Describing Open Source development as cathedrals vs. bazaars has been quite unsatisfactory to many of us ever since Eric Raymond first wrote about it in 1997. It was a good conversation starter, but certainly not the final word on the matter. Perhaps the best distinction to be made is the difference between the restricted-access, high ceremony organization (the priests in the cathedral), and the ad hoc, self-organizing low ceremony group (the merchants in the bazaar). So, really, the paper should be entitled "The Priests and the Merchants".
The distinction may sound subtle, but it is not. A medival cathedral -- the edifice itself -- was built by the local town, and took many decades (sometimes more than a century) to build. Its construction was a collaborative effort. The early builders surely did not live long enough to see the final product. With such a long time to complete and so many hands helping, the overall structure de-emphasizes individual contributions and focuses instead the accomplishments of the group as a whole.
Bazaars, on the other hand, are temporary, ad hoc constructions that have little to no lasting value. They are built, serve their purpose, and torn down. In fact, it is preferable to have a central market square that is empty at 6am, converted into a bazaar by 9am, and returned to its former empty state by 6pm -- with no remnants of the bazaar to be found. If there is more stuff to trade, come back tomorrow and we'll build another one.
So bazaars are temporary organizations that are formed to serve a single time-limited function, while cathedrals are works that take a long time to build and have lasting value. Today, we can build a church -- even a cathedral -- in considerably less time than our medival predecessors, but we still have long term building projects today. Many cite the LA Highway system as a modern cathedral. The internet serves as another example.
I have been thinking about these things for a while now It came to me again most recently when I read this letter to Tim O'Reilly asking "when will Perl 6 be done?". Simon Cozens answered the letter as the editor of Perl.com, saying (among other things):
You said that Perl 6 is "beginning to look like a project with no end;" this is dead right--Perl 6 is just the next version of Perl, a project started over ten years ago and, we hope, with no end to its future. As Damian says, "Perl 5's first production release took the best part of a decade (counting from Perl 1), so if Perl 6 takes less than that I think we're doing okay."Perl is a Cathedral. It is our Cathedral.
Re:cathedrals are for worshipping in
chromatic on 2004-08-15T03:02:11
Honest answers not analogies would be more effective in assuaging their concerns.As if saying "We'll finish Perl 6 sooner when more people write code than complain that it's not done yet" has helped yet.
Re:cathedrals are for worshipping in
hfb on 2004-08-15T09:12:41
No, because many of the people who might have done so have already given up and moved on. P5P's population seems to be disappearing as well.
Lofty metaphors do nothing for drawing a realistic picture as you can't wish it into being. I don't care at all for perl6 and it could arrive in the, more realistic, year 2035 or not at all for all I care, but the rift it has created due to just this kind of fanciful wishful thinking. What's so hard about telling it like it is? Telling people not to complain is just a republican bait and switch smokescreen by not answering the question and trying to shame the querant into silence. People want straight answers, not cathedrals in the clouds.
Re:cathedrals are for worshipping in
ziggy on 2004-08-15T22:32:11
P5P had a certain level of activity before the Perl 6 announcement. There is neither an expectation nor a guarantee that the level of participation on P5P at the time was sustainable, or even likely to increase. Furthermore, it's disingenuous of you to imply that the health of P5P is some barometer of the health of the Perl Community as a whole. While P5P is important in its own right, the center of gravity has long since left P5P and moved onto CPAN -- just as Larry said it should back in 1998.No, because many of the people who might have done so have already given up and moved on. P5P's population seems to be disappearing as well.Open source has always been a revolving door. People get jobs, get married, have kids and other demands on their time that often prevent them from working on open source projects as much as they would like to do otherwise. Sometimes they lose interest, too. But to say that all of the ills you personally perceive in Perl today to be caused solely by Perl 6 is just nonsense. Reality is much more complex.
Re:cathedrals are for worshipping in
hfb on 2004-08-16T08:47:53
Talking heads always seem to have the answers and
,hey, a revolving door seems like the ticket for this one. I think it's an insult to those who actually have contributed in the past but who have moved on coming from someone who hasn't actually been a patching member of P5P. I speak of P5P today, not 1999. And, regardless of whether or not the people who might actually write the code for P6 will come from P5P, there are serious problems, problems noone wants to talk about publicly.
Reality is much more complex, but who would tell that story, eh? There are major issues with P6 and people, when they ask, deserve something better than some high falutin' bullshit about cathedrals and bazaars.
Re:Bwana ESR
hfb on 2004-08-12T22:53:23
and a badass combover with dating tips.:)
