Archive of UserLand's first discussion group, started October 5, 1998.
Re: Thoughts on ESR and Perl
Author: Eric Kidd Posted: 8/27/1999; 8:06:30 AM Topic: Opening Up Linux Journal and O'Reilly Msg #: 10103 (In response to 10084) Prev/Next: 10102 / 10104
That was my view too. He was describing, in an incomplete way, how I work.I suppose I should say something here, for the sake of clarification. The Cathedral and the Bazaar wasn't written to compare proprietary software with open source software. Instead, it was written to compare two ways of building open source software.
To over-simplify the issues a bit, it was written to compare the way Linus runs projects with the way Richard Stallman runs projects. Linus was open, open, open. He let anybody work on anything, merged in whatever set of changes he liked best, and repeatedly offered to step down if anybody could do a better job. He made a new release several times a week (sometimes several times a day). He pointed out that he had no more legal authority than any other contributor, and that people could ignore him if they wished.
Richard Stallman is a very capable programmer, but he used to run projects in a very hierarchial style. Only a tiny number of very dedicated volunteers ever got to see the code, and new releases were made when everything was finally done. When ESR writes about "individual of wizards or small bands or mages working in splendid isolation, with no beta to be released before its time", he's talking about the development teams for the GCC compiler and the HURD kernel. Trust me--I've worked with the Free Software Foundation, and ESR's descriptions are very insightful.
The big question a few years back was why Linus succeeded where Stallman failed. Stallman had been working on a very advanced kernel called the HURD, but it took him many years (and it still isn't done). Linux was a much less ambitious project, but it quickly surpassed the HURD in quality and features. Why did one project succeed where the other failed? That's the question ESR originally tried to answer in The Cathedral and the Bazaar.
What relevance does all this have to the regular software world? Well, open source projects run according to the "Cathedral" model can't really compete with paid developers, at least in the long term. (In the short term, some of these projects do quite well, largely thanks to the skills of people like Stallman.) But projects run according to Linus' "Bazaar" model seem to be able to write millions of lines of high-quality code in a relatively short period of time.
In truth, Dave, The Cathedral and the Bazaar has almost nothing to say about small commercial developers. It's about managing really large projects with hundreds of contributors. Everybody knows (or so I hope) that a six person software company can work just fine--there's lots of evidence of this fact. It's the two-hundred person development teams which seem to have trouble.
Cheers,
EricA note about the Free Software Foundation: Stallman is a smart guy, and he listens to criticism. The GCC compiler is now maintained by a bazaar-style group, and the HURD has been placed on the back burner in favor of Linux. These days, Stallman maintains Emacs and lends help to hundreds of software projects connected to his foundation. Without his efforts over the years, much of the most important Linux software would never have existed.
This page was archived on 6/13/2001; 4:52:12 PM.
© Copyright 1998-2001 UserLand Software, Inc.