1010@unknown@formal@none@1@S@The Cathedral and the Bazaar@@@@1@5@nil@oe@24-11-2010 1020@unknown@formal@none@1@S@Eric Steven Raymond@@@@1@3@nil@oe@24-11-2010 1030@unknown@formal@none@1@S@Thyrsus Enterprises@@@@1@2@nil@oe@24-11-2010 1040@unknown@formal@none@1@S@This is version 3.0@@@@1@4@nil@oe@24-11-2010 1050@unknown@formal@none@1@S@Copyright © 2000 Eric S. Raymond@@@@1@6@nil@oe@24-11-2010 1060@unknown@formal@none@1@S@Permission is granted to copy, distribute and/or modify this document under the terms of the Open Publication License, version 2.0.@@@@1@20@nil@oe@24-11-2010 1070@unknown@formal@none@1@S@Abstract@@@@1@1@nil@oe@24-11-2010 1080@unknown@formal@none@1@S@I anatomize a successful open-source project, fetchmail, that was run as a deliberate test of the surprising theories about software engineering suggested by the history of Linux.@@@@1@27@nil@oe@24-11-2010 1090@unknown@formal@none@1@S@I discuss these theories in terms of two fundamentally different development styles, the ``cathedral'' model of most of the commercial world versus the ``bazaar'' model of the Linux world.@@@@1@29@nil@oe@24-11-2010 1100@unknown@formal@none@1@S@I show that these models derive from opposing assumptions about the nature of the software-debugging task.@@@@1@16@nil@oe@24-11-2010 1110@unknown@formal@none@1@S@I then make a sustained argument from the Linux experience for the proposition that ``Given enough eyeballs, all bugs are shallow'', suggest productive analogies with other self-correcting systems of selfish agents, and conclude with some exploration of the implications of this insight for the future of software.@@@@1@47@nil@oe@24-11-2010 1120@unknown@formal@none@1@S@The Cathedral and the Bazaar@@@@1@5@nil@oe@24-11-2010 1130@unknown@formal@none@1@S@Linux is subversive.@@@@1@3@nil@oe@24-11-2010 1140@unknown@formal@none@1@S@Who would have thought even five years ago (1991) that a world-class operating system could coalesce as if by magic out of part-time hacking by several thousand developers scattered all over the planet, connected only by the tenuous strands of the Internet?@@@@1@42@nil@oe@24-11-2010 1150@unknown@formal@none@1@S@Certainly not I.@@@@1@3@nil@oe@24-11-2010 1160@unknown@formal@none@1@S@By the time Linux swam onto my radar screen in early 1993, I had already been involved in Unix and open-source development for ten years.@@@@1@25@nil@oe@24-11-2010 1170@unknown@formal@none@1@S@I was one of the first GNU contributors in the mid-1980s.@@@@1@11@nil@oe@24-11-2010 1180@unknown@formal@none@1@S@I had released a good deal of open-source software onto the net, developing or co-developing several programs (nethack, Emacs's VC and GUD modes, xlife, and others) that are still in wide use today.@@@@1@33@nil@oe@24-11-2010 1190@unknown@formal@none@1@S@I thought I knew how it was done.@@@@1@8@nil@oe@24-11-2010 1200@unknown@formal@none@1@S@Linux overturned much of what I thought I knew.@@@@1@9@nil@oe@24-11-2010 1210@unknown@formal@none@1@S@I had been preaching the Unix gospel of small tools, rapid prototyping and evolutionary programming for years.@@@@1@17@nil@oe@24-11-2010 1220@unknown@formal@none@1@S@But I also believed there was a certain critical complexity above which a more centralized, a priori approach was required.@@@@1@20@nil@oe@24-11-2010 1230@unknown@formal@none@1@S@I believed that the most important software (operating systems and really large tools like the Emacs programming editor) needed to be built like cathedrals, carefully crafted by individual wizards or small bands of mages working in splendid isolation, with no beta to be released before its time.@@@@1@47@nil@oe@24-11-2010 1240@unknown@formal@none@1@S@Linus Torvalds's style of development—release early and often, delegate everything you can, be open to the point of promiscuity—came as a surprise.@@@@1@22@nil@oe@24-11-2010 1250@unknown@formal@none@1@S@No quiet, reverent cathedral-building here—rather, the Linux community seemed to resemble a great babbling bazaar of differing agendas and approaches (aptly symbolized by the Linux archive sites, who'd take submissions from anyone) out of which a coherent and stable system could seemingly emerge only by a succession of miracles.@@@@1@49@nil@oe@24-11-2010 1260@unknown@formal@none@1@S@The fact that this bazaar style seemed to work, and work well, came as a distinct shock.@@@@1@17@nil@oe@24-11-2010 1270@unknown@formal@none@1@S@As I learned my way around, I worked hard not just at individual projects, but also at trying to understand why the Linux world not only didn't fly apart in confusion but seemed to go from strength to strength at a speed barely imaginable to cathedral-builders.@@@@1@46@nil@oe@24-11-2010 1280@unknown@formal@none@1@S@By mid-1996 I thought I was beginning to understand.@@@@1@9@nil@oe@24-11-2010 1290@unknown@formal@none@1@S@Chance handed me a perfect way to test my theory, in the form of an open-source project that I could consciously try to run in the bazaar style.@@@@1@28@nil@oe@24-11-2010 1300@unknown@formal@none@1@S@So I did—and it was a significant success.@@@@1@8@nil@oe@24-11-2010 1310@unknown@formal@none@1@S@This is the story of that project.@@@@1@7@nil@oe@24-11-2010 1320@unknown@formal@none@1@S@I'll use it to propose some aphorisms about effective open-source development.@@@@1@11@nil@oe@24-11-2010 1330@unknown@formal@none@1@S@Not all of these are things I first learned in the Linux world, but we'll see how the Linux world gives them particular point.@@@@1@24@nil@oe@24-11-2010 1340@unknown@formal@none@1@S@If I'm correct, they'll help you understand exactly what it is that makes the Linux community such a fountain of good software—and, perhaps, they will help you become more productive yourself.@@@@1@31@nil@oe@24-11-2010 1350@unknown@formal@none@1@S@The Mail Must Get Through@@@@1@5@nil@oe@24-11-2010 1360@unknown@formal@none@1@S@Since 1993 I'd been running the technical side of a small free-access Internet service provider called Chester County InterLink (CCIL) in West Chester, Pennsylvania.@@@@1@24@nil@oe@24-11-2010 1370@unknown@formal@none@1@S@I co-founded CCIL and wrote our unique multiuser bulletin-board software—you can check it out by telnetting to locke.ccil.org.@@@@1@18@nil@oe@24-11-2010 1380@unknown@formal@none@1@S@Today it supports almost three thousand users on thirty lines.@@@@1@10@nil@oe@24-11-2010 1390@unknown@formal@none@1@S@The job allowed me 24-hour-a-day access to the net through CCIL's 56K line—in fact, the job practically demanded it!@@@@1@19@nil@oe@24-11-2010 1400@unknown@formal@none@1@S@I had gotten quite used to instant Internet email.@@@@1@9@nil@oe@24-11-2010 1410@unknown@formal@none@1@S@I found having to periodically telnet over to locke to check my mail annoying.@@@@1@14@nil@oe@24-11-2010 1420@unknown@formal@none@1@S@What I wanted was for my mail to be delivered on snark (my home system) so that I would be notified when it arrived and could handle it using all my local tools.@@@@1@33@nil@oe@24-11-2010 1430@unknown@formal@none@1@S@The Internet's native mail forwarding protocol, SMTP (Simple Mail Transfer Protocol), wouldn't suit, because it works best when machines are connected full-time, while my personal machine isn't always on the Internet, and doesn't have a static IP address.@@@@1@38@nil@oe@24-11-2010 1440@unknown@formal@none@1@S@What I needed was a program that would reach out over my intermittent dialup connection and pull across my mail to be delivered locally.@@@@1@24@nil@oe@24-11-2010 1450@unknown@formal@none@1@S@I knew such things existed, and that most of them used a simple application protocol called POP (Post Office Protocol).@@@@1@20@nil@oe@24-11-2010 1460@unknown@formal@none@1@S@POP is now widely supported by most common mail clients, but at the time, it wasn't built in to the mail reader I was using.@@@@1@25@nil@oe@24-11-2010 1470@unknown@formal@none@1@S@I needed a POP3 client.@@@@1@5@nil@oe@24-11-2010 1480@unknown@formal@none@1@S@So I went out on the Internet and found one.@@@@1@10@nil@oe@24-11-2010 1490@unknown@formal@none@1@S@Actually, I found three or four.@@@@1@6@nil@oe@24-11-2010 1500@unknown@formal@none@1@S@I used one of them for a while, but it was missing what seemed an obvious feature, the ability to hack the addresses on fetched mail so replies would work properly.@@@@1@31@nil@oe@24-11-2010 1510@unknown@formal@none@1@S@The problem was this: suppose someone named `joe' on locke sent me mail.@@@@1@13@nil@oe@24-11-2010 1520@unknown@formal@none@1@S@If I fetched the mail to snark and then tried to reply to it, my mailer would cheerfully try to ship it to a nonexistent `joe' on snark.@@@@1@28@nil@oe@24-11-2010 1530@unknown@formal@none@1@S@Hand-editing reply addresses to tack on <\sccil.org> quickly got to be a serious pain.@@@@1@14@nil@oe@24-11-2010 1540@unknown@formal@none@1@S@This was clearly something the computer ought to be doing for me.@@@@1@12@nil@oe@24-11-2010 1550@unknown@formal@none@1@S@But none of the existing POP clients knew how!@@@@1@9@nil@oe@24-11-2010 1560@unknown@formal@none@1@S@And this brings us to the first lesson:@@@@1@8@nil@oe@24-11-2010 1570@unknown@formal@none@1@S@1. Every good work of software starts by scratching a developer's personal itch.@@@@1@13@nil@oe@24-11-2010 1580@unknown@formal@none@1@S@Perhaps this should have been obvious (it's long been proverbial that ``Necessity is the mother of invention'') but too often software developers spend their days grinding away for pay at programs they neither need nor love.@@@@1@36@nil@oe@24-11-2010 1590@unknown@formal@none@1@S@But not in the Linux world—which may explain why the average quality of software originated in the Linux community is so high.@@@@1@22@nil@oe@24-11-2010 1600@unknown@formal@none@1@S@So, did I immediately launch into a furious whirl of coding up a brand-new POP3 client to compete with the existing ones?@@@@1@22@nil@oe@24-11-2010 1610@unknown@formal@none@1@S@Not on your life!@@@@1@4@nil@oe@24-11-2010 1620@unknown@formal@none@1@S@I looked carefully at the POP utilities I had in hand, asking myself ``Which one is closest to what I want?''@@@@1@21@nil@oe@24-11-2010 1630@unknown@formal@none@1@S@Because:@@@@1@1@nil@oe@24-11-2010 1640@unknown@formal@none@1@S@2. Good programmers know what to write.@@@@1@7@nil@oe@24-11-2010 1650@unknown@formal@none@1@S@Great ones know what to rewrite (and reuse).@@@@1@8@nil@oe@24-11-2010 1660@unknown@formal@none@1@S@While I don't claim to be a great programmer, I try to imitate one.@@@@1@14@nil@oe@24-11-2010 1670@unknown@formal@none@1@S@An important trait of the great ones is constructive laziness.@@@@1@10@nil@oe@24-11-2010 1680@unknown@formal@none@1@S@They know that you get an A not for effort but for results, and that it's almost always easier to start from a good partial solution than from nothing at all.@@@@1@31@nil@oe@24-11-2010 1690@unknown@formal@none@1@S@Linus Torvalds, for example, didn't actually try to write Linux from scratch.@@@@1@12@nil@oe@24-11-2010 1700@unknown@formal@none@1@S@Instead, he started by reusing code and ideas from Minix, a tiny Unix-like operating system for PC clones.@@@@1@18@nil@oe@24-11-2010 1710@unknown@formal@none@1@S@Eventually all the Minix code went away or was completely rewritten—but while it was there, it provided scaffolding for the infant that would eventually become Linux.@@@@1@26@nil@oe@24-11-2010 1720@unknown@formal@none@1@S@In the same spirit, I went looking for an existing POP utility that was reasonably well coded, to use as a development base.@@@@1@23@nil@oe@24-11-2010 1730@unknown@formal@none@1@S@The source-sharing tradition of the Unix world has always been friendly to code reuse (this is why the GNU project chose Unix as a base OS, in spite of serious reservations about the OS itself).@@@@1@35@nil@oe@24-11-2010 1740@unknown@formal@none@1@S@The Linux world has taken this tradition nearly to its technological limit; it has terabytes of open sources generally available.@@@@1@20@nil@oe@24-11-2010 1750@unknown@formal@none@1@S@So spending time looking for some else's almost-good-enough is more likely to give you good results in the Linux world than anywhere else.@@@@1@23@nil@oe@24-11-2010 1760@unknown@formal@none@1@S@And it did for me.@@@@1@5@nil@oe@24-11-2010 1770@unknown@formal@none@1@S@With those I'd found earlier, my second search made up a total of nine candidates—fetchpop, PopTart, get-mail, gwpop, pimp, pop-perl, popc, popmail and upop.@@@@1@24@nil@oe@24-11-2010 1780@unknown@formal@none@1@S@The one I first settled on was `fetchpop' by Seung-Hong Oh.@@@@1@11@nil@oe@24-11-2010 1790@unknown@formal@none@1@S@I put my header-rewrite feature in it, and made various other improvements which the author accepted into his 1.9 release.@@@@1@20@nil@oe@24-11-2010 1800@unknown@formal@none@1@S@A few weeks later, though, I stumbled across the code for popclient by Carl Harris, and found I had a problem.@@@@1@21@nil@oe@24-11-2010 1810@unknown@formal@none@1@S@Though fetchpop had some good original ideas in it (such as its background-daemon mode), it could only handle POP3 and was rather amateurishly coded (Seung-Hong was at that time a bright but inexperienced programmer, and both traits showed).@@@@1@38@nil@oe@24-11-2010 1820@unknown@formal@none@1@S@Carl's code was better, quite professional and solid, but his program lacked several important and rather tricky-to-implement fetchpop features (including those I'd coded myself).@@@@1@24@nil@oe@24-11-2010 1830@unknown@formal@none@1@S@Stay or switch?@@@@1@3@nil@oe@24-11-2010 1840@unknown@formal@none@1@S@If I switched, I'd be throwing away the coding I'd already done in exchange for a better development base.@@@@1@19@nil@oe@24-11-2010 1850@unknown@formal@none@1@S@A practical motive to switch was the presence of multiple-protocol support.@@@@1@11@nil@oe@24-11-2010 1860@unknown@formal@none@1@S@POP3 is the most commonly used of the post-office server protocols, but not the only one.@@@@1@16@nil@oe@24-11-2010 1870@unknown@formal@none@1@S@Fetchpop and the other competition didn't do POP2, RPOP, or APOP, and I was already having vague thoughts of perhaps adding IMAP (Internet Message Access Protocol, the most recently designed and most powerful post-office protocol) just for fun.@@@@1@38@nil@oe@24-11-2010 1880@unknown@formal@none@1@S@But I had a more theoretical reason to think switching might be as good an idea as well, something I learned long before Linux.@@@@1@24@nil@oe@24-11-2010 1890@unknown@formal@none@1@S@3. ``Plan to throw one away; you will, anyhow.''@@@@1@9@nil@oe@24-11-2010 1900@unknown@formal@none@1@S@(Fred Brooks, The Mythical Man-Month, Chapter 11)@@@@1@7@nil@oe@24-11-2010 1910@unknown@formal@none@1@S@Or, to put it another way, you often don't really understand the problem until after the first time you implement a solution.@@@@1@22@nil@oe@24-11-2010 1920@unknown@formal@none@1@S@The second time, maybe you know enough to do it right.@@@@1@11@nil@oe@24-11-2010 1930@unknown@formal@none@1@S@So if you want to get it right, be ready to start over at least once.@@@@1@16@nil@oe@24-11-2010 1940@unknown@formal@none@1@S@Well (I told myself) the changes to fetchpop had been my first try.@@@@1@13@nil@oe@24-11-2010 1950@unknown@formal@none@1@S@So I switched.@@@@1@3@nil@oe@24-11-2010 1960@unknown@formal@none@1@S@After I sent my first set of popclient patches to Carl Harris on 25 June 1996, I found out that he had basically lost interest in popclient some time before.@@@@1@30@nil@oe@24-11-2010 1970@unknown@formal@none@1@S@The code was a bit dusty, with minor bugs hanging out.@@@@1@11@nil@oe@24-11-2010 1980@unknown@formal@none@1@S@I had many changes to make, and we quickly agreed that the logical thing for me to do was take over the program.@@@@1@23@nil@oe@24-11-2010 1990@unknown@formal@none@1@S@Without my actually noticing, the project had escalated.@@@@1@8@nil@oe@24-11-2010 2000@unknown@formal@none@1@S@No longer was I just contemplating minor patches to an existing POP client.@@@@1@13@nil@oe@24-11-2010 2010@unknown@formal@none@1@S@I took on maintaining an entire one, and there were ideas bubbling in my head that I knew would probably lead to major changes.@@@@1@24@nil@oe@24-11-2010 2020@unknown@formal@none@1@S@In a software culture that encourages code-sharing, this is a natural way for a project to evolve.@@@@1@17@nil@oe@24-11-2010 2030@unknown@formal@none@1@S@I was acting out this principle:@@@@1@6@nil@oe@24-11-2010 2040@unknown@formal@none@1@S@4. If you have the right attitude, interesting problems will find you.@@@@1@12@nil@oe@24-11-2010 2050@unknown@formal@none@1@S@But Carl Harris's attitude was even more important.@@@@1@8@nil@oe@24-11-2010 2060@unknown@formal@none@1@S@He understood that@@@@1@3@nil@oe@24-11-2010 2070@unknown@formal@none@1@S@5. When you lose interest in a program, your last duty to it is to hand it off to a competent successor.@@@@1@22@nil@oe@24-11-2010 2080@unknown@formal@none@1@S@Without ever having to discuss it, Carl and I knew we had a common goal of having the best solution out there.@@@@1@22@nil@oe@24-11-2010 2090@unknown@formal@none@1@S@The only question for either of us was whether I could establish that I was a safe pair of hands.@@@@1@20@nil@oe@24-11-2010 2100@unknown@formal@none@1@S@Once I did that, he acted with grace and dispatch.@@@@1@10@nil@oe@24-11-2010 2110@unknown@formal@none@1@S@I hope I will do as well when it comes my turn.@@@@1@12@nil@oe@24-11-2010 2120@unknown@formal@none@1@S@The Importance of Having Users@@@@1@5@nil@oe@24-11-2010 2130@unknown@formal@none@1@S@And so I inherited popclient.@@@@1@5@nil@oe@24-11-2010 2140@unknown@formal@none@1@S@Just as importantly, I inherited popclient's user base.@@@@1@8@nil@oe@24-11-2010 2150@unknown@formal@none@1@S@Users are wonderful things to have, and not just because they demonstrate that you're serving a need, that you've done something right.@@@@1@22@nil@oe@24-11-2010 2160@unknown@formal@none@1@S@Properly cultivated, they can become co-developers.@@@@1@6@nil@oe@24-11-2010 2170@unknown@formal@none@1@S@Another strength of the Unix tradition, one that Linux pushes to a happy extreme, is that a lot of users are hackers too.@@@@1@23@nil@oe@24-11-2010 2180@unknown@formal@none@1@S@Because source code is available, they can be effective hackers.@@@@1@10@nil@oe@24-11-2010 2190@unknown@formal@none@1@S@This can be tremendously useful for shortening debugging time.@@@@1@9@nil@oe@24-11-2010 2200@unknown@formal@none@1@S@Given a bit of encouragement, your users will diagnose problems, suggest fixes, and help improve the code far more quickly than you could unaided.@@@@1@24@nil@oe@24-11-2010 2210@unknown@formal@none@1@S@6. Treating your users as co-developers is your least-hassle route to rapid code improvement and effective debugging.@@@@1@17@nil@oe@24-11-2010 2220@unknown@formal@none@1@S@The power of this effect is easy to underestimate.@@@@1@9@nil@oe@24-11-2010 2230@unknown@formal@none@1@S@In fact, pretty well all of us in the open-source world drastically underestimated how well it would scale up with number of users and against system complexity, until Linus Torvalds showed us differently.@@@@1@33@nil@oe@24-11-2010 2240@unknown@formal@none@1@S@In fact, I think Linus's cleverest and most consequential hack was not the construction of the Linux kernel itself, but rather his invention of the Linux development model.@@@@1@28@nil@oe@24-11-2010 2250@unknown@formal@none@1@S@When I expressed this opinion in his presence once, he smiled and quietly repeated something he has often said: ``I'm basically a very lazy person who likes to get credit for things other people actually do.''@@@@1@36@nil@oe@24-11-2010 2260@unknown@formal@none@1@S@Lazy like a fox.@@@@1@4@nil@oe@24-11-2010 2270@unknown@formal@none@1@S@Or, as Robert Heinlein famously wrote of one of his characters, too lazy to fail.@@@@1@15@nil@oe@24-11-2010 2280@unknown@formal@none@1@S@In retrospect, one precedent for the methods and success of Linux can be seen in the development of the GNU Emacs Lisp library and Lisp code archives.@@@@1@27@nil@oe@24-11-2010 2290@unknown@formal@none@1@S@In contrast to the cathedral-building style of the Emacs C core and most other GNU tools, the evolution of the Lisp code pool was fluid and very user-driven.@@@@1@28@nil@oe@24-11-2010 2300@unknown@formal@none@1@S@Ideas and prototype modes were often rewritten three or four times before reaching a stable final form.@@@@1@17@nil@oe@24-11-2010 2310@unknown@formal@none@1@S@And loosely-coupled collaborations enabled by the Internet, a la Linux, were frequent.@@@@1@12@nil@oe@24-11-2010 2320@unknown@formal@none@1@S@Indeed, my own most successful single hack previous to fetchmail was probably Emacs VC (version control) mode, a Linux-like collaboration by email with three other people, only one of whom (Richard Stallman, the author of Emacs and founder of the Free Software Foundation) I have met to this day.@@@@1@49@nil@oe@24-11-2010 2330@unknown@formal@none@1@S@It was a front-end for SCCS, RCS and later CVS from within Emacs that offered ``one-touch'' version control operations.@@@@1@19@nil@oe@24-11-2010 2340@unknown@formal@none@1@S@It evolved from a tiny, crude sccs.el mode somebody else had written.@@@@1@12@nil@oe@24-11-2010 2350@unknown@formal@none@1@S@And the development of VC succeeded because, unlike Emacs itself, Emacs Lisp code could go through release/test/improve generations very quickly.@@@@1@20@nil@oe@24-11-2010 2360@unknown@formal@none@1@S@The Emacs story is not unique.@@@@1@6@nil@oe@24-11-2010 2370@unknown@formal@none@1@S@There have been other software products with a two-level architecture and a two-tier user community that combined a cathedral-mode core and a bazaar-mode toolbox.@@@@1@24@nil@oe@24-11-2010 2380@unknown@formal@none@1@S@One such is MATLAB, a commercial data-analysis and visualization tool.@@@@1@10@nil@oe@24-11-2010 2390@unknown@formal@none@1@S@Users of MATLAB and other products with a similar structure invariably report that the action, the ferment, the innovation mostly takes place in the open part of the tool where a large and varied community can tinker with it.@@@@1@39@nil@oe@24-11-2010 2400@unknown@formal@none@1@S@Release Early, Release Often@@@@1@4@nil@oe@24-11-2010 2410@unknown@formal@none@1@S@Early and frequent releases are a critical part of the Linux development model.@@@@1@13@nil@oe@24-11-2010 2420@unknown@formal@none@1@S@Most developers (including me) used to believe this was bad policy for larger than trivial projects, because early versions are almost by definition buggy versions and you don't want to wear out the patience of your users.@@@@1@37@nil@oe@24-11-2010 2430@unknown@formal@none@1@S@This belief reinforced the general commitment to a cathedral-building style of development.@@@@1@12@nil@oe@24-11-2010 2440@unknown@formal@none@1@S@If the overriding objective was for users to see as few bugs as possible, why then you'd only release a version every six months (or less often), and work like a dog on debugging between releases.@@@@1@36@nil@oe@24-11-2010 2450@unknown@formal@none@1@S@The Emacs C core was developed this way.@@@@1@8@nil@oe@24-11-2010 2460@unknown@formal@none@1@S@The Lisp library, in effect, was not—because there were active Lisp archives outside the FSF's control, where you could go to find new and development code versions independently of Emacs's release cycle.@@@@1@32@nil@oe@24-11-2010 2470@unknown@formal@none@1@S@The most important of these, the Ohio State Emacs Lisp archive, anticipated the spirit and many of the features of today's big Linux archives.@@@@1@24@nil@oe@24-11-2010 2480@unknown@formal@none@1@S@But few of us really thought very hard about what we were doing, or about what the very existence of that archive suggested about problems in the FSF's cathedral-building development model.@@@@1@31@nil@oe@24-11-2010 2490@unknown@formal@none@1@S@I made one serious attempt around 1992 to get a lot of the Ohio code formally merged into the official Emacs Lisp library.@@@@1@23@nil@oe@24-11-2010 2500@unknown@formal@none@1@S@I ran into political trouble and was largely unsuccessful.@@@@1@9@nil@oe@24-11-2010 2510@unknown@formal@none@1@S@But by a year later, as Linux became widely visible, it was clear that something different and much healthier was going on there.@@@@1@23@nil@oe@24-11-2010 2520@unknown@formal@none@1@S@Linus's open development policy was the very opposite of cathedral-building.@@@@1@10@nil@oe@24-11-2010 2530@unknown@formal@none@1@S@Linux's Internet archives were burgeoning, multiple distributions were being floated.@@@@1@10@nil@oe@24-11-2010 2540@unknown@formal@none@1@S@And all of this was driven by an unheard-of frequency of core system releases.@@@@1@14@nil@oe@24-11-2010 2550@unknown@formal@none@1@S@Linus was treating his users as co-developers in the most effective possible way:@@@@1@13@nil@oe@24-11-2010 2560@unknown@formal@none@1@S@7. Release early.@@@@1@3@nil@oe@24-11-2010 2570@unknown@formal@none@1@S@Release often.@@@@1@2@nil@oe@24-11-2010 2580@unknown@formal@none@1@S@And listen to your customers.@@@@1@5@nil@oe@24-11-2010 2590@unknown@formal@none@1@S@Linus's innovation wasn't so much in doing quick-turnaround releases incorporating lots of user feedback (something like this had been Unix-world tradition for a long time), but in scaling it up to a level of intensity that matched the complexity of what he was developing.@@@@1@44@nil@oe@24-11-2010 2600@unknown@formal@none@1@S@In those early times (around 1991) it wasn't unknown for him to release a new kernel more than once a day!@@@@1@21@nil@oe@24-11-2010 2610@unknown@formal@none@1@S@Because he cultivated his base of co-developers and leveraged the Internet for collaboration harder than anyone else, this worked.@@@@1@19@nil@oe@24-11-2010 2620@unknown@formal@none@1@S@But how did it work?@@@@1@5@nil@oe@24-11-2010 2630@unknown@formal@none@1@S@And was it something I could duplicate, or did it rely on some unique genius of Linus Torvalds?@@@@1@18@nil@oe@24-11-2010 2640@unknown@formal@none@1@S@I didn't think so.@@@@1@4@nil@oe@24-11-2010 2650@unknown@formal@none@1@S@Granted, Linus is a damn fine hacker.@@@@1@7@nil@oe@24-11-2010 2660@unknown@formal@none@1@S@How many of us could engineer an entire production-quality operating system kernel from scratch?@@@@1@14@nil@oe@24-11-2010 2670@unknown@formal@none@1@S@But Linux didn't represent any awesome conceptual leap forward.@@@@1@9@nil@oe@24-11-2010 2680@unknown@formal@none@1@S@Linus is not (or at least, not yet) an innovative genius of design in the way that, say, Richard Stallman or James Gosling (of NeWS and Java) are.@@@@1@28@nil@oe@24-11-2010 2690@unknown@formal@none@1@S@Rather, Linus seems to me to be a genius of engineering and implementation, with a sixth sense for avoiding bugs and development dead-ends and a true knack for finding the minimum-effort path from point A to point B.@@@@1@38@nil@oe@24-11-2010 2700@unknown@formal@none@1@S@Indeed, the whole design of Linux breathes this quality and mirrors Linus's essentially conservative and simplifying design approach.@@@@1@18@nil@oe@24-11-2010 2710@unknown@formal@none@1@S@So, if rapid releases and leveraging the Internet medium to the hilt were not accidents but integral parts of Linus's engineering-genius insight into the minimum-effort path, what was he maximizing?@@@@1@30@nil@oe@24-11-2010 2720@unknown@formal@none@1@S@What was he cranking out of the machinery?@@@@1@8@nil@oe@24-11-2010 2730@unknown@formal@none@1@S@Put that way, the question answers itself.@@@@1@7@nil@oe@24-11-2010 2740@unknown@formal@none@1@S@Linus was keeping his hacker/users constantly stimulated and rewarded—stimulated by the prospect of having an ego-satisfying piece of the action, rewarded by the sight of constant (even daily) improvement in their work.@@@@1@32@nil@oe@24-11-2010 2750@unknown@formal@none@1@S@Linus was directly aiming to maximize the number of person-hours thrown at debugging and development, even at the possible cost of instability in the code and user-base burnout if any serious bug proved intractable.@@@@1@34@nil@oe@24-11-2010 2760@unknown@formal@none@1@S@Linus was behaving as though he believed something like this:@@@@1@10@nil@oe@24-11-2010 2770@unknown@formal@none@1@S@8. Given a large enough beta-tester and co-developer base, almost every problem will be characterized quickly and the fix obvious to someone.@@@@1@22@nil@oe@24-11-2010 2780@unknown@formal@none@1@S@Or, less formally, ``Given enough eyeballs, all bugs are shallow.''@@@@1@10@nil@oe@24-11-2010 2790@unknown@formal@none@1@S@I dub this: ``Linus's Law''.@@@@1@5@nil@oe@24-11-2010 2800@unknown@formal@none@1@S@My original formulation was that every problem ``will be transparent to somebody''.@@@@1@12@nil@oe@24-11-2010 2810@unknown@formal@none@1@S@Linus demurred that the person who understands and fixes the problem is not necessarily or even usually the person who first characterizes it.@@@@1@23@nil@oe@24-11-2010 2820@unknown@formal@none@1@S@``Somebody finds the problem,'' he says, ``and somebody else understands it.@@@@1@11@nil@oe@24-11-2010 2830@unknown@formal@none@1@S@And I'll go on record as saying that finding it is the bigger challenge.''@@@@1@14@nil@oe@24-11-2010 2840@unknown@formal@none@1@S@That correction is important; we'll see how in the next section, when we examine the practice of debugging in more detail.@@@@1@21@nil@oe@24-11-2010 2850@unknown@formal@none@1@S@But the key point is that both parts of the process (finding and fixing) tend to happen rapidly.@@@@1@18@nil@oe@24-11-2010 2860@unknown@formal@none@1@S@In Linus's Law, I think, lies the core difference underlying the cathedral-builder and bazaar styles.@@@@1@15@nil@oe@24-11-2010 2870@unknown@formal@none@1@S@In the cathedral-builder view of programming, bugs and development problems are tricky, insidious, deep phenomena.@@@@1@15@nil@oe@24-11-2010 2880@unknown@formal@none@1@S@It takes months of scrutiny by a dedicated few to develop confidence that you've winkled them all out.@@@@1@18@nil@oe@24-11-2010 2890@unknown@formal@none@1@S@Thus the long release intervals, and the inevitable disappointment when long-awaited releases are not perfect.@@@@1@15@nil@oe@24-11-2010 2900@unknown@formal@none@1@S@In the bazaar view, on the other hand, you assume that bugs are generally shallow phenomena—or, at least, that they turn shallow pretty quickly when exposed to a thousand eager co-developers pounding on every single new release.@@@@1@37@nil@oe@24-11-2010 2910@unknown@formal@none@1@S@Accordingly you release often in order to get more corrections, and as a beneficial side effect you have less to lose if an occasional botch gets out the door.@@@@1@29@nil@oe@24-11-2010 2920@unknown@formal@none@1@S@And that's it.@@@@1@3@nil@oe@24-11-2010 2930@unknown@formal@none@1@S@That's enough.@@@@1@2@nil@oe@24-11-2010 2940@unknown@formal@none@1@S@If ``Linus's Law'' is false, then any system as complex as the Linux kernel, being hacked over by as many hands as that kernel was, should at some point have collapsed under the weight of unforeseen bad interactions and undiscovered ``deep'' bugs.@@@@1@42@nil@oe@24-11-2010 2950@unknown@formal@none@1@S@If it's true, on the other hand, it is sufficient to explain Linux's relative lack of bugginess and its continuous uptimes spanning months or even years.@@@@1@26@nil@oe@24-11-2010 2960@unknown@formal@none@1@S@Maybe it shouldn't have been such a surprise, at that.@@@@1@10@nil@oe@24-11-2010 2970@unknown@formal@none@1@S@Sociologists years ago discovered that the averaged opinion of a mass of equally expert (or equally ignorant) observers is quite a bit more reliable a predictor than the opinion of a single randomly-chosen one of the observers.@@@@1@37@nil@oe@24-11-2010 2980@unknown@formal@none@1@S@They called this the Delphi effect.@@@@1@6@nil@oe@24-11-2010 2990@unknown@formal@none@1@S@It appears that what Linus has shown is that this applies even to debugging an operating system—that the Delphi effect can tame development complexity even at the complexity level of an OS kernel.@@@@1@33@nil@oe@24-11-2010 3000@unknown@formal@none@1@S@One special feature of the Linux situation that clearly helps along the Delphi effect is the fact that the contributors for any given project are self-selected.@@@@1@26@nil@oe@24-11-2010 3010@unknown@formal@none@1@S@An early respondent pointed out that contributions are received not from a random sample, but from people who are interested enough to use the software, learn about how it works, attempt to find solutions to problems they encounter, and actually produce an apparently reasonable fix.@@@@1@45@nil@oe@24-11-2010 3020@unknown@formal@none@1@S@Anyone who passes all these filters is highly likely to have something useful to contribute.@@@@1@15@nil@oe@24-11-2010 3030@unknown@formal@none@1@S@Linus's Law can be rephrased as ``Debugging is parallelizable''.@@@@1@9@nil@oe@24-11-2010 3040@unknown@formal@none@1@S@Although debugging requires debuggers to communicate with some coordinating developer, it doesn't require significant coordination between debuggers.@@@@1@17@nil@oe@24-11-2010 3050@unknown@formal@none@1@S@Thus it doesn't fall prey to the same quadratic complexity and management costs that make adding developers problematic.@@@@1@18@nil@oe@24-11-2010 3060@unknown@formal@none@1@S@In practice, the theoretical loss of efficiency due to duplication of work by debuggers almost never seems to be an issue in the Linux world.@@@@1@25@nil@oe@24-11-2010 3070@unknown@formal@none@1@S@One effect of a ``release early and often'' policy is to minimize such duplication by propagating fed-back fixes quickly.@@@@1@19@nil@oe@24-11-2010 3080@unknown@formal@none@1@S@Brooks (the author of The Mythical Man-Month) even made an off-hand observation related to this: ``The total cost of maintaining a widely used program is typically 40 percent or more of the cost of developing it.@@@@1@36@nil@oe@24-11-2010 3090@unknown@formal@none@1@S@Surprisingly this cost is strongly affected by the number of users.@@@@1@11@nil@oe@24-11-2010 3100@unknown@formal@none@1@S@More users find more bugs.''@@@@1@5@nil@oe@24-11-2010 3110@unknown@formal@none@1@S@[emphasis added].@@@@1@2@nil@oe@24-11-2010 3120@unknown@formal@none@1@S@More users find more bugs because adding more users adds more different ways of stressing the program.@@@@1@17@nil@oe@24-11-2010 3130@unknown@formal@none@1@S@This effect is amplified when the users are co-developers.@@@@1@9@nil@oe@24-11-2010 3140@unknown@formal@none@1@S@Each one approaches the task of bug characterization with a slightly different perceptual set and analytical toolkit, a different angle on the problem.@@@@1@23@nil@oe@24-11-2010 3150@unknown@formal@none@1@S@The ``Delphi effect'' seems to work precisely because of this variation.@@@@1@11@nil@oe@24-11-2010 3160@unknown@formal@none@1@S@In the specific context of debugging, the variation also tends to reduce duplication of effort.@@@@1@15@nil@oe@24-11-2010 3170@unknown@formal@none@1@S@So adding more beta-testers may not reduce the complexity of the current ``deepest'' bug from the developer's point of view, but it increases the probability that someone's toolkit will be matched to the problem in such a way that the bug is shallow to that person.@@@@1@46@nil@oe@24-11-2010 3180@unknown@formal@none@1@S@Linus coppers his bets, too.@@@@1@5@nil@oe@24-11-2010 3190@unknown@formal@none@1@S@In case there are serious bugs, Linux kernel version are numbered in such a way that potential users can make a choice either to run the last version designated ``stable'' or to ride the cutting edge and risk bugs in order to get new features.@@@@1@45@nil@oe@24-11-2010 3200@unknown@formal@none@1@S@This tactic is not yet systematically imitated by most Linux hackers, but perhaps it should be; the fact that either choice is available makes both more attractive.@@@@1@27@nil@oe@24-11-2010 3210@unknown@formal@none@1@S@How Many Eyeballs Tame Complexity@@@@1@5@nil@oe@24-11-2010 3220@unknown@formal@none@1@S@It's one thing to observe in the large that the bazaar style greatly accelerates debugging and code evolution.@@@@1@18@nil@oe@24-11-2010 3230@unknown@formal@none@1@S@It's another to understand exactly how and why it does so at the micro-level of day-to-day developer and tester behavior.@@@@1@20@nil@oe@24-11-2010 3240@unknown@formal@none@1@S@In this section (written three years after the original paper, using insights by developers who read it and re-examined their own behavior) we'll take a hard look at the actual mechanisms.@@@@1@31@nil@oe@24-11-2010 3250@unknown@formal@none@1@S@Non-technically inclined readers can safely skip to the next section.@@@@1@10@nil@oe@24-11-2010 3260@unknown@formal@none@1@S@One key to understanding is to realize exactly why it is that the kind of bug report non–source-aware users normally turn in tends not to be very useful.@@@@1@28@nil@oe@24-11-2010 3270@unknown@formal@none@1@S@Non–source-aware users tend to report only surface symptoms; they take their environment for granted, so they (a) omit critical background data, and (b) seldom include a reliable recipe for reproducing the bug.@@@@1@32@nil@oe@24-11-2010 3280@unknown@formal@none@1@S@The underlying problem here is a mismatch between the tester's and the developer's mental models of the program; the tester, on the outside looking in, and the developer on the inside looking out.@@@@1@33@nil@oe@24-11-2010 3290@unknown@formal@none@1@S@In closed-source development they're both stuck in these roles, and tend to talk past each other and find each other deeply frustrating.@@@@1@22@nil@oe@24-11-2010 3300@unknown@formal@none@1@S@Open-source development breaks this bind, making it far easier for tester and developer to develop a shared representation grounded in the actual source code and to communicate effectively about it.@@@@1@30@nil@oe@24-11-2010 3310@unknown@formal@none@1@S@Practically, there is a huge difference in leverage for the developer between the kind of bug report that just reports externally-visible symptoms and the kind that hooks directly to the developer's source-code–based mental representation of the program.@@@@1@37@nil@oe@24-11-2010 3320@unknown@formal@none@1@S@Most bugs, most of the time, are easily nailed given even an incomplete but suggestive characterization of their error conditions at source-code level.@@@@1@23@nil@oe@24-11-2010 3330@unknown@formal@none@1@S@When someone among your beta-testers can point out, "there's a boundary problem in line nnn", or even just "under conditions X, Y, and Z, this variable rolls over", a quick look at the offending code often suffices to pin down the exact mode of failure and generate a fix.@@@@1@49@nil@oe@24-11-2010 3340@unknown@formal@none@1@S@Thus, source-code awareness by both parties greatly enhances both good communication and the synergy between what a beta-tester reports and what the core developer(s) know.@@@@1@25@nil@oe@24-11-2010 3350@unknown@formal@none@1@S@In turn, this means that the core developers' time tends to be well conserved, even with many collaborators.@@@@1@18@nil@oe@24-11-2010 3360@unknown@formal@none@1@S@Another characteristic of the open-source method that conserves developer time is the communication structure of typical open-source projects.@@@@1@18@nil@oe@24-11-2010 3370@unknown@formal@none@1@S@Above I used the term "core developer"; this reflects a distinction between the project core (typically quite small; a single core developer is common, and one to three is typical) and the project halo of beta-testers and available contributors (which often numbers in the hundreds).@@@@1@45@nil@oe@24-11-2010 3380@unknown@formal@none@1@S@The fundamental problem that traditional software-development organization addresses is Brook's Law: ``Adding more programmers to a late project makes it later.''@@@@1@21@nil@oe@24-11-2010 3390@unknown@formal@none@1@S@More generally, Brooks's Law predicts that the complexity and communication costs of a project rise with the square of the number of developers, while work done only rises linearly.@@@@1@29@nil@oe@24-11-2010 3400@unknown@formal@none@1@S@Brooks's Law is founded on experience that bugs tend strongly to cluster at the interfaces between code written by different people, and that communications/coordination overhead on a project tends to rise with the number of interfaces between human beings.@@@@1@39@nil@oe@24-11-2010 3410@unknown@formal@none@1@S@Thus, problems scale with the number of communications paths between developers, which scales as the square of the number of developers (more precisely, according to the formula N*(N - 1)/2 where N is the number of developers).@@@@1@37@nil@oe@24-11-2010 3420@unknown@formal@none@1@S@The Brooks's Law analysis (and the resulting fear of large numbers in development groups) rests on a hidden assumption: that the communications structure of the project is necessarily a complete graph, that everybody talks to everybody else.@@@@1@37@nil@oe@24-11-2010 3430@unknown@formal@none@1@S@But on open-source projects, the halo developers work on what are in effect separable parallel subtasks and interact with each other very little; code changes and bug reports stream through the core group, and only within that small core group do we pay the full Brooksian overhead.@@@@1@47@nil@oe@24-11-2010 3440@unknown@formal@none@1@S@There are still more reasons that source-code–level bug reporting tends to be very efficient.@@@@1@14@nil@oe@24-11-2010 3450@unknown@formal@none@1@S@They center around the fact that a single error can often have multiple possible symptoms, manifesting differently depending on details of the user's usage pattern and environment.@@@@1@27@nil@oe@24-11-2010 3460@unknown@formal@none@1@S@Such errors tend to be exactly the sort of complex and subtle bugs (such as dynamic-memory-management errors or nondeterministic interrupt-window artifacts) that are hardest to reproduce at will or to pin down by static analysis, and which do the most to create long-term problems in software.@@@@1@46@nil@oe@24-11-2010 3470@unknown@formal@none@1@S@A tester who sends in a tentative source-code–level characterization of such a multi-symptom bug (e.g. "It looks to me like there's a window in the signal handling near line 1250" or "Where are you zeroing that buffer?") may give a developer, otherwise too close to the code to see it, the critical clue to a half-dozen disparate symptoms.@@@@1@58@nil@oe@24-11-2010 3480@unknown@formal@none@1@S@In cases like this, it may be hard or even impossible to know which externally-visible misbehaviour was caused by precisely which bug—but with frequent releases, it's unnecessary to know.@@@@1@29@nil@oe@24-11-2010 3490@unknown@formal@none@1@S@Other collaborators will be likely to find out quickly whether their bug has been fixed or not.@@@@1@17@nil@oe@24-11-2010 3500@unknown@formal@none@1@S@In many cases, source-level bug reports will cause misbehaviours to drop out without ever having been attributed to any specific fix.@@@@1@21@nil@oe@24-11-2010 3510@unknown@formal@none@1@S@Complex multi-symptom errors also tend to have multiple trace paths from surface symptoms back to the actual bug.@@@@1@18@nil@oe@24-11-2010 3520@unknown@formal@none@1@S@Which of the trace paths a given developer or tester can chase may depend on subtleties of that person's environment, and may well change in a not obviously deterministic way over time.@@@@1@32@nil@oe@24-11-2010 3530@unknown@formal@none@1@S@In effect, each developer and tester samples a semi-random set of the program's state space when looking for the etiology of a symptom.@@@@1@23@nil@oe@24-11-2010 3540@unknown@formal@none@1@S@The more subtle and complex the bug, the less likely that skill will be able to guarantee the relevance of that sample.@@@@1@22@nil@oe@24-11-2010 3550@unknown@formal@none@1@S@For simple and easily reproducible bugs, then, the accent will be on the "semi" rather than the "random"; debugging skill and intimacy with the code and its architecture will matter a lot.@@@@1@32@nil@oe@24-11-2010 3560@unknown@formal@none@1@S@But for complex bugs, the accent will be on the "random".@@@@1@11@nil@oe@24-11-2010 3570@unknown@formal@none@1@S@Under these circumstances many people running traces will be much more effective than a few people running traces sequentially—even if the few have a much higher average skill level.@@@@1@29@nil@oe@24-11-2010 3580@unknown@formal@none@1@S@This effect will be greatly amplified if the difficulty of following trace paths from different surface symptoms back to a bug varies significantly in a way that can't be predicted by looking at the symptoms.@@@@1@35@nil@oe@24-11-2010 3590@unknown@formal@none@1@S@A single developer sampling those paths sequentially will be as likely to pick a difficult trace path on the first try as an easy one.@@@@1@25@nil@oe@24-11-2010 3600@unknown@formal@none@1@S@On the other hand, suppose many people are trying trace paths in parallel while doing rapid releases.@@@@1@17@nil@oe@24-11-2010 3610@unknown@formal@none@1@S@Then it is likely one of them will find the easiest path immediately, and nail the bug in a much shorter time.@@@@1@22@nil@oe@24-11-2010 3620@unknown@formal@none@1@S@The project maintainer will see that, ship a new release, and the other people running traces on the same bug will be able to stop before having spent too much time on their more difficult traces .@@@@1@37@nil@oe@24-11-2010 3630@unknown@formal@none@1@S@When Is a Rose Not a Rose?@@@@1@7@nil@oe@24-11-2010 3640@unknown@formal@none@1@S@Having studied Linus's behavior and formed a theory about why it was successful, I made a conscious decision to test this theory on my new (admittedly much less complex and ambitious) project.@@@@1@32@nil@oe@24-11-2010 3650@unknown@formal@none@1@S@But the first thing I did was reorganize and simplify popclient a lot.@@@@1@13@nil@oe@24-11-2010 3660@unknown@formal@none@1@S@Carl Harris's implementation was very sound, but exhibited a kind of unnecessary complexity common to many C programmers.@@@@1@18@nil@oe@24-11-2010 3670@unknown@formal@none@1@S@He treated the code as central and the data structures as support for the code.@@@@1@15@nil@oe@24-11-2010 3680@unknown@formal@none@1@S@As a result, the code was beautiful but the data structure design ad-hoc and rather ugly (at least by the high standards of this veteran LISP hacker).@@@@1@27@nil@oe@24-11-2010 3690@unknown@formal@none@1@S@I had another purpose for rewriting besides improving the code and the data structure design, however.@@@@1@16@nil@oe@24-11-2010 3700@unknown@formal@none@1@S@That was to evolve it into something I understood completely.@@@@1@10@nil@oe@24-11-2010 3710@unknown@formal@none@1@S@It's no fun to be responsible for fixing bugs in a program you don't understand.@@@@1@15@nil@oe@24-11-2010 3720@unknown@formal@none@1@S@For the first month or so, then, I was simply following out the implications of Carl's basic design.@@@@1@18@nil@oe@24-11-2010 3730@unknown@formal@none@1@S@The first serious change I made was to add IMAP support.@@@@1@11@nil@oe@24-11-2010 3740@unknown@formal@none@1@S@I did this by reorganizing the protocol machines into a generic driver and three method tables (for POP2, POP3, and IMAP).@@@@1@21@nil@oe@24-11-2010 3750@unknown@formal@none@1@S@This and the previous changes illustrate a general principle that's good for programmers to keep in mind, especially in languages like C that don't naturally do dynamic typing:@@@@1@28@nil@oe@24-11-2010 3760@unknown@formal@none@1@S@9. Smart data structures and dumb code works a lot better than the other way around.@@@@1@16@nil@oe@24-11-2010 3770@unknown@formal@none@1@S@Brooks, Chapter 9: ``Show me your flowchart and conceal your tables, and I shall continue to be mystified.@@@@1@18@nil@oe@24-11-2010 3780@unknown@formal@none@1@S@Show me your tables, and I won't usually need your flowchart; it'll be obvious.''@@@@1@14@nil@oe@24-11-2010 3790@unknown@formal@none@1@S@Allowing for thirty years of terminological/cultural shift, it's the same point.@@@@1@11@nil@oe@24-11-2010 3800@unknown@formal@none@1@S@At this point (early September 1996, about six weeks from zero) I started thinking that a name change might be in order—after all, it wasn't just a POP client any more.@@@@1@31@nil@oe@24-11-2010 3810@unknown@formal@none@1@S@But I hesitated, because there was as yet nothing genuinely new in the design.@@@@1@14@nil@oe@24-11-2010 3820@unknown@formal@none@1@S@My version of popclient had yet to develop an identity of its own.@@@@1@13@nil@oe@24-11-2010 3830@unknown@formal@none@1@S@That changed, radically, when popclient learned how to forward fetched mail to the SMTP port.@@@@1@15@nil@oe@24-11-2010 3840@unknown@formal@none@1@S@I'll get to that in a moment.@@@@1@7@nil@oe@24-11-2010 3850@unknown@formal@none@1@S@But first: I said earlier that I'd decided to use this project to test my theory about what Linus Torvalds had done right.@@@@1@23@nil@oe@24-11-2010 3860@unknown@formal@none@1@S@How (you may well ask) did I do that?@@@@1@9@nil@oe@24-11-2010 3870@unknown@formal@none@1@S@In these ways:@@@@1@3@nil@oe@24-11-2010 3880@unknown@formal@none@1@S@I released early and often (almost never less often than every ten days; during periods of intense development, once a day).@@@@1@21@nil@oe@24-11-2010 3890@unknown@formal@none@1@S@I grew my beta list by adding to it everyone who contacted me about fetchmail.@@@@1@15@nil@oe@24-11-2010 3900@unknown@formal@none@1@S@I sent chatty announcements to the beta list whenever I released, encouraging people to participate.@@@@1@15@nil@oe@24-11-2010 3910@unknown@formal@none@1@S@And I listened to my beta-testers, polling them about design decisions and stroking them whenever they sent in patches and feedback.@@@@1@21@nil@oe@24-11-2010 3920@unknown@formal@none@1@S@The payoff from these simple measures was immediate.@@@@1@8@nil@oe@24-11-2010 3930@unknown@formal@none@1@S@From the beginning of the project, I got bug reports of a quality most developers would kill for, often with good fixes attached.@@@@1@23@nil@oe@24-11-2010 3940@unknown@formal@none@1@S@I got thoughtful criticism, I got fan mail, I got intelligent feature suggestions.@@@@1@13@nil@oe@24-11-2010 3950@unknown@formal@none@1@S@Which leads to:@@@@1@3@nil@oe@24-11-2010 3960@unknown@formal@none@1@S@10. If you treat your beta-testers as if they're your most valuable resource, they will respond by becoming your most valuable resource.@@@@1@22@nil@oe@24-11-2010 3970@unknown@formal@none@1@S@One interesting measure of fetchmail's success is the sheer size of the project beta list, fetchmail-friends.@@@@1@16@nil@oe@24-11-2010 3980@unknown@formal@none@1@S@At the time of latest revision of this paper (November 2000) it has 287 members and is adding two or three a week.@@@@1@23@nil@oe@24-11-2010 3990@unknown@formal@none@1@S@Actually, when I revised in late May 1997 I found the list was beginning to lose members from its high of close to 300 for an interesting reason.@@@@1@28@nil@oe@24-11-2010 4000@unknown@formal@none@1@S@Several people have asked me to unsubscribe them because fetchmail is working so well for them that they no longer need to see the list traffic!@@@@1@26@nil@oe@24-11-2010 4010@unknown@formal@none@1@S@Perhaps this is part of the normal life-cycle of a mature bazaar-style project.@@@@1@13@nil@oe@24-11-2010 4020@unknown@formal@none@1@S@Popclient becomes Fetchmail@@@@1@3@nil@oe@24-11-2010 4030@unknown@formal@none@1@S@The real turning point in the project was when Harry Hochheiser sent me his scratch code for forwarding mail to the client machine's SMTP port.@@@@1@25@nil@oe@24-11-2010 4040@unknown@formal@none@1@S@I realized almost immediately that a reliable implementation of this feature would make all the other mail delivery modes next to obsolete.@@@@1@22@nil@oe@24-11-2010 4050@unknown@formal@none@1@S@For many weeks I had been tweaking fetchmail rather incrementally while feeling like the interface design was serviceable but grubby—inelegant and with too many exiguous options hanging out all over.@@@@1@30@nil@oe@24-11-2010 4060@unknown@formal@none@1@S@The options to dump fetched mail to a mailbox file or standard output particularly bothered me, but I couldn't figure out why.@@@@1@22@nil@oe@24-11-2010 4070@unknown@formal@none@1@S@(If you don't care about the technicalia of Internet mail, the next two paragraphs can be safely skipped.)@@@@1@18@nil@oe@24-11-2010 4080@unknown@formal@none@1@S@What I saw when I thought about SMTP forwarding was that popclient had been trying to do too many things.@@@@1@20@nil@oe@24-11-2010 4090@unknown@formal@none@1@S@It had been designed to be both a mail transport agent (MTA) and a local delivery agent (MDA).@@@@1@18@nil@oe@24-11-2010 4100@unknown@formal@none@1@S@With SMTP forwarding, it could get out of the MDA business and be a pure MTA, handing off mail to other programs for local delivery just as sendmail does.@@@@1@29@nil@oe@24-11-2010 4110@unknown@formal@none@1@S@Why mess with all the complexity of configuring a mail delivery agent or setting up lock-and-append on a mailbox when port 25 is almost guaranteed to be there on any platform with TCP/IP support in the first place?@@@@1@38@nil@oe@24-11-2010 4120@unknown@formal@none@1@S@Especially when this means retrieved mail is guaranteed to look like normal sender-initiated SMTP mail, which is really what we want anyway.@@@@1@22@nil@oe@24-11-2010 4130@unknown@formal@none@1@S@(Back to a higher level....)@@@@1@5@nil@oe@24-11-2010 4140@unknown@formal@none@1@S@Even if you didn't follow the preceding technical jargon, there are several important lessons here.@@@@1@15@nil@oe@24-11-2010 4150@unknown@formal@none@1@S@First, this SMTP-forwarding concept was the biggest single payoff I got from consciously trying to emulate Linus's methods.@@@@1@18@nil@oe@24-11-2010 4160@unknown@formal@none@1@S@A user gave me this terrific idea—all I had to do was understand the implications.@@@@1@15@nil@oe@24-11-2010 4170@unknown@formal@none@1@S@11. The next best thing to having good ideas is recognizing good ideas from your users.@@@@1@16@nil@oe@24-11-2010 4180@unknown@formal@none@1@S@Sometimes the latter is better.@@@@1@5@nil@oe@24-11-2010 4190@unknown@formal@none@1@S@Interestingly enough, you will quickly find that if you are completely and self-deprecatingly truthful about how much you owe other people, the world at large will treat you as though you did every bit of the invention yourself and are just being becomingly modest about your innate genius.@@@@1@48@nil@oe@24-11-2010 4200@unknown@formal@none@1@S@We can all see how well this worked for Linus!@@@@1@10@nil@oe@24-11-2010 4210@unknown@formal@none@1@S@(When I gave my talk at the first Perl Conference in August 1997, hacker extraordinaire Larry Wall was in the front row.@@@@1@22@nil@oe@24-11-2010 4220@unknown@formal@none@1@S@As I got to the last line above he called out, religious-revival style, ``Tell it, tell it, brother!''.@@@@1@18@nil@oe@24-11-2010 4230@unknown@formal@none@1@S@The whole audience laughed, because they knew this had worked for the inventor of Perl, too.)@@@@1@16@nil@oe@24-11-2010 4240@unknown@formal@none@1@S@After a very few weeks of running the project in the same spirit, I began to get similar praise not just from my users but from other people to whom the word leaked out.@@@@1@34@nil@oe@24-11-2010 4250@unknown@formal@none@1@S@I stashed away some of that email; I'll look at it again sometime if I ever start wondering whether my life has been worthwhile :-).@@@@1@25@nil@oe@24-11-2010 4260@unknown@formal@none@1@S@But there are two more fundamental, non-political lessons here that are general to all kinds of design.@@@@1@17@nil@oe@24-11-2010 4270@unknown@formal@none@1@S@12. Often, the most striking and innovative solutions come from realizing that your concept of the problem was wrong.@@@@1@19@nil@oe@24-11-2010 4280@unknown@formal@none@1@S@I had been trying to solve the wrong problem by continuing to develop popclient as a combined MTA/MDA with all kinds of funky local delivery modes.@@@@1@26@nil@oe@24-11-2010 4290@unknown@formal@none@1@S@Fetchmail's design needed to be rethought from the ground up as a pure MTA, a part of the normal SMTP-speaking Internet mail path.@@@@1@23@nil@oe@24-11-2010 4300@unknown@formal@none@1@S@When you hit a wall in development—when you find yourself hard put to think past the next patch—it's often time to ask not whether you've got the right answer, but whether you're asking the right question.@@@@1@36@nil@oe@24-11-2010 4310@unknown@formal@none@1@S@Perhaps the problem needs to be reframed.@@@@1@7@nil@oe@24-11-2010 4320@unknown@formal@none@1@S@Well, I had reframed my problem.@@@@1@6@nil@oe@24-11-2010 4330@unknown@formal@none@1@S@Clearly, the right thing to do was (1) hack SMTP forwarding support into the generic driver, (2) make it the default mode, and (3) eventually throw out all the other delivery modes, especially the deliver-to-file and deliver-to-standard-output options.@@@@1@38@nil@oe@24-11-2010 4340@unknown@formal@none@1@S@I hesitated over step 3 for some time, fearing to upset long-time popclient users dependent on the alternate delivery mechanisms.@@@@1@20@nil@oe@24-11-2010 4350@unknown@formal@none@1@S@In theory, they could immediately switch to .forward files or their non-sendmail equivalents to get the same effects.@@@@1@18@nil@oe@24-11-2010 4360@unknown@formal@none@1@S@In practice the transition might have been messy.@@@@1@8@nil@oe@24-11-2010 4370@unknown@formal@none@1@S@But when I did it, the benefits proved huge.@@@@1@9@nil@oe@24-11-2010 4380@unknown@formal@none@1@S@The cruftiest parts of the driver code vanished.@@@@1@8@nil@oe@24-11-2010 4390@unknown@formal@none@1@S@Configuration got radically simpler—no more grovelling around for the system MDA and user's mailbox, no more worries about whether the underlying OS supports file locking.@@@@1@25@nil@oe@24-11-2010 4400@unknown@formal@none@1@S@Also, the only way to lose mail vanished.@@@@1@8@nil@oe@24-11-2010 4410@unknown@formal@none@1@S@If you specified delivery to a file and the disk got full, your mail got lost.@@@@1@16@nil@oe@24-11-2010 4420@unknown@formal@none@1@S@This can't happen with SMTP forwarding because your SMTP listener won't return OK unless the message can be delivered or at least spooled for later delivery.@@@@1@26@nil@oe@24-11-2010 4430@unknown@formal@none@1@S@Also, performance improved (though not so you'd notice it in a single run).@@@@1@13@nil@oe@24-11-2010 4440@unknown@formal@none@1@S@Another not insignificant benefit of this change was that the manual page got a lot simpler.@@@@1@16@nil@oe@24-11-2010 4450@unknown@formal@none@1@S@Later, I had to bring delivery via a user-specified local MDA back in order to allow handling of some obscure situations involving dynamic SLIP.@@@@1@24@nil@oe@24-11-2010 4460@unknown@formal@none@1@S@But I found a much simpler way to do it.@@@@1@10@nil@oe@24-11-2010 4470@unknown@formal@none@1@S@The moral?@@@@1@2@nil@oe@24-11-2010 4480@unknown@formal@none@1@S@Don't hesitate to throw away superannuated features when you can do it without loss of effectiveness.@@@@1@16@nil@oe@24-11-2010 4490@unknown@formal@none@1@S@Antoine de Saint-Exupéry (who was an aviator and aircraft designer when he wasn't authoring classic children's books) said:@@@@1@18@nil@oe@24-11-2010 4500@unknown@formal@none@1@S@13. ``Perfection (in design) is achieved not when there is nothing more to add, but rather when there is nothing more to take away.''@@@@1@24@nil@oe@24-11-2010 4510@unknown@formal@none@1@S@When your code is getting both better and simpler, that is when you know it's right.@@@@1@16@nil@oe@24-11-2010 4520@unknown@formal@none@1@S@And in the process, the fetchmail design acquired an identity of its own, different from the ancestral popclient.@@@@1@18@nil@oe@24-11-2010 4530@unknown@formal@none@1@S@It was time for the name change.@@@@1@7@nil@oe@24-11-2010 4540@unknown@formal@none@1@S@The new design looked much more like a dual of sendmail than the old popclient had; both are MTAs, but where sendmail pushes then delivers, the new popclient pulls then delivers.@@@@1@31@nil@oe@24-11-2010 4550@unknown@formal@none@1@S@So, two months off the blocks, I renamed it fetchmail.@@@@1@10@nil@oe@24-11-2010 4560@unknown@formal@none@1@S@There is a more general lesson in this story about how SMTP delivery came to fetchmail.@@@@1@16@nil@oe@24-11-2010 4570@unknown@formal@none@1@S@It is not only debugging that is parallelizable; development and (to a perhaps surprising extent) exploration of design space is, too.@@@@1@21@nil@oe@24-11-2010 4580@unknown@formal@none@1@S@When your development mode is rapidly iterative, development and enhancement may become special cases of debugging—fixing `bugs of omission' in the original capabilities or concept of the software.@@@@1@28@nil@oe@24-11-2010 4590@unknown@formal@none@1@S@Even at a higher level of design, it can be very valuable to have lots of co-developers random-walking through the design space near your product.@@@@1@25@nil@oe@24-11-2010 4600@unknown@formal@none@1@S@Consider the way a puddle of water finds a drain, or better yet how ants find food: exploration essentially by diffusion, followed by exploitation mediated by a scalable communication mechanism.@@@@1@30@nil@oe@24-11-2010 4610@unknown@formal@none@1@S@This works very well; as with Harry Hochheiser and me, one of your outriders may well find a huge win nearby that you were just a little too close-focused to see.@@@@1@31@nil@oe@24-11-2010 4620@unknown@formal@none@1@S@Fetchmail Grows Up@@@@1@3@nil@oe@24-11-2010 4630@unknown@formal@none@1@S@There I was with a neat and innovative design, code that I knew worked well because I used it every day, and a burgeoning beta list.@@@@1@26@nil@oe@24-11-2010 4640@unknown@formal@none@1@S@It gradually dawned on me that I was no longer engaged in a trivial personal hack that might happen to be useful to few other people.@@@@1@26@nil@oe@24-11-2010 4650@unknown@formal@none@1@S@I had my hands on a program that every hacker with a Unix box and a SLIP/PPP mail connection really needs.@@@@1@21@nil@oe@24-11-2010 4660@unknown@formal@none@1@S@With the SMTP forwarding feature, it pulled far enough in front of the competition to potentially become a ``category killer'', one of those classic programs that fills its niche so competently that the alternatives are not just discarded but almost forgotten.@@@@1@41@nil@oe@24-11-2010 4670@unknown@formal@none@1@S@I think you can't really aim or plan for a result like this.@@@@1@13@nil@oe@24-11-2010 4680@unknown@formal@none@1@S@You have to get pulled into it by design ideas so powerful that afterward the results just seem inevitable, natural, even foreordained.@@@@1@22@nil@oe@24-11-2010 4690@unknown@formal@none@1@S@The only way to try for ideas like that is by having lots of ideas—or by having the engineering judgment to take other peoples' good ideas beyond where the originators thought they could go.@@@@1@34@nil@oe@24-11-2010 4700@unknown@formal@none@1@S@Andy Tanenbaum had the original idea to build a simple native Unix for IBM PCs, for use as a teaching tool (he called it Minix).@@@@1@25@nil@oe@24-11-2010 4710@unknown@formal@none@1@S@Linus Torvalds pushed the Minix concept further than Andrew probably thought it could go—and it grew into something wonderful.@@@@1@19@nil@oe@24-11-2010 4720@unknown@formal@none@1@S@In the same way (though on a smaller scale), I took some ideas by Carl Harris and Harry Hochheiser and pushed them hard.@@@@1@23@nil@oe@24-11-2010 4730@unknown@formal@none@1@S@Neither of us was `original' in the romantic way people think is genius.@@@@1@13@nil@oe@24-11-2010 4740@unknown@formal@none@1@S@But then, most science and engineering and software development isn't done by original genius, hacker mythology to the contrary.@@@@1@19@nil@oe@24-11-2010 4750@unknown@formal@none@1@S@The results were pretty heady stuff all the same—in fact, just the kind of success every hacker lives for!@@@@1@19@nil@oe@24-11-2010 4760@unknown@formal@none@1@S@And they meant I would have to set my standards even higher.@@@@1@12@nil@oe@24-11-2010 4770@unknown@formal@none@1@S@To make fetchmail as good as I now saw it could be, I'd have to write not just for my own needs, but also include and support features necessary to others but outside my orbit.@@@@1@35@nil@oe@24-11-2010 4780@unknown@formal@none@1@S@And do that while keeping the program simple and robust.@@@@1@10@nil@oe@24-11-2010 4790@unknown@formal@none@1@S@The first and overwhelmingly most important feature I wrote after realizing this was multidrop support—the ability to fetch mail from mailboxes that had accumulated all mail for a group of users, and then route each piece of mail to its individual recipients.@@@@1@42@nil@oe@24-11-2010 4800@unknown@formal@none@1@S@I decided to add the multidrop support partly because some users were clamoring for it, but mostly because I thought it would shake bugs out of the single-drop code by forcing me to deal with addressing in full generality.@@@@1@39@nil@oe@24-11-2010 4810@unknown@formal@none@1@S@And so it proved.@@@@1@4@nil@oe@24-11-2010 4820@unknown@formal@none@1@S@Getting RFC 822 address parsing right took me a remarkably long time, not because any individual piece of it is hard but because it involved a pile of interdependent and fussy details.@@@@1@32@nil@oe@24-11-2010 4830@unknown@formal@none@1@S@But multidrop addressing turned out to be an excellent design decision as well.@@@@1@13@nil@oe@24-11-2010 4840@unknown@formal@none@1@S@Here's how I knew:@@@@1@4@nil@oe@24-11-2010 4850@unknown@formal@none@1@S@14. Any tool should be useful in the expected way, but a truly great tool lends itself to uses you never expected.@@@@1@22@nil@oe@24-11-2010 4860@unknown@formal@none@1@S@The unexpected use for multidrop fetchmail is to run mailing lists with the list kept, and alias expansion done, on the client side of the Internet connection.@@@@1@27@nil@oe@24-11-2010 4870@unknown@formal@none@1@S@This means someone running a personal machine through an ISP account can manage a mailing list without continuing access to the ISP's alias files.@@@@1@24@nil@oe@24-11-2010 4880@unknown@formal@none@1@S@Another important change demanded by my beta-testers was support for 8-bit MIME (Multipurpose Internet Mail Extensions) operation.@@@@1@17@nil@oe@24-11-2010 4890@unknown@formal@none@1@S@This was pretty easy to do, because I had been careful to keep the code 8-bit clean (that is, to not press the 8th bit, unused in the ASCII character set, into service to carry information within the program).@@@@1@39@nil@oe@24-11-2010 4900@unknown@formal@none@1@S@Not because I anticipated the demand for this feature, but rather in obedience to another rule:@@@@1@16@nil@oe@24-11-2010 4910@unknown@formal@none@1@S@15. When writing gateway software of any kind, take pains to disturb the data stream as little as possible—and never throw away information unless the recipient forces you to!@@@@1@29@nil@oe@24-11-2010 4920@unknown@formal@none@1@S@Had I not obeyed this rule, 8-bit MIME support would have been difficult and buggy.@@@@1@15@nil@oe@24-11-2010 4930@unknown@formal@none@1@S@As it was, all I had to do is read the MIME standard (RFC 1652) and add a trivial bit of header-generation logic.@@@@1@23@nil@oe@24-11-2010 4940@unknown@formal@none@1@S@Some European users bugged me into adding an option to limit the number of messages retrieved per session (so they can control costs from their expensive phone networks).@@@@1@28@nil@oe@24-11-2010 4950@unknown@formal@none@1@S@I resisted this for a long time, and I'm still not entirely happy about it.@@@@1@15@nil@oe@24-11-2010 4960@unknown@formal@none@1@S@But if you're writing for the world, you have to listen to your customers—this doesn't change just because they're not paying you in money.@@@@1@24@nil@oe@24-11-2010 4970@unknown@formal@none@1@S@A Few More Lessons from Fetchmail@@@@1@6@nil@oe@24-11-2010 4980@unknown@formal@none@1@S@Before we go back to general software-engineering issues, there are a couple more specific lessons from the fetchmail experience to ponder.@@@@1@21@nil@oe@24-11-2010 4990@unknown@formal@none@1@S@Nontechnical readers can safely skip this section.@@@@1@7@nil@oe@24-11-2010 5000@unknown@formal@none@1@S@The rc (control) file syntax includes optional `noise' keywords that are entirely ignored by the parser.@@@@1@16@nil@oe@24-11-2010 5010@unknown@formal@none@1@S@The English-like syntax they allow is considerably more readable than the traditional terse keyword-value pairs you get when you strip them all out.@@@@1@23@nil@oe@24-11-2010 5020@unknown@formal@none@1@S@These started out as a late-night experiment when I noticed how much the rc file declarations were beginning to resemble an imperative minilanguage.@@@@1@23@nil@oe@24-11-2010 5030@unknown@formal@none@1@S@(This is also why I changed the original popclient ``server'' keyword to ``poll'').@@@@1@13@nil@oe@24-11-2010 5040@unknown@formal@none@1@S@It seemed to me that trying to make that imperative minilanguage more like English might make it easier to use.@@@@1@20@nil@oe@24-11-2010 5050@unknown@formal@none@1@S@Now, although I'm a convinced partisan of the ``make it a language'' school of design as exemplified by Emacs and HTML and many database engines, I am not normally a big fan of ``English-like'' syntaxes.@@@@1@35@nil@oe@24-11-2010 5060@unknown@formal@none@1@S@Traditionally programmers have tended to favor control syntaxes that are very precise and compact and have no redundancy at all.@@@@1@20@nil@oe@24-11-2010 5070@unknown@formal@none@1@S@This is a cultural legacy from when computing resources were expensive, so parsing stages had to be as cheap and simple as possible.@@@@1@23@nil@oe@24-11-2010 5080@unknown@formal@none@1@S@English, with about 50% redundancy, looked like a very inappropriate model then.@@@@1@12@nil@oe@24-11-2010 5090@unknown@formal@none@1@S@This is not my reason for normally avoiding English-like syntaxes; I mention it here only to demolish it.@@@@1@18@nil@oe@24-11-2010 5100@unknown@formal@none@1@S@With cheap cycles and core, terseness should not be an end in itself.@@@@1@13@nil@oe@24-11-2010 5110@unknown@formal@none@1@S@Nowadays it's more important for a language to be convenient for humans than to be cheap for the computer.@@@@1@19@nil@oe@24-11-2010 5120@unknown@formal@none@1@S@There remain, however, good reasons to be wary.@@@@1@8@nil@oe@24-11-2010 5130@unknown@formal@none@1@S@One is the complexity cost of the parsing stage—you don't want to raise that to the point where it's a significant source of bugs and user confusion in itself.@@@@1@29@nil@oe@24-11-2010 5140@unknown@formal@none@1@S@Another is that trying to make a language syntax English-like often demands that the ``English'' it speaks be bent seriously out of shape, so much so that the superficial resemblance to natural language is as confusing as a traditional syntax would have been.@@@@1@43@nil@oe@24-11-2010 5150@unknown@formal@none@1@S@(You see this bad effect in a lot of so-called ``fourth generation'' and commercial database-query languages.)@@@@1@16@nil@oe@24-11-2010 5160@unknown@formal@none@1@S@The fetchmail control syntax seems to avoid these problems because the language domain is extremely restricted.@@@@1@16@nil@oe@24-11-2010 5170@unknown@formal@none@1@S@It's nowhere near a general-purpose language; the things it says simply are not very complicated, so there's little potential for confusion in moving mentally between a tiny subset of English and the actual control language.@@@@1@35@nil@oe@24-11-2010 5180@unknown@formal@none@1@S@I think there may be a broader lesson here:@@@@1@9@nil@oe@24-11-2010 5190@unknown@formal@none@1@S@16. When your language is nowhere near Turing-complete, syntactic sugar can be your friend.@@@@1@14@nil@oe@24-11-2010 5200@unknown@formal@none@1@S@Another lesson is about security by obscurity.@@@@1@7@nil@oe@24-11-2010 5210@unknown@formal@none@1@S@Some fetchmail users asked me to change the software to store passwords encrypted in the rc file, so snoopers wouldn't be able to casually see them.@@@@1@26@nil@oe@24-11-2010 5220@unknown@formal@none@1@S@I didn't do it, because this doesn't actually add protection.@@@@1@10@nil@oe@24-11-2010 5230@unknown@formal@none@1@S@Anyone who's acquired permissions to read your rc file will be able to run fetchmail as you anyway—and if it's your password they're after, they'd be able to rip the necessary decoder out of the fetchmail code itself to get it.@@@@1@41@nil@oe@24-11-2010 5240@unknown@formal@none@1@S@All .fetchmailrc password encryption would have done is give a false sense of security to people who don't think very hard.@@@@1@21@nil@oe@24-11-2010 5250@unknown@formal@none@1@S@The general rule here is:@@@@1@5@nil@oe@24-11-2010 5260@unknown@formal@none@1@S@17. A security system is only as secure as its secret.@@@@1@11@nil@oe@24-11-2010 5270@unknown@formal@none@1@S@Beware of pseudo-secrets.@@@@1@3@nil@oe@24-11-2010 5280@unknown@formal@none@1@S@Necessary Preconditions for the Bazaar Style@@@@1@6@nil@oe@24-11-2010 5290@unknown@formal@none@1@S@Early reviewers and test audiences for this essay consistently raised questions about the preconditions for successful bazaar-style development, including both the qualifications of the project leader and the state of code at the time one goes public and starts to try to build a co-developer community.@@@@1@46@nil@oe@24-11-2010 5300@unknown@formal@none@1@S@It's fairly clear that one cannot code from the ground up in bazaar style.@@@@1@14@nil@oe@24-11-2010 5310@unknown@formal@none@1@S@One can test, debug and improve in bazaar style, but it would be very hard to originate a project in bazaar mode.@@@@1@22@nil@oe@24-11-2010 5320@unknown@formal@none@1@S@Linus didn't try it.@@@@1@4@nil@oe@24-11-2010 5330@unknown@formal@none@1@S@I didn't either.@@@@1@3@nil@oe@24-11-2010 5340@unknown@formal@none@1@S@Your nascent developer community needs to have something runnable and testable to play with.@@@@1@14@nil@oe@24-11-2010 5350@unknown@formal@none@1@S@When you start community-building, what you need to be able to present is a plausible promise.@@@@1@16@nil@oe@24-11-2010 5360@unknown@formal@none@1@S@Your program doesn't have to work particularly well.@@@@1@8@nil@oe@24-11-2010 5370@unknown@formal@none@1@S@It can be crude, buggy, incomplete, and poorly documented.@@@@1@9@nil@oe@24-11-2010 5380@unknown@formal@none@1@S@What it must not fail to do is (a) run, and (b) convince potential co-developers that it can be evolved into something really neat in the foreseeable future.@@@@1@28@nil@oe@24-11-2010 5390@unknown@formal@none@1@S@Linux and fetchmail both went public with strong, attractive basic designs.@@@@1@11@nil@oe@24-11-2010 5400@unknown@formal@none@1@S@Many people thinking about the bazaar model as I have presented it have correctly considered this critical, then jumped from that to the conclusion that a high degree of design intuition and cleverness in the project leader is indispensable.@@@@1@39@nil@oe@24-11-2010 5410@unknown@formal@none@1@S@But Linus got his design from Unix.@@@@1@7@nil@oe@24-11-2010 5420@unknown@formal@none@1@S@I got mine initially from the ancestral popclient (though it would later change a great deal, much more proportionately speaking than has Linux).@@@@1@23@nil@oe@24-11-2010 5430@unknown@formal@none@1@S@So does the leader/coordinator for a bazaar-style effort really have to have exceptional design talent, or can he get by through leveraging the design talent of others?@@@@1@27@nil@oe@24-11-2010 5440@unknown@formal@none@1@S@I think it is not critical that the coordinator be able to originate designs of exceptional brilliance, but it is absolutely critical that the coordinator be able to recognize good design ideas from others.@@@@1@34@nil@oe@24-11-2010 5450@unknown@formal@none@1@S@Both the Linux and fetchmail projects show evidence of this.@@@@1@10@nil@oe@24-11-2010 5460@unknown@formal@none@1@S@Linus, while not (as previously discussed) a spectacularly original designer, has displayed a powerful knack for recognizing good design and integrating it into the Linux kernel.@@@@1@26@nil@oe@24-11-2010 5470@unknown@formal@none@1@S@And I have already described how the single most powerful design idea in fetchmail (SMTP forwarding) came from somebody else.@@@@1@20@nil@oe@24-11-2010 5480@unknown@formal@none@1@S@Early audiences of this essay complimented me by suggesting that I am prone to undervalue design originality in bazaar projects because I have a lot of it myself, and therefore take it for granted.@@@@1@34@nil@oe@24-11-2010 5490@unknown@formal@none@1@S@There may be some truth to this; design (as opposed to coding or debugging) is certainly my strongest skill.@@@@1@19@nil@oe@24-11-2010 5500@unknown@formal@none@1@S@But the problem with being clever and original in software design is that it gets to be a habit—you start reflexively making things cute and complicated when you should be keeping them robust and simple.@@@@1@35@nil@oe@24-11-2010 5510@unknown@formal@none@1@S@I have had projects crash on me because I made this mistake, but I managed to avoid this with fetchmail.@@@@1@20@nil@oe@24-11-2010 5520@unknown@formal@none@1@S@So I believe the fetchmail project succeeded partly because I restrained my tendency to be clever; this argues (at least) against design originality being essential for successful bazaar projects.@@@@1@29@nil@oe@24-11-2010 5530@unknown@formal@none@1@S@And consider Linux.@@@@1@3@nil@oe@24-11-2010 5540@unknown@formal@none@1@S@Suppose Linus Torvalds had been trying to pull off fundamental innovations in operating system design during the development; does it seem at all likely that the resulting kernel would be as stable and successful as what we have?@@@@1@38@nil@oe@24-11-2010 5550@unknown@formal@none@1@S@A certain base level of design and coding skill is required, of course, but I expect almost anybody seriously thinking of launching a bazaar effort will already be above that minimum.@@@@1@31@nil@oe@24-11-2010 5560@unknown@formal@none@1@S@The open-source community's internal market in reputation exerts subtle pressure on people not to launch development efforts they're not competent to follow through on.@@@@1@24@nil@oe@24-11-2010 5570@unknown@formal@none@1@S@So far this seems to have worked pretty well.@@@@1@9@nil@oe@24-11-2010 5580@unknown@formal@none@1@S@There is another kind of skill not normally associated with software development which I think is as important as design cleverness to bazaar projects—and it may be more important.@@@@1@29@nil@oe@24-11-2010 5590@unknown@formal@none@1@S@A bazaar project coordinator or leader must have good people and communications skills.@@@@1@13@nil@oe@24-11-2010 5600@unknown@formal@none@1@S@This should be obvious.@@@@1@4@nil@oe@24-11-2010 5610@unknown@formal@none@1@S@In order to build a development community, you need to attract people, interest them in what you're doing, and keep them happy about the amount of work they're doing.@@@@1@29@nil@oe@24-11-2010 5620@unknown@formal@none@1@S@Technical sizzle will go a long way towards accomplishing this, but it's far from the whole story.@@@@1@17@nil@oe@24-11-2010 5630@unknown@formal@none@1@S@The personality you project matters, too.@@@@1@6@nil@oe@24-11-2010 5640@unknown@formal@none@1@S@It is not a coincidence that Linus is a nice guy who makes people like him and want to help him.@@@@1@21@nil@oe@24-11-2010 5650@unknown@formal@none@1@S@It's not a coincidence that I'm an energetic extrovert who enjoys working a crowd and has some of the delivery and instincts of a stand-up comic.@@@@1@26@nil@oe@24-11-2010 5660@unknown@formal@none@1@S@To make the bazaar model work, it helps enormously if you have at least a little skill at charming people.@@@@1@20@nil@oe@24-11-2010 5670@unknown@formal@none@1@S@The Social Context of Open-Source Software@@@@1@6@nil@oe@24-11-2010 5680@unknown@formal@none@1@S@It is truly written: the best hacks start out as personal solutions to the author's everyday problems, and spread because the problem turns out to be typical for a large class of users.@@@@1@33@nil@oe@24-11-2010 5690@unknown@formal@none@1@S@This takes us back to the matter of rule 1, restated in a perhaps more useful way:@@@@1@17@nil@oe@24-11-2010 5700@unknown@formal@none@1@S@18. To solve an interesting problem, start by finding a problem that is interesting to you.@@@@1@16@nil@oe@24-11-2010 5710@unknown@formal@none@1@S@So it was with Carl Harris and the ancestral popclient, and so with me and fetchmail.@@@@1@16@nil@oe@24-11-2010 5720@unknown@formal@none@1@S@But this has been understood for a long time.@@@@1@9@nil@oe@24-11-2010 5730@unknown@formal@none@1@S@The interesting point, the point that the histories of Linux and fetchmail seem to demand we focus on, is the next stage—the evolution of software in the presence of a large and active community of users and co-developers.@@@@1@38@nil@oe@24-11-2010 5740@unknown@formal@none@1@S@In The Mythical Man-Month, Fred Brooks observed that programmer time is not fungible; adding developers to a late software project makes it later.@@@@1@23@nil@oe@24-11-2010 5750@unknown@formal@none@1@S@As we've seen previously, he argued that the complexity and communication costs of a project rise with the square of the number of developers, while work done only rises linearly.@@@@1@30@nil@oe@24-11-2010 5760@unknown@formal@none@1@S@Brooks's Law has been widely regarded as a truism.@@@@1@9@nil@oe@24-11-2010 5770@unknown@formal@none@1@S@But we've examined in this essay a number of ways in which the process of open-source development falsifies the assumptions behind it—and, empirically, if Brooks's Law were the whole picture Linux would be impossible.@@@@1@34@nil@oe@24-11-2010 5780@unknown@formal@none@1@S@Gerald Weinberg's classic The Psychology of Computer Programming supplied what, in hindsight, we can see as a vital correction to Brooks.@@@@1@21@nil@oe@24-11-2010 5790@unknown@formal@none@1@S@In his discussion of ``egoless programming'', Weinberg observed that in shops where developers are not territorial about their code, and encourage other people to look for bugs and potential improvements in it, improvement happens dramatically faster than elsewhere.@@@@1@38@nil@oe@24-11-2010 5800@unknown@formal@none@1@S@(Recently, Kent Beck's `extreme programming' technique of deploying coders in pairs looking over one anothers' shoulders might be seen as an attempt to force this effect.)@@@@1@26@nil@oe@24-11-2010 5810@unknown@formal@none@1@S@Weinberg's choice of terminology has perhaps prevented his analysis from gaining the acceptance it deserved—one has to smile at the thought of describing Internet hackers as ``egoless''.@@@@1@27@nil@oe@24-11-2010 5820@unknown@formal@none@1@S@But I think his argument looks more compelling today than ever.@@@@1@11@nil@oe@24-11-2010 5830@unknown@formal@none@1@S@The bazaar method, by harnessing the full power of the ``egoless programming'' effect, strongly mitigates the effect of Brooks's Law.@@@@1@20@nil@oe@24-11-2010 5840@unknown@formal@none@1@S@The principle behind Brooks's Law is not repealed, but given a large developer population and cheap communications its effects can be swamped by competing nonlinearities that are not otherwise visible.@@@@1@30@nil@oe@24-11-2010 5850@unknown@formal@none@1@S@This resembles the relationship between Newtonian and Einsteinian physics—the older system is still valid at low energies, but if you push mass and velocity high enough you get surprises like nuclear explosions or Linux.@@@@1@34@nil@oe@24-11-2010 5860@unknown@formal@none@1@S@The history of Unix should have prepared us for what we're learning from Linux (and what I've verified experimentally on a smaller scale by deliberately copying Linus's methods).@@@@1@28@nil@oe@24-11-2010 5870@unknown@formal@none@1@S@That is, while coding remains an essentially solitary activity, the really great hacks come from harnessing the attention and brainpower of entire communities.@@@@1@23@nil@oe@24-11-2010 5880@unknown@formal@none@1@S@The developer who uses only his or her own brain in a closed project is going to fall behind the developer who knows how to create an open, evolutionary context in which feedback exploring the design space, code contributions, bug-spotting, and other improvements come from hundreds (perhaps thousands) of people.@@@@1@50@nil@oe@24-11-2010 5890@unknown@formal@none@1@S@But the traditional Unix world was prevented from pushing this approach to the ultimate by several factors.@@@@1@17@nil@oe@24-11-2010 5900@unknown@formal@none@1@S@One was the legal contraints of various licenses, trade secrets, and commercial interests.@@@@1@13@nil@oe@24-11-2010 5910@unknown@formal@none@1@S@Another (in hindsight) was that the Internet wasn't yet good enough.@@@@1@11@nil@oe@24-11-2010 5920@unknown@formal@none@1@S@Before cheap Internet, there were some geographically compact communities where the culture encouraged Weinberg's ``egoless'' programming, and a developer could easily attract a lot of skilled kibitzers and co-developers.@@@@1@29@nil@oe@24-11-2010 5930@unknown@formal@none@1@S@Bell Labs, the MIT AI and LCS labs, UC Berkeley—these became the home of innovations that are legendary and still potent.@@@@1@21@nil@oe@24-11-2010 5940@unknown@formal@none@1@S@Linux was the first project for which a conscious and successful effort to use the entire world as its talent pool was made.@@@@1@23@nil@oe@24-11-2010 5950@unknown@formal@none@1@S@I don't think it's a coincidence that the gestation period of Linux coincided with the birth of the World Wide Web, and that Linux left its infancy during the same period in 1993–1994 that saw the takeoff of the ISP industry and the explosion of mainstream interest in the Internet.@@@@1@50@nil@oe@24-11-2010 5960@unknown@formal@none@1@S@Linus was the first person who learned how to play by the new rules that pervasive Internet access made possible.@@@@1@20@nil@oe@24-11-2010 5970@unknown@formal@none@1@S@While cheap Internet was a necessary condition for the Linux model to evolve, I think it was not by itself a sufficient condition.@@@@1@23@nil@oe@24-11-2010 5980@unknown@formal@none@1@S@Another vital factor was the development of a leadership style and set of cooperative customs that could allow developers to attract co-developers and get maximum leverage out of the medium.@@@@1@30@nil@oe@24-11-2010 5990@unknown@formal@none@1@S@But what is this leadership style and what are these customs?@@@@1@11@nil@oe@24-11-2010 6000@unknown@formal@none@1@S@They cannot be based on power relationships—and even if they could be, leadership by coercion would not produce the results we see.@@@@1@22@nil@oe@24-11-2010 6010@unknown@formal@none@1@S@Weinberg quotes the autobiography of the 19th-century Russian anarchist Pyotr Alexeyvich Kropotkin's Memoirs of a Revolutionist to good effect on this subject:@@@@1@22@nil@oe@24-11-2010 6020@unknown@formal@none@1@S@Having been brought up in a serf-owner's family, I entered active life, like all young men of my time, with a great deal of confidence in the necessity of commanding, ordering, scolding, punishing and the like.@@@@1@36@nil@oe@24-11-2010 6030@unknown@formal@none@1@S@But when, at an early stage, I had to manage serious enterprises and to deal with [free] men, and when each mistake would lead at once to heavy consequences, I began to appreciate the difference between acting on the principle of command and discipline and acting on the principle of common understanding.@@@@1@52@nil@oe@24-11-2010 6040@unknown@formal@none@1@S@The former works admirably in a military parade, but it is worth nothing where real life is concerned, and the aim can be achieved only through the severe effort of many converging wills.@@@@1@33@nil@oe@24-11-2010 6050@unknown@formal@none@1@S@The ``severe effort of many converging wills'' is precisely what a project like Linux requires—and the ``principle of command'' is effectively impossible to apply among volunteers in the anarchist's paradise we call the Internet.@@@@1@34@nil@oe@24-11-2010 6060@unknown@formal@none@1@S@To operate and compete effectively, hackers who want to lead collaborative projects have to learn how to recruit and energize effective communities of interest in the mode vaguely suggested by Kropotkin's ``principle of understanding''.@@@@1@34@nil@oe@24-11-2010 6070@unknown@formal@none@1@S@They must learn to use Linus's Law.@@@@1@7@nil@oe@24-11-2010 6080@unknown@formal@none@1@S@Earlier I referred to the ``Delphi effect'' as a possible explanation for Linus's Law.@@@@1@14@nil@oe@24-11-2010 6090@unknown@formal@none@1@S@But more powerful analogies to adaptive systems in biology and economics also irresistably suggest themselves.@@@@1@15@nil@oe@24-11-2010 6100@unknown@formal@none@1@S@The Linux world behaves in many respects like a free market or an ecology, a collection of selfish agents attempting to maximize utility which in the process produces a self-correcting spontaneous order more elaborate and efficient than any amount of central planning could have achieved.@@@@1@45@nil@oe@24-11-2010 6110@unknown@formal@none@1@S@Here, then, is the place to seek the ``principle of understanding''.@@@@1@11@nil@oe@24-11-2010 6120@unknown@formal@none@1@S@The ``utility function'' Linux hackers are maximizing is not classically economic, but is the intangible of their own ego satisfaction and reputation among other hackers.@@@@1@25@nil@oe@24-11-2010 6130@unknown@formal@none@1@S@(One may call their motivation ``altruistic'', but this ignores the fact that altruism is itself a form of ego satisfaction for the altruist).@@@@1@23@nil@oe@24-11-2010 6140@unknown@formal@none@1@S@Voluntary cultures that work this way are not actually uncommon; one other in which I have long participated is science fiction fandom, which unlike hackerdom has long explicitly recognized ``egoboo'' (ego-boosting, or the enhancement of one's reputation among other fans) as the basic drive behind volunteer activity.@@@@1@47@nil@oe@24-11-2010 6150@unknown@formal@none@1@S@Linus, by successfully positioning himself as the gatekeeper of a project in which the development is mostly done by others, and nurturing interest in the project until it became self-sustaining, has shown an acute grasp of Kropotkin's ``principle of shared understanding''.@@@@1@41@nil@oe@24-11-2010 6160@unknown@formal@none@1@S@This quasi-economic view of the Linux world enables us to see how that understanding is applied.@@@@1@16@nil@oe@24-11-2010 6170@unknown@formal@none@1@S@We may view Linus's method as a way to create an efficient market in ``egoboo''—to connect the selfishness of individual hackers as firmly as possible to difficult ends that can only be achieved by sustained cooperation.@@@@1@36@nil@oe@24-11-2010 6180@unknown@formal@none@1@S@With the fetchmail project I have shown (albeit on a smaller scale) that his methods can be duplicated with good results.@@@@1@21@nil@oe@24-11-2010 6190@unknown@formal@none@1@S@Perhaps I have even done it a bit more consciously and systematically than he.@@@@1@14@nil@oe@24-11-2010 6200@unknown@formal@none@1@S@Many people (especially those who politically distrust free markets) would expect a culture of self-directed egoists to be fragmented, territorial, wasteful, secretive, and hostile.@@@@1@24@nil@oe@24-11-2010 6210@unknown@formal@none@1@S@But this expectation is clearly falsified by (to give just one example) the stunning variety, quality, and depth of Linux documentation.@@@@1@21@nil@oe@24-11-2010 6220@unknown@formal@none@1@S@It is a hallowed given that programmers hate documenting; how is it, then, that Linux hackers generate so much documentation?@@@@1@20@nil@oe@24-11-2010 6230@unknown@formal@none@1@S@Evidently Linux's free market in egoboo works better to produce virtuous, other-directed behavior than the massively-funded documentation shops of commercial software producers.@@@@1@22@nil@oe@24-11-2010 6240@unknown@formal@none@1@S@Both the fetchmail and Linux kernel projects show that by properly rewarding the egos of many other hackers, a strong developer/coordinator can use the Internet to capture the benefits of having lots of co-developers without having a project collapse into a chaotic mess.@@@@1@43@nil@oe@24-11-2010 6250@unknown@formal@none@1@S@So to Brooks's Law I counter-propose the following:@@@@1@8@nil@oe@24-11-2010 6260@unknown@formal@none@1@S@19: Provided the development coordinator has a communications medium at least as good as the Internet, and knows how to lead without coercion, many heads are inevitably better than one.@@@@1@30@nil@oe@24-11-2010 6270@unknown@formal@none@1@S@I think the future of open-source software will increasingly belong to people who know how to play Linus's game, people who leave behind the cathedral and embrace the bazaar.@@@@1@29@nil@oe@24-11-2010 6280@unknown@formal@none@1@S@This is not to say that individual vision and brilliance will no longer matter; rather, I think that the cutting edge of open-source software will belong to people who start from individual vision and brilliance, then amplify it through the effective construction of voluntary communities of interest.@@@@1@47@nil@oe@24-11-2010 6290@unknown@formal@none@1@S@Perhaps this is not only the future of open-source software.@@@@1@10@nil@oe@24-11-2010 6300@unknown@formal@none@1@S@No closed-source developer can match the pool of talent the Linux community can bring to bear on a problem.@@@@1@19@nil@oe@24-11-2010 6310@unknown@formal@none@1@S@Very few could afford even to hire the more than 200 (1999: 600, 2000: 800) people who have contributed to fetchmail!@@@@1@21@nil@oe@24-11-2010 6320@unknown@formal@none@1@S@Perhaps in the end the open-source culture will triumph not because cooperation is morally right or software ``hoarding'' is morally wrong (assuming you believe the latter, which neither Linus nor I do), but simply because the closed-source world cannot win an evolutionary arms race with open-source communities that can put orders of magnitude more skilled time into a problem.@@@@1@59@nil@oe@24-11-2010 6330@unknown@formal@none@1@S@On Management and the Maginot Line@@@@1@6@nil@oe@24-11-2010 6340@unknown@formal@none@1@S@The original Cathedral and Bazaar paper of 1997 ended with the vision above—that of happy networked hordes of programmer/anarchists outcompeting and overwhelming the hierarchical world of conventional closed software.@@@@1@29@nil@oe@24-11-2010 6350@unknown@formal@none@1@S@A good many skeptics weren't convinced, however; and the questions they raise deserve a fair engagement.@@@@1@16@nil@oe@24-11-2010 6360@unknown@formal@none@1@S@Most of the objections to the bazaar argument come down to the claim that its proponents have underestimated the productivity-multiplying effect of conventional management.@@@@1@24@nil@oe@24-11-2010 6370@unknown@formal@none@1@S@Traditionally-minded software-development managers often object that the casualness with which project groups form and change and dissolve in the open-source world negates a significant part of the apparent advantage of numbers that the open-source community has over any single closed-source developer.@@@@1@41@nil@oe@24-11-2010 6380@unknown@formal@none@1@S@They would observe that in software development it is really sustained effort over time and the degree to which customers can expect continuing investment in the product that matters, not just how many people have thrown a bone in the pot and left it to simmer.@@@@1@46@nil@oe@24-11-2010 6390@unknown@formal@none@1@S@There is something to this argument, to be sure; in fact, I have developed the idea that expected future service value is the key to the economics of software production in the essay The Magic Cauldron.@@@@1@36@nil@oe@24-11-2010 6400@unknown@formal@none@1@S@But this argument also has a major hidden problem; its implicit assumption that open-source development cannot deliver such sustained effort.@@@@1@20@nil@oe@24-11-2010 6410@unknown@formal@none@1@S@In fact, there have been open-source projects that maintained a coherent direction and an effective maintainer community over quite long periods of time without the kinds of incentive structures or institutional controls that conventional management finds essential.@@@@1@37@nil@oe@24-11-2010 6420@unknown@formal@none@1@S@The development of the GNU Emacs editor is an extreme and instructive example; it has absorbed the efforts of hundreds of contributors over 15 years into a unified architectural vision, despite high turnover and the fact that only one person (its author) has been continuously active during all that time.@@@@1@50@nil@oe@24-11-2010 6430@unknown@formal@none@1@S@No closed-source editor has ever matched this longevity record.@@@@1@9@nil@oe@24-11-2010 6440@unknown@formal@none@1@S@This suggests a reason for questioning the advantages of conventionally-managed software development that is independent of the rest of the arguments over cathedral vs. bazaar mode.@@@@1@26@nil@oe@24-11-2010 6450@unknown@formal@none@1@S@If it's possible for GNU Emacs to express a consistent architectural vision over 15 years, or for an operating system like Linux to do the same over 8 years of rapidly changing hardware and platform technology; and if (as is indeed the case) there have been many well-architected open-source projects of more than 5 years duration -- then we are entitled to wonder what, if anything, the tremendous overhead of conventionally-managed development is actually buying us.@@@@1@76@nil@oe@24-11-2010 6460@unknown@formal@none@1@S@Whatever it is certainly doesn't include reliable execution by deadline, or on budget, or to all features of the specification; it's a rare `managed' project that meets even one of these goals, let alone all three.@@@@1@36@nil@oe@24-11-2010 6470@unknown@formal@none@1@S@It also does not appear to be ability to adapt to changes in technology and economic context during the project lifetime, either; the open-source community has proven far more effective on that score (as one can readily verify, for example, by comparing the 30-year history of the Internet with the short half-lives of proprietary networking technologies—or the cost of the 16-bit to 32-bit transition in Microsoft Windows with the nearly effortless upward migration of Linux during the same period, not only along the Intel line of development but to more than a dozen other hardware platforms, including the 64-bit Alpha as well).@@@@1@102@nil@oe@24-11-2010 6480@unknown@formal@none@1@S@One thing many people think the traditional mode buys you is somebody to hold legally liable and potentially recover compensation from if the project goes wrong.@@@@1@26@nil@oe@24-11-2010 6490@unknown@formal@none@1@S@But this is an illusion; most software licenses are written to disclaim even warranty of merchantability, let alone performance—and cases of successful recovery for software nonperformance are vanishingly rare.@@@@1@29@nil@oe@24-11-2010 6500@unknown@formal@none@1@S@Even if they were common, feeling comforted by having somebody to sue would be missing the point.@@@@1@17@nil@oe@24-11-2010 6510@unknown@formal@none@1@S@You didn't want to be in a lawsuit; you wanted working software.@@@@1@12@nil@oe@24-11-2010 6520@unknown@formal@none@1@S@So what is all that management overhead buying?@@@@1@8@nil@oe@24-11-2010 6530@unknown@formal@none@1@S@In order to understand that, we need to understand what software development managers believe they do.@@@@1@16@nil@oe@24-11-2010 6540@unknown@formal@none@1@S@A woman I know who seems to be very good at this job says software project management has five functions:@@@@1@20@nil@oe@24-11-2010 6550@unknown@formal@none@1@S@To define goals and keep everybody pointed in the same direction@@@@1@11@nil@oe@24-11-2010 6560@unknown@formal@none@1@S@To monitor and make sure crucial details don't get skipped@@@@1@10@nil@oe@24-11-2010 6570@unknown@formal@none@1@S@To motivate people to do boring but necessary drudgework@@@@1@9@nil@oe@24-11-2010 6580@unknown@formal@none@1@S@To organize the deployment of people for best productivity@@@@1@9@nil@oe@24-11-2010 6590@unknown@formal@none@1@S@To marshal resources needed to sustain the project@@@@1@8@nil@oe@24-11-2010 6600@unknown@formal@none@1@S@Apparently worthy goals, all of these; but under the open-source model, and in its surrounding social context, they can begin to seem strangely irrelevant.@@@@1@24@nil@oe@24-11-2010 6610@unknown@formal@none@1@S@We'll take them in reverse order.@@@@1@6@nil@oe@24-11-2010 6620@unknown@formal@none@1@S@My friend reports that a lot of resource marshalling is basically defensive; once you have your people and machines and office space, you have to defend them from peer managers competing for the same resources, and from higher-ups trying to allocate the most efficient use of a limited pool.@@@@1@49@nil@oe@24-11-2010 6630@unknown@formal@none@1@S@But open-source developers are volunteers, self-selected for both interest and ability to contribute to the projects they work on (and this remains generally true even when they are being paid a salary to hack open source.)@@@@1@36@nil@oe@24-11-2010 6640@unknown@formal@none@1@S@The volunteer ethos tends to take care of the `attack' side of resource-marshalling automatically; people bring their own resources to the table.@@@@1@22@nil@oe@24-11-2010 6650@unknown@formal@none@1@S@And there is little or no need for a manager to `play defense' in the conventional sense.@@@@1@17@nil@oe@24-11-2010 6660@unknown@formal@none@1@S@Anyway, in a world of cheap PCs and fast Internet links, we find pretty consistently that the only really limiting resource is skilled attention.@@@@1@24@nil@oe@24-11-2010 6670@unknown@formal@none@1@S@Open-source projects, when they founder, essentially never do so for want of machines or links or office space; they die only when the developers themselves lose interest.@@@@1@27@nil@oe@24-11-2010 6680@unknown@formal@none@1@S@That being the case, it's doubly important that open-source hackers organize themselves for maximum productivity by self-selection—and the social milieu selects ruthlessly for competence.@@@@1@24@nil@oe@24-11-2010 6690@unknown@formal@none@1@S@My friend, familiar with both the open-source world and large closed projects, believes that open source has been successful partly because its culture only accepts the most talented 5% or so of the programming population.@@@@1@35@nil@oe@24-11-2010 6700@unknown@formal@none@1@S@She spends most of her time organizing the deployment of the other 95%, and has thus observed first-hand the well-known variance of a factor of one hundred in productivity between the most able programmers and the merely competent.@@@@1@38@nil@oe@24-11-2010 6710@unknown@formal@none@1@S@The size of that variance has always raised an awkward question: would individual projects, and the field as a whole, be better off without more than 50% of the least able in it?@@@@1@33@nil@oe@24-11-2010 6720@unknown@formal@none@1@S@Thoughtful managers have understood for a long time that if conventional software management's only function were to convert the least able from a net loss to a marginal win, the game might not be worth the candle.@@@@1@37@nil@oe@24-11-2010 6730@unknown@formal@none@1@S@The success of the open-source community sharpens this question considerably, by providing hard evidence that it is often cheaper and more effective to recruit self-selected volunteers from the Internet than it is to manage buildings full of people who would rather be doing something else.@@@@1@45@nil@oe@24-11-2010 6740@unknown@formal@none@1@S@Which brings us neatly to the question of motivation.@@@@1@9@nil@oe@24-11-2010 6750@unknown@formal@none@1@S@An equivalent and often-heard way to state my friend's point is that traditional development management is a necessary compensation for poorly motivated programmers who would not otherwise turn out good work.@@@@1@31@nil@oe@24-11-2010 6760@unknown@formal@none@1@S@This answer usually travels with a claim that the open-source community can only be relied on only to do work that is `sexy' or technically sweet; anything else will be left undone (or done only poorly) unless it's churned out by money-motivated cubicle peons with managers cracking whips over them.@@@@1@50@nil@oe@24-11-2010 6770@unknown@formal@none@1@S@I address the psychological and social reasons for being skeptical of this claim in Homesteading the Noosphere.@@@@1@17@nil@oe@24-11-2010 6780@unknown@formal@none@1@S@For present purposes, however, I think it's more interesting to point out the implications of accepting it as true.@@@@1@19@nil@oe@24-11-2010 6790@unknown@formal@none@1@S@If the conventional, closed-source, heavily-managed style of software development is really defended only by a sort of Maginot Line of problems conducive to boredom, then it's going to remain viable in each individual application area for only so long as nobody finds those problems really interesting and nobody else finds any way to route around them.@@@@1@56@nil@oe@24-11-2010 6800@unknown@formal@none@1@S@Because the moment there is open-source competition for a `boring' piece of software, customers are going to know that it was finally tackled by someone who chose that problem to solve because of a fascination with the problem itself—which, in software as in other kinds of creative work, is a far more effective motivator than money alone.@@@@1@57@nil@oe@24-11-2010 6810@unknown@formal@none@1@S@Having a conventional management structure solely in order to motivate, then, is probably good tactics but bad strategy; a short-term win, but in the longer term a surer loss.@@@@1@29@nil@oe@24-11-2010 6820@unknown@formal@none@1@S@So far, conventional development management looks like a bad bet now against open source on two points (resource marshalling, organization), and like it's living on borrowed time with respect to a third (motivation).@@@@1@33@nil@oe@24-11-2010 6830@unknown@formal@none@1@S@And the poor beleaguered conventional manager is not going to get any succour from the monitoring issue; the strongest argument the open-source community has is that decentralized peer review trumps all the conventional methods for trying to ensure that details don't get slipped.@@@@1@43@nil@oe@24-11-2010 6840@unknown@formal@none@1@S@Can we save defining goals as a justification for the overhead of conventional software project management?@@@@1@16@nil@oe@24-11-2010 6850@unknown@formal@none@1@S@Perhaps; but to do so, we'll need good reason to believe that management committees and corporate roadmaps are more successful at defining worthy and widely shared goals than the project leaders and tribal elders who fill the analogous role in the open-source world.@@@@1@43@nil@oe@24-11-2010 6860@unknown@formal@none@1@S@That is on the face of it a pretty hard case to make.@@@@1@13@nil@oe@24-11-2010 6870@unknown@formal@none@1@S@And it's not so much the open-source side of the balance (the longevity of Emacs, or Linus Torvalds's ability to mobilize hordes of developers with talk of ``world domination'') that makes it tough.@@@@1@33@nil@oe@24-11-2010 6880@unknown@formal@none@1@S@Rather, it's the demonstrated awfulness of conventional mechanisms for defining the goals of software projects.@@@@1@15@nil@oe@24-11-2010 6890@unknown@formal@none@1@S@One of the best-known folk theorems of software engineering is that 60% to 75% of conventional software projects either are never completed or are rejected by their intended users.@@@@1@29@nil@oe@24-11-2010 6900@unknown@formal@none@1@S@If that range is anywhere near true (and I've never met a manager of any experience who disputes it) then more projects than not are being aimed at goals that are either (a) not realistically attainable, or (b) just plain wrong.@@@@1@41@nil@oe@24-11-2010 6910@unknown@formal@none@1@S@This, more than any other problem, is the reason that in today's software engineering world the very phrase ``management committee'' is likely to send chills down the hearer's spine—even (or perhaps especially) if the hearer is a manager.@@@@1@38@nil@oe@24-11-2010 6920@unknown@formal@none@1@S@The days when only programmers griped about this pattern are long past; Dilbert cartoons hang over executives' desks now.@@@@1@19@nil@oe@24-11-2010 6930@unknown@formal@none@1@S@Our reply, then, to the traditional software development manager, is simple—if the open-source community has really underestimated the value of conventional management, why do so many of you display contempt for your own process?@@@@1@34@nil@oe@24-11-2010 6940@unknown@formal@none@1@S@Once again the example of the open-source community sharpens this question considerably—because we have fun doing what we do.@@@@1@19@nil@oe@24-11-2010 6950@unknown@formal@none@1@S@Our creative play has been racking up technical, market-share, and mind-share successes at an astounding rate.@@@@1@16@nil@oe@24-11-2010 6960@unknown@formal@none@1@S@We're proving not only that we can do better software, but that joy is an asset.@@@@1@16@nil@oe@24-11-2010 6970@unknown@formal@none@1@S@Two and a half years after the first version of this essay, the most radical thought I can offer to close with is no longer a vision of an open-source–dominated software world; that, after all, looks plausible to a lot of sober people in suits these days.@@@@1@47@nil@oe@24-11-2010 6980@unknown@formal@none@1@S@Rather, I want to suggest what may be a wider lesson about software, (and probably about every kind of creative or professional work).@@@@1@23@nil@oe@24-11-2010 6990@unknown@formal@none@1@S@Human beings generally take pleasure in a task when it falls in a sort of optimal-challenge zone; not so easy as to be boring, not too hard to achieve.@@@@1@29@nil@oe@24-11-2010 7000@unknown@formal@none@1@S@A happy programmer is one who is neither underutilized nor weighed down with ill-formulated goals and stressful process friction.@@@@1@19@nil@oe@24-11-2010 7010@unknown@formal@none@1@S@Enjoyment predicts efficiency.@@@@1@3@nil@oe@24-11-2010 7020@unknown@formal@none@1@S@Relating to your own work process with fear and loathing (even in the displaced, ironic way suggested by hanging up Dilbert cartoons) should therefore be regarded in itself as a sign that the process has failed.@@@@1@36@nil@oe@24-11-2010 7030@unknown@formal@none@1@S@Joy, humor, and playfulness are indeed assets; it was not mainly for the alliteration that I wrote of "happy hordes" above, and it is no mere joke that the Linux mascot is a cuddly, neotenous penguin.@@@@1@36@nil@oe@24-11-2010 7040@unknown@formal@none@1@S@It may well turn out that one of the most important effects of open source's success will be to teach us that play is the most economically efficient mode of creative work.@@@@1@32@nil@oe@24-11-2010 7050@unknown@formal@none@1@S@Epilog: Netscape Embraces the Bazaar@@@@1@5@nil@oe@24-11-2010 7060@unknown@formal@none@1@S@It's a strange feeling to realize you're helping make history....@@@@1@10@nil@oe@24-11-2010 7070@unknown@formal@none@1@S@On January 22 1998, approximately seven months after I first published The Cathedral and the Bazaar, Netscape Communications, Inc. announced plans to give away the source for Netscape Communicator.@@@@1@29@nil@oe@24-11-2010 7080@unknown@formal@none@1@S@I had had no clue this was going to happen before the day of the announcement.@@@@1@16@nil@oe@24-11-2010 7090@unknown@formal@none@1@S@Eric Hahn, executive vice president and chief technology officer at Netscape, emailed me shortly afterwards as follows: ``On behalf of everyone at Netscape, I want to thank you for helping us get to this point in the first place.@@@@1@39@nil@oe@24-11-2010 7100@unknown@formal@none@1@S@Your thinking and writings were fundamental inspirations to our decision.''@@@@1@10@nil@oe@24-11-2010 7110@unknown@formal@none@1@S@The following week I flew out to Silicon Valley at Netscape's invitation for a day-long strategy conference (on 4 Feb 1998) with some of their top executives and technical people.@@@@1@30@nil@oe@24-11-2010 7120@unknown@formal@none@1@S@We designed Netscape's source-release strategy and license together.@@@@1@8@nil@oe@24-11-2010 7130@unknown@formal@none@1@S@A few days later I wrote the following:@@@@1@8@nil@oe@24-11-2010 7140@unknown@formal@none@1@S@Netscape is about to provide us with a large-scale, real-world test of the bazaar model in the commercial world.@@@@1@19@nil@oe@24-11-2010 7150@unknown@formal@none@1@S@The open-source culture now faces a danger; if Netscape's execution doesn't work, the open-source concept may be so discredited that the commercial world won't touch it again for another decade.@@@@1@30@nil@oe@24-11-2010 7160@unknown@formal@none@1@S@On the other hand, this is also a spectacular opportunity.@@@@1@10@nil@oe@24-11-2010 7170@unknown@formal@none@1@S@Initial reaction to the move on Wall Street and elsewhere has been cautiously positive.@@@@1@14@nil@oe@24-11-2010 7180@unknown@formal@none@1@S@We're being given a chance to prove ourselves, too.@@@@1@9@nil@oe@24-11-2010 7190@unknown@formal@none@1@S@If Netscape regains substantial market share through this move, it just may set off a long-overdue revolution in the software industry.@@@@1@21@nil@oe@24-11-2010 7200@unknown@formal@none@1@S@The next year should be a very instructive and interesting time.@@@@1@11@nil@oe@24-11-2010 7210@unknown@formal@none@1@S@And indeed it was.@@@@1@4@nil@oe@24-11-2010 7220@unknown@formal@none@1@S@As I write in mid-2000, the development of what was later named Mozilla has been only a qualified success.@@@@1@19@nil@oe@24-11-2010 7230@unknown@formal@none@1@S@It achieved Netscape's original goal, which was to deny Microsoft a monopoly lock on the browser market.@@@@1@17@nil@oe@24-11-2010 7240@unknown@formal@none@1@S@It has also achieved some dramatic successes (notably the release of the next-generation Gecko rendering engine).@@@@1@16@nil@oe@24-11-2010 7250@unknown@formal@none@1@S@However, it has not yet garnered the massive development effort from outside Netscape that the Mozilla founders had originally hoped for.@@@@1@21@nil@oe@24-11-2010 7260@unknown@formal@none@1@S@The problem here seems to be that for a long time the Mozilla distribution actually broke one of the basic rules of the bazaar model; it didn't ship with something potential contributors could easily run and see working.@@@@1@38@nil@oe@24-11-2010 7270@unknown@formal@none@1@S@(Until more than a year after release, building Mozilla from source required a license for the proprietary Motif library.)@@@@1@19@nil@oe@24-11-2010 7280@unknown@formal@none@1@S@Most negatively (from the point of view of the outside world) the Mozilla group didn't ship a production-quality browser for two and a half years after the project launch—and in 1999 one of the project's principals caused a bit of a sensation by resigning, complaining of poor management and missed opportunities.@@@@1@51@nil@oe@24-11-2010 7290@unknown@formal@none@1@S@``Open source,'' he correctly observed, ``is not magic pixie dust.''@@@@1@10@nil@oe@24-11-2010 7300@unknown@formal@none@1@S@And indeed it is not.@@@@1@5@nil@oe@24-11-2010 7310@unknown@formal@none@1@S@The long-term prognosis for Mozilla looks dramatically better now (in November 2000) than it did at the time of Jamie Zawinski's resignation letter—in the last few weeks the nightly releases have finally passed the critical threshold to production usability.@@@@1@39@nil@oe@24-11-2010 7320@unknown@formal@none@1@S@But Jamie was right to point out that going open will not necessarily save an existing project that suffers from ill-defined goals or spaghetti code or any of the software engineering's other chronic ills.@@@@1@34@nil@oe@24-11-2010 7330@unknown@formal@none@1@S@Mozilla has managed to provide an example simultaneously of how open source can succeed and how it could fail.@@@@1@19@nil@oe@24-11-2010 7340@unknown@formal@none@1@S@In the mean time, however, the open-source idea has scored successes and found backers elsewhere.@@@@1@15@nil@oe@24-11-2010 7350@unknown@formal@none@1@S@Since the Netscape release we've seen a tremendous explosion of interest in the open-source development model, a trend both driven by and driving the continuing success of the Linux operating system.@@@@1@31@nil@oe@24-11-2010 7360@unknown@formal@none@1@S@The trend Mozilla touched off is continuing at an accelerating rate.@@@@1@11@nil@oe@24-11-2010 7370@unknown@formal@none@1@S@Notes@@@@1@1@nil@oe@24-11-2010 7380@unknown@formal@none@1@S@In Programming Pearls, the noted computer-science aphorist Jon Bentley comments on Brooks's observation with ``If you plan to throw one away, you will throw away two.''.@@@@1@26@nil@oe@24-11-2010 7390@unknown@formal@none@1@S@He is almost certainly right.@@@@1@5@nil@oe@24-11-2010 7400@unknown@formal@none@1@S@The point of Brooks's observation, and Bentley's, isn't merely that you should expect the first attempt to be wrong, it's that starting over with the right idea is usually more effective than trying to salvage a mess.@@@@1@37@nil@oe@24-11-2010 7410@unknown@formal@none@1@S@Examples of successful open-source, bazaar development predating the Internet explosion and unrelated to the Unix and Internet traditions have existed.@@@@1@20@nil@oe@24-11-2010 7420@unknown@formal@none@1@S@The development of the info-Zip compression utility during 1990–1992, primarily for DOS machines, was one such example.@@@@1@17@nil@oe@24-11-2010 7430@unknown@formal@none@1@S@Another was the RBBS bulletin board system (again for DOS), which began in 1983 and developed a sufficiently strong community that there have been fairly regular releases up to the present (mid-1999) despite the huge technical advantages of Internet mail and file-sharing over local BBSs.@@@@1@45@nil@oe@24-11-2010 7440@unknown@formal@none@1@S@While the info-Zip community relied to some extent on Internet mail, the RBBS developer culture was actually able to base a substantial on-line community on RBBS that was completely independent of the TCP/IP infrastructure.@@@@1@34@nil@oe@24-11-2010 7450@unknown@formal@none@1@S@That transparency and peer review are valuable for taming the complexity of OS development turns out, after all, not to be a new concept.@@@@1@24@nil@oe@24-11-2010 7460@unknown@formal@none@1@S@In 1965, very early in the history of time-sharing operating systems, Corbató and Vyssotsky, co-designers of the Multics operating system, wrote@@@@1@21@nil@oe@24-11-2010 7470@unknown@formal@none@1@S@It is expected that the Multics system will be published when it is operating substantially...@@@@1@15@nil@oe@24-11-2010 7480@unknown@formal@none@1@S@Such publication is desirable for two reasons: First, the system should withstand public scrutiny and criticism volunteered by interested readers; second, in an age of increasing complexity, it is an obligation to present and future system designers to make the inner operating system as lucid as possible so as to reveal the basic system issues.@@@@1@55@nil@oe@24-11-2010 7490@unknown@formal@none@1@S@John Hasler has suggested an interesting explanation for the fact that duplication of effort doesn't seem to be a net drag on open-source development.@@@@1@24@nil@oe@24-11-2010 7500@unknown@formal@none@1@S@He proposes what I'll dub ``Hasler's Law'': the costs of duplicated work tend to scale sub-quadratically with team size—that is, more slowly than the planning and management overhead that would be needed to eliminate them.@@@@1@35@nil@oe@24-11-2010 7510@unknown@formal@none@1@S@This claim actually does not contradict Brooks's Law.@@@@1@8@nil@oe@24-11-2010 7520@unknown@formal@none@1@S@It may be the case that total complexity overhead and vulnerability to bugs scales with the square of team size, but that the costs from duplicated work are nevertheless a special case that scales more slowly.@@@@1@36@nil@oe@24-11-2010 7530@unknown@formal@none@1@S@It's not hard to develop plausible reasons for this, starting with the undoubted fact that it is much easier to agree on functional boundaries between different developers' code that will prevent duplication of effort than it is to prevent the kinds of unplanned bad interactions across the whole system that underly most bugs.@@@@1@53@nil@oe@24-11-2010 7540@unknown@formal@none@1@S@The combination of Linus's Law and Hasler's Law suggests that there are actually three critical size regimes in software projects.@@@@1@20@nil@oe@24-11-2010 7550@unknown@formal@none@1@S@On small projects (I would say one to at most three developers) no management structure more elaborate than picking a lead programmer is needed.@@@@1@24@nil@oe@24-11-2010 7560@unknown@formal@none@1@S@And there is some intermediate range above that in which the cost of traditional management is relatively low, so its benefits from avoiding duplication of effort, bug-tracking, and pushing to see that details are not overlooked actually net out positive.@@@@1@40@nil@oe@24-11-2010 7570@unknown@formal@none@1@S@Above that, however, the combination of Linus's Law and Hasler's Law suggests there is a large-project range in which the costs and problems of traditional management rise much faster than the expected cost from duplication of effort.@@@@1@37@nil@oe@24-11-2010 7580@unknown@formal@none@1@S@Not the least of these costs is a structural inability to harness the many-eyeballs effect, which (as we've seen) seems to do a much better job than traditional management at making sure bugs and details are not overlooked.@@@@1@38@nil@oe@24-11-2010 7590@unknown@formal@none@1@S@Thus, in the large-project case, the combination of these laws effectively drives the net payoff of traditional management to zero.@@@@1@20@nil@oe@24-11-2010 7600@unknown@formal@none@1@S@The split between Linux's experimental and stable versions has another function related to, but distinct from, hedging risk.@@@@1@18@nil@oe@24-11-2010 7610@unknown@formal@none@1@S@The split attacks another problem: the deadliness of deadlines.@@@@1@9@nil@oe@24-11-2010 7620@unknown@formal@none@1@S@When programmers are held both to an immutable feature list and a fixed drop-dead date, quality goes out the window and there is likely a colossal mess in the making.@@@@1@30@nil@oe@24-11-2010 7630@unknown@formal@none@1@S@I am indebted to Marco Iansiti and Alan MacCormack of the Harvard Business School for showing me evidence that relaxing either one of these constraints can make scheduling workable.@@@@1@29@nil@oe@24-11-2010 7640@unknown@formal@none@1@S@One way to do this is to fix the deadline but leave the feature list flexible, allowing features to drop off if not completed by deadline.@@@@1@26@nil@oe@24-11-2010 7650@unknown@formal@none@1@S@This is essentially the strategy of the "stable" kernel branch; Alan Cox (the stable-kernel maintainer) puts out releases at fairly regular intervals, but makes no guarantees about when particular bugs will be fixed or what features will beback-ported from the experimental branch.@@@@1@42@nil@oe@24-11-2010 7660@unknown@formal@none@1@S@The other way to do this is to set a desired feature list and deliver only when it is done.@@@@1@20@nil@oe@24-11-2010 7670@unknown@formal@none@1@S@This is essentially the strategy of the "experimental" kernel branch.@@@@1@10@nil@oe@24-11-2010 7680@unknown@formal@none@1@S@De Marco and Lister cited research showing that this scheduling policy ("wake me up when it's done") produces not only the highest quality but, on average, shorter delivery times than either "realistic" or "aggressive" scheduling.@@@@1@35@nil@oe@24-11-2010 7690@unknown@formal@none@1@S@I have come to suspect (as of early 2000) that in earlier versions of this essay I severely underestimated the importance of the "wake me up when it's done" anti-deadline policy to the open-source community's productivity and quality.@@@@1@38@nil@oe@24-11-2010 7700@unknown@formal@none@1@S@General experience with the rushed GNOME 1.0 release in 1999 suggests that pressure for a premature release can neutralize many of the quality benefits open source normally confers.@@@@1@28@nil@oe@24-11-2010 7710@unknown@formal@none@1@S@It may well turn out to be that the process transparency of open source is one of three co-equal drivers of its quality, along with "wake me up when it's done" scheduling and developer self-selection.@@@@1@35@nil@oe@24-11-2010 7720@unknown@formal@none@1@S@It's tempting, and not entirely inaccurate, to see the core-plus-halo organization characteristic of open-source projects as an Internet-enabled spin on Brooks's own recommendation for solving the N-squared complexity problem, the "surgical-team" organization—but the differences are significant.@@@@1@36@nil@oe@24-11-2010 7730@unknown@formal@none@1@S@The constellation of specialist roles such as "code librarian" that Brooks envisioned around the team leader doesn't really exist; those roles are executed instead by generalists aided by toolsets quite a bit more powerful than those of Brooks's day.@@@@1@39@nil@oe@24-11-2010 7740@unknown@formal@none@1@S@Also, the open-source culture leans heavily on strong Unix traditions of modularity, APIs, and information hiding—none of which were elements of Brooks's prescription.@@@@1@23@nil@oe@24-11-2010 7750@unknown@formal@none@1@S@The respondent who pointed out to me the effect of widely varying trace path lengths on the difficulty of characterizing a bug speculated that trace-path difficulty for multiple symptoms of the same bug varies "exponentially" (which I take to mean on a Gaussian or Poisson distribution, and agree seems very plausible).@@@@1@51@nil@oe@24-11-2010 7760@unknown@formal@none@1@S@If it is experimentally possible to get a handle on the shape of this distribution, that would be extremely valuable data.@@@@1@21@nil@oe@24-11-2010 7770@unknown@formal@none@1@S@Large departures from a flat equal-probability distribution of trace difficulty would suggest that even solo developers should emulate the bazaar strategy by bounding the time they spend on tracing a given symptom before they switch to another.@@@@1@37@nil@oe@24-11-2010 7780@unknown@formal@none@1@S@Persistence may not always be a virtue...@@@@1@7@nil@oe@24-11-2010 7790@unknown@formal@none@1@S@An issue related to whether one can start projects from zero in the bazaar style is whether the bazaar style is capable of supporting truly innovative work.@@@@1@27@nil@oe@24-11-2010 7800@unknown@formal@none@1@S@Some claim that, lacking strong leadership, the bazaar can only handle the cloning and improvement of ideas already present at the engineering state of the art, but is unable to push the state of the art.@@@@1@36@nil@oe@24-11-2010 7810@unknown@formal@none@1@S@This argument was perhaps most infamously made by the Halloween Documents, two embarrassing internal Microsoft memoranda written about the open-source phenomenon.@@@@1@21@nil@oe@24-11-2010 7820@unknown@formal@none@1@S@The authors compared Linux's development of a Unix-like operating system to ``chasing taillights'', and opined ``(once a project has achieved "parity" with the state-of-the-art), the level of management necessary to push towards new frontiers becomes massive.''@@@@1@36@nil@oe@24-11-2010 7830@unknown@formal@none@1@S@There are serious errors of fact implied in this argument.@@@@1@10@nil@oe@24-11-2010 7840@unknown@formal@none@1@S@One is exposed when the Halloween authors themselves later observe that ``often [...] new research ideas are first implemented and available on Linux before they are available / incorporated into other platforms.''@@@@1@32@nil@oe@24-11-2010 7850@unknown@formal@none@1@S@If we read ``open source'' for ``Linux'', we see that this is far from a new phenomenon.@@@@1@17@nil@oe@24-11-2010 7860@unknown@formal@none@1@S@Historically, the open-source community did not invent Emacs or the World Wide Web or the Internet itself by chasing taillights or being massively managed—and in the present, there is so much innovative work going on in open source that one is spoiled for choice.@@@@1@44@nil@oe@24-11-2010 7870@unknown@formal@none@1@S@The GNOME project (to pick one of many) is pushing the state of the art in GUIs and object technology hard enough to have attracted considerable notice in the computer trade press well outside the Linux community.@@@@1@37@nil@oe@24-11-2010 7880@unknown@formal@none@1@S@Other examples are legion, as a visit to Freshmeat on any given day will quickly prove.@@@@1@16@nil@oe@24-11-2010 7890@unknown@formal@none@1@S@But there is a more fundamental error in the implicit assumption that the cathedral model (or the bazaar model, or any other kind of management structure) can somehow make innovation happen reliably.@@@@1@32@nil@oe@24-11-2010 7900@unknown@formal@none@1@S@This is nonsense.@@@@1@3@nil@oe@24-11-2010 7910@unknown@formal@none@1@S@Gangs don't have breakthrough insights—even volunteer groups of bazaar anarchists are usually incapable of genuine originality, let alone corporate committees of people with a survival stake in some status quo ante.@@@@1@31@nil@oe@24-11-2010 7920@unknown@formal@none@1@S@Insight comes from individuals.@@@@1@4@nil@oe@24-11-2010 7930@unknown@formal@none@1@S@The most their surrounding social machinery can ever hope to do is to be responsive to breakthrough insights—to nourish and reward and rigorously test them instead of squashing them.@@@@1@29@nil@oe@24-11-2010 7940@unknown@formal@none@1@S@Some will characterize this as a romantic view, a reversion to outmoded lone-inventor stereotypes.@@@@1@14@nil@oe@24-11-2010 7950@unknown@formal@none@1@S@Not so; I am not asserting that groups are incapable of developing breakthrough insights once they have been hatched; indeed, we learn from the peer-review process that such development groups are essential to producing a high-quality result.@@@@1@37@nil@oe@24-11-2010 7960@unknown@formal@none@1@S@Rather I am pointing out that every such group development starts from—is necessarily sparked by—one good idea in one person's head.@@@@1@21@nil@oe@24-11-2010 7970@unknown@formal@none@1@S@Cathedrals and bazaars and other social structures can catch that lightning and refine it, but they cannot make it on demand.@@@@1@21@nil@oe@24-11-2010 7980@unknown@formal@none@1@S@Therefore the root problem of innovation (in software, or anywhere else) is indeed how not to squash it—but, even more fundamentally, it is how to grow lots of people who can have insights in the first place.@@@@1@37@nil@oe@24-11-2010 7990@unknown@formal@none@1@S@To suppose that cathedral-style development could manage this trick but the low entry barriers and process fluidity of the bazaar cannot would be absurd.@@@@1@24@nil@oe@24-11-2010 8000@unknown@formal@none@1@S@If what it takes is one person with one good idea, then a social milieu in which one person can rapidly attract the cooperation of hundreds or thousands of others with that good idea is going inevitably to out-innovate any in which the person has to do a political sales job to a hierarchy before he can work on his idea without risk of getting fired.@@@@1@66@nil@oe@24-11-2010 8010@unknown@formal@none@1@S@And, indeed, if we look at the history of software innovation by organizations using the cathedral model, we quickly find it is rather rare.@@@@1@24@nil@oe@24-11-2010 8020@unknown@formal@none@1@S@Large corporations rely on university research for new ideas (thus the Halloween Documents authors' unease about Linux's facility at coopting that research more rapidly).@@@@1@24@nil@oe@24-11-2010 8030@unknown@formal@none@1@S@Or they buy out small companies built around some innovator's brain.@@@@1@11@nil@oe@24-11-2010 8040@unknown@formal@none@1@S@In neither case is the innovation native to the cathedral culture; indeed, many innovations so imported end up being quietly suffocated under the "massive level of management" the Halloween Documents' authors so extol.@@@@1@33@nil@oe@24-11-2010 8050@unknown@formal@none@1@S@That, however, is a negative point.@@@@1@6@nil@oe@24-11-2010 8060@unknown@formal@none@1@S@The reader would be better served by a positive one.@@@@1@10@nil@oe@24-11-2010 8070@unknown@formal@none@1@S@I suggest, as an experiment, the following:@@@@1@7@nil@oe@24-11-2010 8080@unknown@formal@none@1@S@Pick a criterion for originality that you believe you can apply consistently.@@@@1@12@nil@oe@24-11-2010 8090@unknown@formal@none@1@S@If your definition is ``I know it when I see it'', that's not a problem for purposes of this test.@@@@1@20@nil@oe@24-11-2010 8100@unknown@formal@none@1@S@Pick any closed-source operating system competing with Linux, and a best source for accounts of current development work on it.@@@@1@20@nil@oe@24-11-2010 8110@unknown@formal@none@1@S@Watch that source and Freshmeat for one month.@@@@1@8@nil@oe@24-11-2010 8120@unknown@formal@none@1@S@Every day, count the number of release announcements on Freshmeat that you consider `original' work.@@@@1@15@nil@oe@24-11-2010 8130@unknown@formal@none@1@S@Apply the same definition of `original' to announcements for that other OS and count them.@@@@1@15@nil@oe@24-11-2010 8140@unknown@formal@none@1@S@Thirty days later, total up both figures.@@@@1@7@nil@oe@24-11-2010 8150@unknown@formal@none@1@S@The day I wrote this, Freshmeat carried twenty-two release announcements, of which three appear they might push state of the art in some respect.@@@@1@24@nil@oe@24-11-2010 8160@unknown@formal@none@1@S@This was a slow day for Freshmeat, but I will be astonished if any reader reports as many as three likely innovations a month in any closed-source channel.@@@@1@28@nil@oe@24-11-2010 8170@unknown@formal@none@1@S@We now have history on a project that, in several ways, may provide a more indicative test of the bazaar premise than fetchmail; EGCS, the Experimental GNU Compiler System.@@@@1@29@nil@oe@24-11-2010 8180@unknown@formal@none@1@S@This project was announced in mid-August of 1997 as a conscious attempt to apply the ideas in the early public versions of The Cathedral and the Bazaar.@@@@1@27@nil@oe@24-11-2010 8190@unknown@formal@none@1@S@The project founders felt that the development of GCC, the Gnu C Compiler, had been stagnating.@@@@1@16@nil@oe@24-11-2010 8200@unknown@formal@none@1@S@For about twenty months afterwards, GCC and EGCS continued as parallel products—both drawing from the same Internet developer population, both starting from the same GCC source base, both using pretty much the same Unix toolsets and development environment.@@@@1@38@nil@oe@24-11-2010 8210@unknown@formal@none@1@S@The projects differed only in that EGCS consciously tried to apply the bazaar tactics I have previously described, while GCC retained a more cathedral-like organization with a closed developer group and infrequent releases.@@@@1@33@nil@oe@24-11-2010 8220@unknown@formal@none@1@S@This was about as close to a controlled experiment as one could ask for, and the results were dramatic.@@@@1@19@nil@oe@24-11-2010 8230@unknown@formal@none@1@S@Within months, the EGCS versions had pulled substantially ahead in features; better optimization, better support for FORTRAN and C++.@@@@1@19@nil@oe@24-11-2010 8240@unknown@formal@none@1@S@Many people found the EGCS development snapshots to be more reliable than the most recent stable version of GCC, and major Linux distributions began to switch to EGCS.@@@@1@28@nil@oe@24-11-2010 8250@unknown@formal@none@1@S@In April of 1999, the Free Software Foundation (the official sponsors of GCC) dissolved the original GCC development group and officially handed control of the project to the the EGCS steering team.@@@@1@32@nil@oe@24-11-2010 8260@unknown@formal@none@1@S@Of course, Kropotkin's critique and Linus's Law raise some wider issues about the cybernetics of social organizations.@@@@1@17@nil@oe@24-11-2010 8270@unknown@formal@none@1@S@Another folk theorem of software engineering suggests one of them; Conway's Law—commonly stated as ``If you have four groups working on a compiler, you'll get a 4-pass compiler''.@@@@1@28@nil@oe@24-11-2010 8280@unknown@formal@none@1@S@The original statement was more general: ``Organizations which design systems are constrained to produce designs which are copies of the communication structures of these organizations.''@@@@1@25@nil@oe@24-11-2010 8290@unknown@formal@none@1@S@We might put it more succinctly as ``The means determine the ends'', or even ``Process becomes product''.@@@@1@17@nil@oe@24-11-2010 8300@unknown@formal@none@1@S@It is accordingly worth noting that in the open-source community organizational form and function match on many levels.@@@@1@18@nil@oe@24-11-2010 8310@unknown@formal@none@1@S@The network is everything and everywhere: not just the Internet, but the people doing the work form a distributed, loosely coupled, peer-to-peer network that provides multiple redundancy and degrades very gracefully.@@@@1@31@nil@oe@24-11-2010 8320@unknown@formal@none@1@S@In both networks, each node is important only to the extent that other nodes want to cooperate with it.@@@@1@19@nil@oe@24-11-2010 8330@unknown@formal@none@1@S@The peer-to-peer part is essential to the community's astonishing productivity.@@@@1@10@nil@oe@24-11-2010 8340@unknown@formal@none@1@S@The point Kropotkin was trying to make about power relationships is developed further by the `SNAFU Principle': ``True communication is possible only between equals, because inferiors are more consistently rewarded for telling their superiors pleasant lies than for telling the truth.''@@@@1@41@nil@oe@24-11-2010 8350@unknown@formal@none@1@S@Creative teamwork utterly depends on true communication and is thus very seriously hindered by the presence of power relationships.@@@@1@19@nil@oe@24-11-2010 8360@unknown@formal@none@1@S@The open-source community, effectively free of such power relationships, is teaching us by contrast how dreadfully much they cost in bugs, in lowered productivity, and in lost opportunities.@@@@1@28@nil@oe@24-11-2010 8370@unknown@formal@none@1@S@Further, the SNAFU principle predicts in authoritarian organizations a progressive disconnect between decision-makers and reality, as more and more of the input to those who decide tends to become pleasant lies.@@@@1@31@nil@oe@24-11-2010 8380@unknown@formal@none@1@S@The way this plays out in conventional software development is easy to see; there are strong incentives for the inferiors to hide, ignore, and minimize problems.@@@@1@26@nil@oe@24-11-2010 8390@unknown@formal@none@1@S@When this process becomes product, software is a disaster.@@@@1@9@nil@oe@24-11-2010 8400@unknown@formal@none@1@S@Bibliography@@@@1@1@nil@oe@24-11-2010 8410@unknown@formal@none@1@S@I quoted several bits from Frederick P. Brooks's classic The Mythical Man-Month because, in many respects, his insights have yet to be improved upon.@@@@1@24@nil@oe@24-11-2010 8420@unknown@formal@none@1@S@I heartily recommend the 25th Anniversary edition from Addison-Wesley (ISBN 0-201-83595-9), which adds his 1986 ``No Silver Bullet'' paper.@@@@1@19@nil@oe@24-11-2010 8430@unknown@formal@none@1@S@The new edition is wrapped up by an invaluable 20-years-later retrospective in which Brooks forthrightly admits to the few judgements in the original text which have not stood the test of time.@@@@1@32@nil@oe@24-11-2010 8440@unknown@formal@none@1@S@I first read the retrospective after the first public version of this essay was substantially complete, and was surprised to discover that Brooks attributed bazaar-like practices to Microsoft!@@@@1@28@nil@oe@24-11-2010 8450@unknown@formal@none@1@S@(In fact, however, this attribution turned out to be mistaken.@@@@1@10@nil@oe@24-11-2010 8460@unknown@formal@none@1@S@In 1998 we learned from the Halloween Documents that Microsoft's internal developer community is heavily balkanized, with the kind of general source access needed to support a bazaar not even truly possible.)@@@@1@32@nil@oe@24-11-2010 8470@unknown@formal@none@1@S@Gerald M. Weinberg's The Psychology Of Computer Programming (New York, Van Nostrand Reinhold 1971) introduced the rather unfortunately-labeled concept of ``egoless programming''.@@@@1@22@nil@oe@24-11-2010 8480@unknown@formal@none@1@S@While he was nowhere near the first person to realize the futility of the ``principle of command'', he was probably the first to recognize and argue the point in particular connection with software development.@@@@1@34@nil@oe@24-11-2010 8490@unknown@formal@none@1@S@Richard P. Gabriel, contemplating the Unix culture of the pre-Linux era, reluctantly argued for the superiority of a primitive bazaar-like model in his 1989 paper ``LISP: Good News, Bad News, and How To Win Big''.@@@@1@35@nil@oe@24-11-2010 8500@unknown@formal@none@1@S@Though dated in some respects, this essay is still rightly celebrated among LISP fans (including me).@@@@1@16@nil@oe@24-11-2010 8510@unknown@formal@none@1@S@A correspondent reminded me that the section titled ``Worse Is Better'' reads almost as an anticipation of Linux.@@@@1@18@nil@oe@24-11-2010 8520@unknown@formal@none@1@S@The paper is accessible on the World Wide Web at http://www.naggum.no/worse-is-better.html.@@@@1@11@nil@oe@24-11-2010 8530@unknown@formal@none@1@S@De Marco and Lister's Peopleware: Productive Projects and Teams (New York; Dorset House, 1987; ISBN 0-932633-05-6) is an underappreciated gem which I was delighted to see Fred Brooks cite in his retrospective.@@@@1@32@nil@oe@24-11-2010 8540@unknown@formal@none@1@S@While little of what the authors have to say is directly applicable to the Linux or open-source communities, the authors' insight into the conditions necessary for creative work is acute and worthwhile for anyone attempting to import some of the bazaar model's virtues into a commercial context.@@@@1@47@nil@oe@24-11-2010 8550@unknown@formal@none@1@S@Finally, I must admit that I very nearly called this essay ``The Cathedral and the Agora'', the latter term being the Greek for an open market or public meeting place.@@@@1@30@nil@oe@24-11-2010 8560@unknown@formal@none@1@S@The seminal ``agoric systems'' papers by Mark Miller and Eric Drexler, by describing the emergent properties of market-like computational ecologies, helped prepare me to think clearly about analogous phenomena in the open-source culture when Linux rubbed my nose in them five years later.@@@@1@43@nil@oe@24-11-2010 8570@unknown@formal@none@1@S@These papers are available on the Web at http://www.agorics.com/agorpapers.html.@@@@1@9@nil@oe@24-11-2010 8580@unknown@formal@none@1@S@Acknowledgements@@@@1@1@nil@oe@24-11-2010 8590@unknown@formal@none@1@S@This essay was improved by conversations with a large number of people who helped debug it.@@@@1@16@nil@oe@24-11-2010 8600@unknown@formal@none@1@S@Particular thanks to Jeff Dutky , who suggested the ``debugging is parallelizable'' formulation, and helped develop the analysis that proceeds from it.@@@@1@22@nil@oe@24-11-2010 8610@unknown@formal@none@1@S@Also to Nancy Lebovitz for her suggestion that I emulate Weinberg by quoting Kropotkin.@@@@1@15@nil@oe@24-11-2010 8620@unknown@formal@none@1@S@Perceptive criticisms also came from Joan Eslinger and Marty Franz of the General Technics list.@@@@1@17@nil@oe@24-11-2010 8630@unknown@formal@none@1@S@Glen Vandenburg pointed out the importance of self-selection in contributor populations and suggested the fruitful idea that much development rectifies `bugs of omission'; Daniel Upper suggested the natural analogies for this.@@@@1@33@nil@oe@24-11-2010 8640@unknown@formal@none@1@S@I'm grateful to the members of PLUG, the Philadelphia Linux User's group, for providing the first test audience for the first public version of this essay.@@@@1@26@nil@oe@24-11-2010 8650@unknown@formal@none@1@S@Paula Matuszek enlightened me about the practice of software management.@@@@1@11@nil@oe@24-11-2010 8660@unknown@formal@none@1@S@Phil Hudson reminded me that the social organization of the hacker culture mirrors the organization of its software, and vice-versa.@@@@1@21@nil@oe@24-11-2010 8670@unknown@formal@none@1@S@John Buck pointed out that MATLAB makes an instructive parallel to Emacs.@@@@1@13@nil@oe@24-11-2010 8680@unknown@formal@none@1@S@Russell Johnston brought me to consciousness about some of the mechanisms discussed in ``How Many Eyeballs Tame Complexity.''@@@@1@19@nil@oe@24-11-2010 8690@unknown@formal@none@1@S@Finally, Linus Torvalds's comments were helpful and his early endorsement very encouraging.@@@@1@12@nil@oe@24-11-2010