Lars Wirzenius: June, 2006


Wednesday, June 28, 2006

Debian: Re-qualification of Debian developers

The Debian New Maintainer process (NM) is well-known to be strict and tedious. It tests the skills, aptitudes, and attitudes of candidates in a number of areas, including technical, philosophical, and license interpreation. The process keeps the minimum level of competency of new maintainers at a reasonably high level.

We have, however, no process for ensuring that old farts don't stink.

Take myself for an example. I have been with Debian for over a decade, but I still find new surprises in things that are from the early days. Just recently, I learned about dch, which has existed since 1997 or before. This discover made my life much easier: I had grown tired of updating debian/changelog dates by hand, so they tended to be rather inexact. Now, when I do an upload, the changelog date is no longer a lie.

The problem is not that I am incapable of learning, or that I don't want to learn. It's just that I don't, unless I make an effort to do so. Given the small amount of free time I have, I tend to spend it on other things.

Most Debian developers probably keep up with most changes to policy, tools, procedures, best practices, and such, without anyone prodding them. However, during the past year and a half, when I've been doing quality assurance work, I've noticed that there seems to be a number of people who don't give any indication of current competence.

There's two approaches we can choose from to deal with this. We can educate, or we can re-test.

Education means writing documentation, announcing important changes, maybe giving IRC or video tutorials over the Internet, and so on. It is quite a bit of work, but if done properly, it would have quite an impact.

Re-testing means putting everyone through NM at regular intervals. Certainly not the entire NM, that would be way too annoying, but a quick slimmed down version would be possible. Something that ensures that every DD still has at least a minimal level of competence, but doesn't take more than an hour or two to do, unless one is a decade out of date.

There is of course the problem of what to do with people who fail the test. The simple approach would be to give them, say, two months of time to get up to date, and then they are re-tested, and if that too fails, they're retired.

I'm not sure whether we should do either. Perhaps we could try both.

Monday, June 19, 2006

Debian: Packaging tutorial (again)

I gave my Debian packaging tutorial today. The Computer Science department of the University of Helsinki provided the space (a classroom with networking), for which I hereby offer grateful thanks.

There were fourteen students, and everyone made a package out of my hello world program (especially written for this tutorial). The slides used my liwc package as an example, and during the talk I noticed a couple of problems, which promptly resulted in a bug report. Excellent!

Nobody wanted to do a victory dance, people are too shy.

I've now given this tutorial twice: once for three of my friends privately, and once for a larger group. It only covers the very basics of creating packages, and there's much more to learn. Some day I might do a second part, and cover things like maintainer scripts, advanced dependencies, debhelper, pbuilder, and piuparts (the last three were only briefly touched upon today).

I might be willing to do the first part again, if there's enough interest. Not very soon, though, teaching is surprisingly exhausting and time-consuming. However, if you'd be interested, please do mail me (at

Saturday, June 17, 2006

Random thought: Dick and Ollie

My friend Dick rarely asks anyone anything. If he needs some information, he looks it up himself, on the Internet, in a book, from source code, or wherever. Only after spending considerable effort himself will he ask someone else. He takes this so far that he'll get very upset if you give him the information. It ruins his day to get derailed, and then he sulks the rest of the day.

My other friend Ollie, on the other hand, always asks immediately when he has a problem. Either on IRC, by e-mail, by calling on the phone, or by shouting over a cubicle wall. He'll even run to the other side of the building to find someone he suspects has the answer. Only if he can't find anyone with the answer will he start looking for the answer himself.

Both Dick and Ollie are smart. Both are successful in their careers and both are regarded well enough by their colleagues, but only Dick is considered an expert.

Looking things up yourself is slower and harder, in the short run. You spend a lot of energy finding the little bit of information that solves your immediate problem. On the other hand, you learn things on the way. You learn about things related to or surrounding the immediate bit, and most importantly, you learn where to find more information. You might even learn that the bit you're looking for is the wrong bit, and there's an altogether different way of looking at the problem that lets you do things in a much better way.

There's nothing wrong in asking questions, but on the whole, it's better for yourself to try finding the answer yourself first.

In a hacker culture, where being good at something is valued exceptionally highly, someone asking frequent questions is often considered annoying. Doubly so, if they seem to be an Ollie.

Don't be an Ollie. Do yourself a favor and be a Dick.

(This entry was written some time ago, in reaction to a certain someone's umpteenth question of the day, and seeing people responding simply with Google's domain. I've postponed publishing it because I didn't want to attack that person directly.)

Wednesday, June 14, 2006

Random thought: Nettiaddressi tiedostomuodoista

This is in Finnish, since it internal to Finland. Sorry.

FLUG, FIMUG, FUUG ja Linux-Aktivaattori pistivät pystyyn nettiaddressin sen puolesta, että valtionhallinto käyttäisi vapaita tiedostomuotoja. Kannatan. Menkää lukemaan ja allekirjoittamaan: addressi.

Tuesday, June 13, 2006

Debian: Packaging tutorial/workshop

My Debian packaging tutorial will happen next Monday, June 19. The info page for this is at and you should monitor that page for updates.

This is a fairly short notice, sorry. I suck at organizing things.

Friday, June 09, 2006

Travel report: Debconf6, plus rest and relaxation in Los Angeles

This May I spent four consecutive weeks abroad, visiting first Debconf6 (including the Debcamp that preceded it), and then Los Angeles afterwards. I won't write a detailed report of everything that happened; nobody likes to read pages after pages of descriptions of meals and such. Instead, I'll jot down a few memories of moments. It's still long, sorry about that.

May 6, about 3 AM, at home. Waiting outside for a taxi. Pretty tired already: spent the night frantically weighing all my clothes to optimize my packing. It is warm enough to wait outside without a jacket. The night is calm and beautiful. Mankind is so much nicer when it sleeps.

May 6, about 4 AM, the Helsinki airport. I bump into Tuukka and Tulitar, who are on the same plane. Kaol also joins us, but that I knew beforehand. I'm suddenly much less apprehensive about the long flight.

May 6, Mexico City airport. We meet more Debian people. The first taxi driver we're assigned seems to be suspect somehow. The other taxi drivers rebel and our luggage is put into another car, with another driver. I have no idea what is actually going on, but that develops into a theme for the Mexican part of the trip.

May 6, Oaxtepec. We're finally here, but it seems very difficult for us to get rooms. While we wait, we and some of the people who arrived earlier go to a restaurant. The Germans want to order the same thing for everyone, but fail to agree on what, so eventually we order all the steaks on the menu. The food is good, but the restaurant staff would probably have preferred to go home at closing time. We eventually get rooms and I collapse into bed.

Debcamp week, Oaxtepec. We have network. We don't have network. We have network. We don't have network. We have network. We don't have network. We're frustrated all the time. I was supposed to do billable work (related to Debian, even), and can't. Should've prepared for this possiblity. I'm too dependent on a good Internet connection, need to figure out to reduce this dependency.

To have something else to do, I start the Mugshots project: take a photo of everyone together with a paper with their name, or IRC nick, or anything else they want to say about themselves.

Debconf, late one evening. Everyone is partying one way or another, drinking, dancing, discussing. Lots of people around already. Nice people. Still I withdraw into myself. I'm physically present, but not interacting. Never know what to say; exhausted by having to remember who everyone is. I'm introverted, and deal better with smaller gatherings. My life's ambition is to become a multi-millionaire recluse.

Debconf, late another evening, Oaxtepec. Everyone else is getting drunk, and it affects me. I make the silly bet about Debian releasing on time: if we do, I'll get a Debian tattoo. Not to worry, I win either way.

Debconf, some evening or another. I participate in a game of Mao. It is an unpleasant game, for me. I suspected this beforehand, but now I've verified it. It is best for everyone if I never play again, now that I know enough to at least try to reach my personal, alternative goals in the game, which make the game less fun for everyone else. (Better for me too, since I make fewer enemies.)

Debconf, day after formal dinner. I learn that Jonathan (Ted) Walther is gone. Good riddance. I suprise myself with the amount of glee I feel when removing him from the Mugshots gallery. (That, incidentally, was something I decided all by myself. It's my gallery, after all.)

Debconf. The sponsored food is quite acceptable, but eventually becomes boring and bland. Mass-kitchen food can't but help to be like that, I guess, unless it gets expensive. I don't want to see another chicken leg ever.

Debconf, early in the week. Droidy (Leena) and Burger (Ville) arrive. Very nice. Helps massively with my withdrawnness to have real life friends around. I still go early to bed, though, instead of participating in the partying.

Debconf, late in the week. I get Ville and Frans Pop to talk and Ville agrees to have look at the installation manual. He's an actual (former) professional tech writer, so hopefully will be able to improve it a lot.

May 22, Oaxtepec. Last day. I spend all morning packing, checking out, and waiting for a taxi. Strangely relaxing. Debconf is finally over, and although I did enjoy parts of it, and don't regret coming, I'm not sad it's over.

May 22, evening, Los Angeles. Julie and Kristian meet me at the airport. Hugging happens. Happiness happens. I get my first taste of LA traffic, which seems to mostly consist of people complaining about how slow it is while driving very fast.

May 23, the Promenade, Los Angeles. A three block long walking street of shops, with three (count them: three) bookstores. The two I have time for even smelled nice. I am in hog heaven. Colossal credit card cringing.

May 24, Getty Center. A very nice, very impressive place architecturally. The art is good too. As usual, the impressionists make the strongest impression on me, but there's lots of other good stuff, too.

Not everything, though: Robert Adams's photographs leave me unimpressed, possibly I lack the cultural reference points necessary to appreciate them. He has many photos of shabby temporary shelter-like houses that in reality seem to be nice south-western US houses from the 60s and 70s. My intuition just screams that the houses and their inhabitants won't survive their first winter. But of course, they don't have a cold winter there.

The rest of the exhibition (or what I have time to see) has way too many portraits for one session. Portraits get boring after the first dozen or so, I'm afraid, unless there's something extraordinary about them.

They are several rooms rebuilt from 18th century French aristocratic houses. Very nice, very pretty, but I would not like to live that uncomfortably.

I have a small epiphany about nudity in art in one of the rooms with statues. I've always known nudity could be portrayed erotically, such as having two lovers embrace passionately; or that it could be portrayed more or less abstractly, using the human body, or parts of it, just because it can be a very beautiful subject, just like a flower, but without giving it a greater significance. It's decoration, nothing more.

There are two statues in the Getty collection that show me a third way. They are exquisitely sensual, evoking a sense of form and texture that make me think of human skin and flesh, a sense that if one were to touch the statue, it would feel like touching a live human. Yet none of this is erotic at all.

I'm not sure I can explain this, one may have to experience it for oneself. Unfortunately, I was too overwhelmed to write down the names of the statues.

May 23, Los Angeles. The food takes me by surprise. Julie has always told me that food in LA is good, much better than in Finland, but she hadn't prepared me for an all-out assault on my palate. Dine well in LA and die happy. (When I go back to Finland, I will have to learn again how to cook, dammit.)

May 25, Los Angeles. Jesus reminds me via SMS about the Towel Day. I participate. My hosts think I'm a bit weird, but in a funny way.

Many happy days in May, Los Angeles. I visit the Los Angeles County Museum of Art (exhibiting a weird Italian designer, some boring Klimt paintings, and a very nice general collection); the Natural history museum of LA (dinosaurs! real skeletons! big ones! T.Rex! for the first time in my life! they're huge!); the Museum of Jurassic Technology (weird); had dinner with some local Debian people, and visited the local Google office afterwards; the Hollywood Boulevard at night (funny, for a while). And other places I've forgotten already. LA is not small.

I visit a political theater play that is brilliant. I also visit my first American BBQ, and have my first experience with the LA gay community at the same time. I'm later told that these two are separable. I enjoy them both a lot.

I visit more bookstores. My hosts keep finding new ones they want to recommend to me.

Parking would be a pain in LA, except that it means that people do actually walk a lot, since the nearest parking space is always three blocks away, or five levels up without an elevator.

May 28, Los Angeles. I innocently mention that LA is such a nice place, with nice people and gorgeous food, that I wouldn't mind living there. Julie's brother hears this and immediately connects me with someone he knows whose company is looking for new people. During the next three days, we meet and talk and things seem favorable.

Even if I decide not to go (it may be too huge a change for me now), it's nice to feel appreciated.

June 1, Los Angeles. Leaving for home. Not a sad thing, but I'm definitely feeling wistful. I get hugged about fourteen thousand times. Parting is such sweet sorrow.

I will definitely have to go back to LA, one day.

Tuesday, June 06, 2006

Rant: SMS and e-mail tax in the EU

The European Union wants to tax SMS and e-mail messages (see digitoday news in Finnish). Or at least the French MEP Alain Lamassoure has suggested that, it seems. The tax would be 1.5 cents per SMS, and 0.00001 (one hundred thousandth) cent per e-mail.

Current prices in Finnish operators seem to be around 7-10 cents per SMS message. A 1.5 cent tax on top of that seems pretty high. Collecting it would be simple: all SMS messages go through operators.

For e-mail, however, the thing is going to be a nightmare. Much e-mail goes not via operators, and there is no good way of tracking it. This could be fixed (from the tax collecting viewpoint) by requiring all e-mail servers to be registered for proper accounting (nevermind freedom and privacy).

Update: This paragraph is rubbish, due to my inability to calculate simple things; I'll keep it here in the interest of historical accuracy (randomly changing already published log entries is impolite, to say the least). Anyone operating a busy mailing list will want to move it outside the EU. Even at such a low price, a list with a thousand subscribers and a hundred messages a day is going cause a tax burden of around four hundred euros a year. A company might be willing to pay that; a free software project would probably not.