
kernel has become more hierarchical while the development of GNU and BSD
software has become more distributed and adapted to the Internet. (Viz.
the now-standard use of networked version control systems.)
While not using the term "Open Source" in its initial version, the essay
preempts the later Open Source-vs.-Free Software debate by discussing
open, distributed development processes as technically superior to
closed processes. This is its main point [with, as pointed out, striking
similarities to Bertalanffy's theory of open systems and Popper's theory
of the open society as the counter-model to societies founded on
philosophical idealism.] Again: While the distributed model has its
advantages - most obvious in the fact that, thanks to BSD, GNU and
Linux, Unix hasn't died, but improved and blessed us with tools that
don't offend the human intellect such as zsh and vim never mind the
complete lack of proprietary commercial interest in developing such
software -, it is not the answer to all questions. Raymond's conclusion that
"given enough eyes, all bugs are shallow", is a bit loudmouthed
considering - for example - the issue of MD5 hash collisions.
The reverse is true as well: If there are not enough eyes, bugs can bite
you, for example in FLOSS multimedia authoring software from Cinelerra
to PD with its minuscule communities of often non-professional
programmers.
All critique of "open systems" ideology pales, however, in comparison to
the issues of (contemporary visual) art. Art literally wears the
emperor's new clothes, and suffers from a severely if not pathologically
distorted self-perception of its actual contemporariness. It is the only
of the modern arts that is still structurally feudalist, with an economy
firmly based on the notion of one material fetish object, with
reproduction - unlike in books, music records, films, software - being
merely a second-rate, plebeian illustration of the aristocratic
"original". It is financed by the modern successors to the old feudal
authorities; back then, the church and the courts, today, the rich as
the successors to the aristocracy and the state as the grant-giving
successor to the church.
