cathedral bazaar

Introducing the paperThe Lessons

Criticism

The Cathedral and the BazaarPresentation about Eric Raymond’s paper

Juanjo Amor, Gregorio Robles

{jjamor,grex}@gsyc.escet.urjc.esGSyC/Libresoft

2-3 November 2007

Juanjo Amor, Gregorio Robles The Cathedral and the Bazaar


Criticism

(cc) 2007 Juanjo Amor, Gregorio Robles

Some rights reserved. This work licensed under Creative Commons Attribution-ShareAlike License. To view a copy

of full license, see http://creativecommons.org/licenses/by-sa/2.0/ or write to Creative Commons, 559 Nathan

Abbott Way, Stanford, California 94305, USA.



Criticism

Summary

Origins

The lessons

Criticism



Criticism

Origins

First version of the paper written in 1997.

Several revisions appeared up to 2000.

The author discovers a “development model” through thehistory of the Linux kernel and an own tool.

This model is presented as revolutionary, since it is useful tobuild large systems without apparently any or few organizationat all.



Criticism

The models

The Cathedral: The “classic” model.

Closed environment.Small group of leaders/developers.Only “stable” releases on, in some cases, “betas”.Used both in classic development models, such as waterfall,spiral etc; and in classic OSS projects.Examples: GCC, GNU Emacs.

The Bazaar: The model introduced by Linus Torvalds.

Open environment, almost any person can participate.There are no clear leaders, undefined number of developers.However, there is a benevolent-dictator figure.“Release early, Release often”.Examples: Linux.



Criticism

The surprise

The Bazaar style of development:

with a community seemed to resemble a large babbling bazaarof diverse agendas and approaches

with archive repositories where anyone can propose amodification

but from this, a stable and coherent large system emerges.

This was surprising:

Why Linux world did not fly apart in confusion?

and why Linux seemed to go from strength to strength at aspeed barely imaginable to cathedral-builders?



Criticism

Lesson 1

Every good work of software starts by scratching a developer’spersonal itch.

Most successful free software projects have been started bydevelopers with needs addressed by their “pet” project.

In the world of proprietary software, programmers spend theirtime building programs that they neither need nor want.

This motivation could explain the high quality of results givenby Linux.



Criticism

Lesson 2

Good programmers know what to write. Great ones know whatto rewrite (and reuse).

Linus Torvalds did not try to write Linux from scratch.Instead, he started by reusing Minix code and ideas.

Although today all reused Minix code has been removed orrewritten, while it was there, it provided scaffolding for theinfant that would eventually become Linux.

The source-sharing tradition of the Unix world has alwaysbeen friendly to code reuse.



Criticism

Lesson 3

“Plan to throw one away; you will, anyhow.”

from chapter 11, The Mythical Man-Month, by Fred Brooks.

We really do not understand the problem, until the firstimplementation is done.

So if you want to get it right, be ready to start over at leastonce.



Criticism

Lesson 4

If you have the right attitude, interesting problems will findyou.

Eric’s problem was that he needed a POP protocol client towork with.

And he found an abandoned one.

The problem was the continuation of the abandoned client,and Eric took it over and started to coordinate it.



Criticism

Lesson 5

When you lose interest in a program, your last duty to it is tohand it off to a competent successor.

Before abandoning the development of a free software, weshould find another person to continue its development.

Fortunately, in the bazaar world, some other hacker will findyour abandoned work soon, and will start to develop it for hisown needs.



Criticism

Lesson 6

Treating your users as co-developers is your least-hassle routeto rapid code improvement and effective debugging.

In Linux, many users are hackers too.

Because source code availability, these users can be effectivehackers.

This can be useful for shortening debugging time.

These users will diagnose problems, suggest fixes, and help inimprovements.



Criticism

Lesson 7

Release early. Release often. And listen to your customers.

This is another Linux characteristic: during very activedevelopment periods, lots of versions were released.

Sometimes, more than one in a day.

This maintains the hackers constantly stimulated andrewarded:

stimulated by the prospect of having an ego-satisfying piece ofthe action,

and rewarded by the sight of constant (even daily)improvement of their work.



Criticism

Lesson 8

Given a large enough beta-tester and co-developer base,almost every problem will be characterized quickly and the fix

obvious to someone.

“Given enough eyeballs, all bugs are shallow.” (named“Linus’s Law” by Eric S. Raymond).

Somebody finds the problem, and somebody else understands(and fixes) it.

The differences between the cathedral and the bazaar: In thecathedral-builder view of programming, bugs and developmentproblems are tricky, insidious, deep phenomena.

In the bazaar, bugs are generally shallow phenomena (theyturn shallow when exposed to a thousand eager co-developerscollaborating in next release).



Criticism

Lesson 9

Smart data structures and dumb code works a lot better thanthe other way around.

It is difficult to understand the code written by others,

but when we understand the data structures, understandingthe code is easier.



Criticism

Lesson 10: Eric tries the model

Eric S. Raymond tries successfully the model, with the followingprinciples:

Releasing early and often.

Adding to the beta list to everyone who contacted him aboutthe implemented program.

Announcing new releases to the beta list, stimulating peopleto participate.

Listening to beta-testers, polling them about design decisionsand taking in consideration patches and other feedback sentby them.

The consequence:

If you treat your beta-testers as if they’re your most valuableresource, they will respond by becoming your most valuable

resource.



Criticism

Lessons 11 and 12

The next best thing to having good ideas is recognizing goodideas from your users. Sometimes the latter is better.

Often, the most striking and innovative solutions come fromrealizing that your concept of the problem was wrong.

Eric learned this when saw good ideas for his applicationsuggested by users.

These ideas helped him to understand that, initially, he waslooking the solution for the wrong problem.



Criticism

Lesson 13

“Perfection (in design) is achieved not when there is nothingmore to add, but rather when there is nothing more to take

away.”

by Antonie de Saint-Exupery.

When the code is getting both better and simpler, that iswhen we know it is right.

At this moment, the software maintained by Eric was, notonly very different, but also simpler and better. It was time tochange its name and give it its new identity: “fetchmail”instead of “popclient”.



Criticism

Lessons 14, 15

Any tool should be useful in the expected way, but a trulygreat tool lends itself to uses you never expected.

When writing gateway software of any kind, take pains todisturb the data stream as little as possible – and never throw

away information unless the recipient forces you to!.



Criticism

Lessons 16, 17

When your language is nowhere near Turing-complete,syntactic sugar can be your friend.

A security system is only as secure as its secret. Beware ofpseudo-secrets.

These are lessons learned directly from the concreteapplication: fetchmail. Not related with SE.



Criticism

Lessons 18, 19

To solve an interesting problem, start by finding a problemthat is interesting to you.

Provided the development coordinator has a communicationsmedium at least as good as the Internet, and knows how tolead without coercion, many heads are inevitably better than

one.



Criticism

Criticism

No scientific rigor, because the document is an essay.

Not systematic.

Raymond tries to generalize the case of Linux to all freesoftware projects.

Some critics say that Linux is, in fact, an example of cathedralprocess: there is a leader, and a hierarchic structure of peoplewith delegated tasks. Also, responsibilities are distributed inthat structure although not explicitly.



Criticism

Criticism

Free Software = Bazaar?

Is there one development model for free software?Are all projects really with many contributions of *many*developers?

Bazaar definition may not be accurate.

Free Software is not bazaar.

Free software is software released with the four freedoms.



Criticism

References

Eric S. Raymond, “The Cathedral and the Bazaar” (1997-2006)http://www.catb.org/~esr/writings/cathedral-bazaar/

N. Bezroukov, “A Second Look to the Cathedral and the Bazaar”(1999)http://www.firstmonday.org/issues/issue4_12/bezroukov/


http://www.catb.org/~esr/writings/cathedral-bazaar/

http://www.firstmonday.org/issues/issue4_12/bezroukov/

cathedral bazaar

Technology