This week's interview guest with Eric S. Raymond. We got a *lot* of good questions, forwarded the moderators' favorites to Eric, and he not only answered the ones we sent him but - extra cool - picked some more out of the crowd and answered them, too. Read the complete session (below) and if you have something you want to add, go ahead. If Eric has time, he'll jump in and respond, because, well, he's just that kind of guy. ;) Note: questions marked with * are the ones Eric added to the moderators' selections.
Astute readers know why you've reluctantly taken a position as a Linux evangelist, open source sociologist, and prime target. Taking the opposite approach, is there anything which would convince you to step down, that your posts were no longer necessary?
This is not meant to be inflammatory ... it's just a roundabout way of asking how far along your goals are, and what your plans will be if you ever meet them.
Three things could cause me to step down:
- One: someone emerging to do the public-advocate job clearly better than I do.
- Two: Linux's market share going over 50%. (Cool down, BSD guys -- I'd be equally pleased to see some other open-source Unix win, it just doesn't seen very likely at this point).
- Three: a collapse in Microsoft's stock price. That would mean the end of effective FUD and countermarketing against open source.
A while ago, we read from you that being the Open Source advocate you are was wearing you down and influencing your life very badly. Did you cut down on advocating and did it help? In other words, did you get your life back?
Not really. Something more remarkable happened instead; the community responded to my distress call by growing up a little. I got letters of apology from some of the worst flamers. Many people in the rest of the community started pressuring the pinheads who had been making my job harder to shut up or get constructive.
I have also cut down somewhat on my travel schedule, but not as much as I thought earlier this year I would have to. I'm also demanding (and getting) better travel conditions -- business class instead of the cheap seats in coach. It makes a difference, more of one than I would have thought.
Stephen Williams asks*:
I'm glad to see that, after a three-year break, the Jargon File has been updated over the past few months. Is version 5.0.0 in the works? Are there any plans to release an update to the print version, The New Hacker's Dictionary, any time soon?
I've discussed the possibility with people at O'Reilly. That might be my second-to-next book, after "The Cathedral and the Bazaar" and before "The Art Of Unix Programming" (which is about half-done now but could take me another nine months to finish). Whether I go with O'Reilly or the publisher of the previous editions (MIT Press) the fourth edition of TNHD seems likely to come out next year sometime.
Tom Christiansen asks:
I don't know how to ask this question without it sounding like stirring the pot, but what about the growing chasm between free software (giftware) and GNU software (the viral kind, not the nice LGPL kind)? This is a real issue for some people in some situations. Think about the many BSD resellers and vendors who have custom packaging in highly competitive fields, like video editing? Doesn't the friction hurt everyone? Apple has turned to BSD not Linux, and the GPL is cited as one reason why. This seems to be devisive. There are no end of flamewars on /. and elsewhere, and the heat diminishes the light. What kind of reconciliation is possible? Or is "take no prisoners" just the way it has to work?
I don't see a chasm there, Tom. After all, we're all still writing and exchanging code. We're all using basically the same set of licenses. I don't think there are properly two different movements at all, outside the imaginations of a few rather fanatical partisans on both sides.
Here is the reality test: if you're running a project and someone sends you a patch, will you stop to enquire whether that person is a member of the correct faction before you apply it? I don't think so...
So despite the verbal fireworks and philosophical disputes, we're all hackers together. What unites us is more important than what divides us.
You say you want to live in a world where software doesn't suck. I couldn't agree more. However, do you see closed source software on an open source OS as a step in the right direction, or just likely to be a more stable platform on which to run your potentially bug-ridden software?
Step in the right direction, definitely. As more and more infrastructure goes open, and the remaimning closed-source applications increasingly use it for leverage, the overall quality of the applications will go up.
Recent interest shown by large commercial tech companies (IBM, SGI, Sun) seems to signal a new chapter in the history of Linux. Do you see the participation of these companies strengthening the linux communitity? Destroying it? Or transforming it into something completely different?
Look around you. What do you see, compared to a year ago?
Do you see fewer Linux hackers writing open source, or more? Do you see fewer hackers getting *paid* to write open source, or more? I think the answer is pretty clear.
Do you see our designs, or our licenses, or our coding practices being changed in any significant way by corporate participation? Again, I think the answer is pretty clear.
The truth is, they're not transforming us. We're transforming them.
I know that you are on the board of directors at VA Linux, what does your job entail?
My job at VA mostly involves sitting in a board meeting once a month asking searching questions about what the firm is doing and why. My role there (as Larry Augustin describes it) is to be the official corporate conscience. This mainly involves nipping bad ideas in the bud, before they flower into something that would piss off the hacker community. I have not had to do this often.
You've always been involved in hacker projects outside of just coding (eg. the Jargon File), but over the last year or so the spokesperson role seems to have grown into a fulltime job. How long is it since you last sat down to write a major piece of software? Do you expect to go back to fulltime development work anytime soon, and if so, what would you work on? How do you manage to cope with the withdrawal symptoms?
An astute question ;-). I haven't sat down to write a major piece of software from scratch in months, but I am continuing to maintain fetchmail. I just took over the gif2png beta code with Greg Roelofs's consent; the 1.0.0 version might be out by the time you read this. Today I did some work on gnuplot, bringing the PNG driver up to date.
If I get to go full-time again soon, I want to go back to work on Trove, the distributed web-based code-archiving system I designed last year. I'd also like to work with Guido van Rossum on Python 1.6; there are some long-time wishlist features like rich comparisons and a full lambda facility that I care enough about to implement myself. I also have a strategy-gaming system I wrote back in the 1980s that I'd like to put a modern (Web-based) interface on. Finally, having contributed a bit of code to GNOME (the network-monitor applet) I'd like to balance things by doing something for KDE.
This has probably been asked before, but I can't recall seeing the answer to it anywhere. What originally led you to write The Cathedral and the Bazaar? -- what I'm interested in is if there was some event or impetus that prompted you to write it down. Obviously you'd have no way of predicting the firestorm that followed, but it's always intriguing to know about the spark that started it all
I wrote CatB as a way of coping with my astonishment in the face of the Linux phenomenon. What I observed was that the community around Linux had evolved a way to write software that (a) was tremendously effective, (b) violated the classic Brooks's Law rules, and (c) was completely unconscious! Nobody reflected on what they were doing; it was practice without theory. I wrote CatB as an attempt to help my tribe become more conscious about what it has been doing.
We all know that you are a staunch advocate of libertarianism. Do you see the open-source / free-software movement turning into a larger political push for libertarian, minimal government?
What conferences are you planning to attend this year? Do you have plans for organizing Geeks with Guns outings during them? If so, is there a mailing list or some other source of information about how to join?
No comment on that first question. But, if you could see my face, I'm wearing a very evil grin....
See my speaking calendar for the conferences I plan to attend. As for GWG, there's no mailing list; would you like to host one? I rely on local organizers to find a range, and I don't have one for Atlanta Linux Showcase yet.
Linux, like all things in the computer world, will eventually become obsolete or maybe just too much work to keep "up to date". Linus (er, Dr. Torvalds) even said in his "Open Sources" essay that (paraphrasing) someone else could come along and write something better which will take Linux's place. How long do you think before someone will have an offering that will obsolete (or at least prove a competitor to) Linux and the BSD's?
I doubt Linux will have a real technical competitor for a long time, because I think it will probably just absorb new architectural ideas, amoeba-like, as they evolve. Twenty years from now the core APIs may have grown and changed tremendously, but we'll still think of it as the `same' codebase and call it Linux :-).
Is the friction between Gnome and KDE, BSD and GPL, Free Software and Open Source, and the other sources of flame war a bad thing or a good thing for the movement? Many people seem to feel that the competition is devisive, but isn't it the opposite? We're always preaching that competition is a good thing for the entire market, but then we complain when any of our pet projects are pitted head to head with another. The passion felt by the proponents of each philosophy seems to result in better, more quality work. Isn't this proof that competition is the Good Thing we've been saying it is all along?
I think you answered your own question :-).
Which of the coders working on open source projects do you admire the most? A particular big name like Linus, or someone less well-known?
Hmmm. I don't think there's anyone I can say I admire the most. There's a level of ability beyond which trying to make comparisons between people just gets silly, because each of the people that good has become a sort of perfect master of his own domain. Linus. RMS. Larry Wall. Guido van Rossum. James Gosling. Going further back, Ken Thompson or Dennis Ritchie. Anyway, I find these guys have gotten their fill of being admired, so I try to be friends with them instead.
Why isn't there an entry for "free software" in the Jargon Dictionary? Was this a politically-motivated decision?
Zounds! You know, until this moment, I didn't realize that entry was missing.
I don't think you want me to write it, though. I would find it hard to avoid using phrases like "rhetorical millstone around our necks" and "held us back for fifteen years". Care to submit one yourself?
Paul Crowley asks:
In Understand my job, please! you described Bruce Perens's proposal that we have a team of Linux advocates sharing the load as "glib". Could you say more about why you feel this way - isn't it more likely that a job where the load is shared would be more attractive?
I think I answered that question in the same paragraph you quoted. What makes the job rough isn't the workload, it's the second-guessers and snipers from the sidelines -- among whom Bruce was, at the time, nearly the worst. Connect the dots yourself.
Starting an open source project from nothing but people with a common interest is difficult. It's been my experience that it is very easy to founder with a bazaar approach to architecture and design. The issues tend to get confused with religious wars about toolkits and license choice, and just a lot of differing opinions about how to best structure a program, no one of which may be *obviously* better.
Is it essential for individuals to first create a working model, incomplete and buggy it may be, before applying bazaar development? Or what would you suggest in terms of managing a bazaar approach to creating programs from a bare idea?
I wouldn't. I think you're right; the successful projects have a core of individual vision around which the bazaar community nucleates.
Since, as we all know, cheese is the most powerful substance in the universe, I was wondering what your favorite source of ultimate power is?
That would have to be sex, because I'm allergic to cheese.
Next week: Bruce Sterling.