:46:05
The next major event was one that
I had directed hand in.
:46:10
I wrote a paper, called
"The Cathedral & the Bazaar".
:46:14
which was my observations,
my anthropological analysis
:46:18
of what it was that made
the open source world work.
:46:21
We didn't call it that then. We were still
using the term "Free Software" primarily.
:46:27
So it was my observation of
what made the Free Software world work
:46:30
and why we were able to
produce extremely high quality software
:46:36
in spite of constantly violating all of
the standard rules of software engineering
:46:42
In that paper, I was setting up a contrast
:46:45
between two different styles of development,
:46:48
two opposed styles of development.
:46:50
One, which is the
conventional closed development style,
:46:57
which I called the "Cathedral" style.
:47:00
In that one, you have
tight specification of objectives.
:47:04
Small project groups which are run
in a fairly hierarchical authoritarian manner.
:47:11
And you have long release intervals
:47:15
On the other hand,what I identified
is happening in the Linux world
:47:19
was a much more peer to peer decentralized,
market or bazaar-like style,
:47:26
which has a very short release intervals
:47:28
and constant solicitation of feedback from people
who are formally outside of the project.
:47:34
A very intense peer review process.
:47:39
And the startling thing was that the more I looked at this,
:47:42
the more it seemed that trading away
all the supposed advantages
:47:48
of conventional closed development,
:47:51
for that one single advantage
of massive independent peer review
:47:55
actually seemed to win,
actually seemed to get you good results.