frog hop

Geese!

Geese!

So Many Geese!

Huh. The people in charge of the facilities on Grouse Mountain (which is directly out my living-room window) appear to have placed a 1.5 megawatt wind turbine on top of the mountain.

They even put an observation deck on top of it. You can elevator up and physically stand in a wind turbine (incidentally overlooking one of the most beautiful views in the world).

That'll be some fun PR for wind energy during the olympics, I imagine.

I'm amused by the topic of hypercorrection. See if you recognize some!

After work yesterday I rode down to Richmond and met Frances at the east end of River road. We then rode back along it, turned south and rode down to Steveston, and home via UBC; total for me about 85km. Half in the dark, so an interesting ride. Coming across 38th (I think?) we went down a surprisingly steep and smooth hill, and just as I got up quite a bit of speed on the descent I met the nail pictured to the left here. An alarming time to get a puncture flat! Wham, thumpa-thumpa-thumpa-thumpa ...

Luckily Frances had spares and I had the pump I got in Portland, and the tube switch only took a few minutes. Coming down UBC on 4th was also a little harrowing as I was worried the bike would vibrate and try to throw me off again, but it did not. Guess the new rack has better weight distribution or something. Annoyingly, it doesn't fully hold the pannier on. Wrong-shaped stays at the bottom. Ed just cannot get the rear rack correct on this thing.

The rest of the week was quiet and pleasant, aside form the lost election. Some books on programming language history showed up. A number of bugs came and went. I rode the commuter around a few times, and got a bit back on track in terms of regular meditating.

Last week we did the pacific populaire (100km in-event, 130km total). I came in at 4:30 for the event, which is a good clip for me. An hour better than last year! It was a total zoo, hundreds of people, very crowded road for the first half or so, until the crowds spread out.

Today we rode up and down Burnaby Mountain a couple times (40km-ish). During the first descent, my bike became unstable. Wobbling, first just a bit but then quickly building up a serious thrashing resonance that threatened to throw me off the bike (at 60km/h). I went from "in control" to "handlebars attempting to spontaneously turn themselves sideways" in a matter of moments. Luckily the brakes were ok, so I brought it to as abrupt a halt as I could on such a hill, then kept it below 30 the rest of the descent. This was very scary. I think its rear rack is too long, and this makes it flex too much or something. I'll take it into the shop.

Otherwise a slow week. Read some more Soviet history. Working on a ridiculous multi-layered set of bugs in the spidermonkey decompiler and bytecode assembler. Trying to keep my meditation schedule up. Generally enjoying the transition to spring.

This weekend was 100% bike crazy. Saturday we rode through Surrey to the US border, then back up to Delta and around Boundary Bay, up through Ladner, over to Alex Fraser bridge, back to Arthur Lang bridge, then home. Today we rode up Cypress mountain then along the sea-to-sky Upper Levels Highway (you pedant) to Horseshoe Bay, and back home (too much heavy Sunday traffic).

Total distance about 245km. The hardest part was the 900m continuous climb up Cypress. I know Kris has no sympathy here at all.

Yesterday I had a chat about depression. It reminded me of one of the more interesting results I heard in a social psychology course I listened to this year. The result may excite you too. It is about "identity complexity". For a variety of theoretical reasons, some researchers set out to measure the number of ways a person views themselves. Contexts a person feels they "have an identity" in. So if you take a dance class, and you fly airplanes for a living, and you're a parent, and you do some political activism, that might be 4 reasonably distinct identity views.

They plugged the number of views and a somewhat arbitrary measurement of similarity between these views into an equation to come up with a measurement of a person's "identity complexity". If you have lots of identity views, or a few very different ones, you get a higher score. They then looked at other parts of life to try to correlate it.

What they found was the interesting part: the complexity measure is proportional to your emotional stability. The richer the network of different views you have on yourself, the less prone you are to either extremes of emotion: you're less easily made absurdly giddy and less easily made horribly depressed, as life's events turf you around.

Remembering this, and remembering that it's been 4 months since I did any pottery and I'm feeling a little down, I went to the pottery studio after work. Amazingly, I threw a very good 3lb bowl my first try back on a wheel. Using 4 month old clay, even. It may be the best, or second-best, piece I've done. Yay.

After I finished that I went to walk home, but encountered an odd thing on the sea wall. At the marina nearby, two sea otters -- big things, probably 50lbs each -- had climbed out of the ocean and were rolling around on the dock. At first they seemed merely to be goofing around, scratching their backs and enjoying the weather. Everyone stopped to look at the cute and fluffy otters! Daaaawww!

But then one hissed loud at the other and they commenced trying to tear each other to pieces. It was insanely violent. They thrashed around for several minutes, knocking things over and screeching at each other while biting necks and flanks, frequently producing unfortunate ripping sounds. I eventually tired of the blood-sport and walked on. So far as I know, they may have fought all night.

Mammals: they're sort of bad ass.

I want to write something about Doron Zeilberger, whose work I just discovered last month and who has finally -- after years of my struggling with it -- shown me a sensible way of thinking about symbolic expressions that do not denote finite objects. Namely: to affirm that "infinite" objects don't exist, but that several interesting families of integer-like finite expressions that happen to not canonicalize to finite integers do exist, and that they occupy interesting finite subspaces of the set of all finite symbolic strings, so they're worthy of study anyways.

Perhaps that's all I need to say about it. I fought with the issue of infinite things for a long time -- long enough that it's mostly the reason I left math during undergrad -- but now I feel like I'm at peace with it. My finitism bigotry feels reconciled with useful mathematics. Infinite things don't "exist" the same way finite things exist, but that doesn't make the expressions people use to refer to them useless. Yay.

After a leisurely sleep-in, wander and cappuccino, I set off to take advantage of the gorgeous weather today by riding out to horseshoe bay again. This is becoming a nice regular sort of practice route for me. Did it in about 130 minutes so even slightly faster than last time; but this time it was particularly notable for not being difficult. I even dropped a couple roadies along the way (admittedly sort of half-hearted sorts). Felt weird. Making it out there I sort of felt like I had ridden ... to the end of the block? I remember when this was a crazy far ride.

While I was stopped at horseshoe bay there was a gull that had just discovered a box containing someone's half-eaten lunch. Probably something good and greasy from Troll's. It actually found the box on top of a garbage pail but gave it a good whack to get it to fall off and spill on the ground. It was very excited about this discovery, and attracted many of its friends and competitors.

Things got ugly. Feathers flew.

The debian/openSSL case has been well covered, most recently by Russ Cox, but I have yet to see anyone raise what seems to me to be the central issue.

Most writing on this has talked about the inappropriate use of "fickle" tools, of developers running around "blindly fixing code they do not understand". That's not the point: nobody actually proposed a fix, and the people who "understood" the code did not appear the least concerned that it contained a UMR. The UMR was "intentional". In fact they had tried to hide it with #ifndef PURIFY. The fiasco is just that someone else messed up adjusting the wallpaper that was covering it.

Analogy: imagine you are an accountant auditing a big company, and they have two sets of books: the "real books" and the "books we give to auditors". This is the sort of situation you're looking at. It should set off serious alarms. The whole point of doing the diagnostics is to find the faults. A UMR is a bug, whether or not you're in a PRNG (perhaps even more so there).

Listen to your bug-killer tools. They are there to help.

Warning: only of interest to language nerds.

I have had a number of engaging conversations since beginning work on ES4, including many with our "visiting academics" Dave Herman and Cormac Flanagan. Recently Dave and I were discussing module systems. I told him I was impressed with the Owens and Flatt Unit system, and he replied that while it was pretty good he was most impressed by Dreyer and Rossberg's MixML system, which is slightly more recent. He's right, it's really amazingly good. It's still only in "submitted for publication" state, but it's implemented and extremely elegant. Tidy, usable, very powerful, subsumes nearly every other system, and yet is smaller and simpler than most of them.

The two "teams" are clearly competing with one another, and cross-pollinating every step of the way. Hurray for quality research!

00:00, May 13.
Achewood posts a comic showing wikipedia lying about Finland.


10:32, May 13.
Wikipedia begins to lie.


Minor edit-war follows.

13:55, May 13.
Wikieditor Wafulz semi-protects page.


All such edits are reverted.

When I was in elementary school, those few of us who were odd enough to like computers generally wanted to play video games. But to play a game on a machine back then meant talking, at least briefly, to a language interpreter. These were simple, simple machines: Sinclairs, Apple IIs, C-64s, etc. They booted into ROM-based language interpreters. You had to learn at least a smattering of BASIC or such to convince them to do anything, even load a game.

Along the way, some number of kids stopped to notice that the language interpreter could do more than just load games: it could, within the confines of its display and sound system, do anything you asked it to do. Kids started to script up little toys. At first these would be simple toys, like cycling the colors on the screen or making gurgling noises or such. But over time they became complex: many months and years of after school fiddling, some of those kids dropped beneath the language interpreter into the assembler, and then (as the machines got better and better) organized a little subculture of artist/programmers called demo coders. Demo work has always been a very pure vein of algorithmic, mechanical art; but it's also one that has fewer channels of entry into these days.

As the machines got better, they also got more complex, and their operating systems hid more and more of the machine beneath pretty abstractions that reduced users to menu-selectors, passive audiences. Language interpreters got pushed into terminal windows and professional IDEs, and in many cases these things were banished from normal users' vocabulary altogether. This is a shame, because the formative experience of interacting with a language processor -- particularly one hooked up to a live, imperative visual display -- is oddly compelling. It gives you a really visceral feeling of breathing life into a bit of inert text.

I wondered what the big deal was about the "Processing" language for a long time, largely ignored it as a sort of toy for multimedia artists. I took another look on John's recent port of it to javascript, and ... I think I understand it in a different way now, in an educational context. I think the point is to make a tool that focuses a non-programmer's attention on that basic, formative experience: put very minimal bits of text in window #1, and something visual, fluid, alive and magic will occur as output in window #2.

If you look at the sorts of things that people tend to do with Processing, they're usually a lot like demo code. Cycle the colors, spin the object, bounce and flicker, make noises. Some get quite complex, generating terrains and 3d animations and dreamy, moody work. Not many of the results are practical, but I imagine they're serving as the point of first contact for a lot of people with that fleeting, odd, powerful experience of saying something in program code, and having it leap to life in front of you.

I thought the web would be the place most young people learn to program these days. The web, after all, has its own good points: you can inspect everyone's code, and there's a ubiquitous universal access and distribution system. But I wonder if an environment like Processing will serve as a similar aid, owing to its highly engaging visual experience for the proto-coder. Hopefully John's port will help blur the distinction, taking the best of both worlds.

(Particularly *cough* if we bolt a dynamic JIT on the back of it and give it access to openGL hardware :)

I washed the marinoni tonight. It had a hairball in the rear derailleur. It was that bad. But now it looks like a shiny newborn.

Tomorrow there will be some cranking and such, so it needs to be in a good mood.

I was thinking of seeing standard operating procedure tonight because I am a big Morris fan, but the reviews are decidedly mixed, and this is a topic that (having been done to death by others) he has to do very well on to make it worthwhile. So I'm undecided. Anyone seen it? I figure it'll be worth an actual full-size screen if it's good, but is it?

Today I did the laptop-vagrant thing and migrated from coffee shop to coffee shop. Sushi lunch was delicious. It's fun to work this way about once every two weeks. The rest of the time an office is nicer.

Right now I am drinking local beer, listening to very bad early 90s techno, and considering what to make for dinner.

kiwano came to stay with us last week. We fed him, helped him get his bike ready, and tried to make him strong for the madness ahead: he's attempting to ride from here to Halifax.

( Some photos and commentary on the first leg )

We rode to Golden Ears Provincial Park today. It's about 60km outside of town (121km circuit, says Mr. Bike Computer). It was nice, but tiring, and my lungs sort of hurt. Frances installed a feed bag on my bike recently, so I ate an appropriate (and constant) amount this time. It certainly helps avoid glucose-depletion exhaustion.

Definitely less exhausting than the populaire. I guess I'm getting stronger.

I went for a nice ride today: out to horseshoe bay and back, 58km in 140 minutes (including a 15 minute break to eat and drink), so this is somewhere around 25 km/h avg. It's a very hilly route, so that was actually pretty a awesome speed for me: I sprinted up several hills. I'll surely feel it tomorrow. The weather was wonderful for it: I didn't need anything extra to keep me warm. The scenery was also pretty special.

The reason I took such a pace -- I was intending for a leisure ride -- was that I acquired an informal pace bunny just outside downtown and found myself wanting to keep up. So we leap-frogged a few times through west van and I finally settled into following him through the hills. Guy was a bit of a beast, but I guess that's what I should aim for! Alas, I had not cleaned or greased my chain since last weekend, and unfortunately it was quite muddy. So the bike squeaked and groaned like crazy while it was moving. I fixed it all up when I got home.

I've been listening to lectures from Berkeley recently, since they put a number of them online, a la MIT opencourseware (if you follow that link, be sure to twiddle the combo box selecting archived semesters in the upper right corner: there are 13 semesters worth archived, and they've different stuff. The podcasts dry up before 2005 though). I'm just finishing up an amusing one from 2006 called "drugs and the brain", a sort of survey thing on the history, uses, biology and psychology of all the major psychoactive things. It's pretty fun. Quite light-hearted -- the prof is casual and frequently jokey about the material -- but for a topic that's usually so taboo, I think the approach really works.

The rest of the day was chores. Tomorrow, more chores. Yet domestic life often makes me happy. It's usually easy, not nerve-wracking. Maybe some cooking or such, or I might drop by the mozilla camp in the afternoon.

I did my longest ride ever today, which was only about 120km door-to-door, but still. Big step for me. I haven't done anything that size since the one last summer that ended with my face torn open.

Took 5:22, which is slow. I had a bad cold, 6 hours of sleep, was high on pseudoephedrine, and it was cold and raining for the last third of it. But mainly it took so long because I didn't ride during the winter, and I'm a bit out of shape. Oh well. Next time will be easier and faster. Despite really hurting near the end, this one felt much shorter than the canada day one.

The scenery was nice, at points. Spanish banks and river road at least. And it was nice to ride with a mob for the first while.

I am now going to fall down.

Three years ago I abandoned a bicycle in Cologne, partly because it was a crappy bike in the first place and partly because the pedals fell off.

Over the past 3 months I have taken a course that taught me to put a bike together "from the parts up", including all the bearings, brackets, cables, spokes and such. I was riding it home to give it to Jack -- I don't actually need another bike -- when I realized that I was feeling the same thing I felt back in Cologne. Only this time I know what it is: a loose bottom bracket. I stopped at the nearest bike shop, asked for a stand, a hook spanner and a crescent wrench, and fixed it. While I was there I also grabbed a cable puller and tightened the rear brake, that I'd noticed was a little loose.

It feels a lot better to be in charge of the machine I'm riding.

This week I shifted back to working on the ES4 RI. I deleted a thousand or so lines of code, which was fun. I have my eyes on quite a bit more. I also seem to be sliding back into the travel obligation: by mid April, depending on the luck of the border, I will probably be in Boston again. I haven't been there in a year. Let me know if you want to squeeze in dinner or something. Or breakfast at, oh lord, 4am on the way to the airport?

Now that I'm done building the bike I will probably shift back to doing some more pottery, although there's a weekly volunteer bike-repair night on Tuesdays that I might fall into instead. I enjoy the social atmosphere of a bike shop. It's mostly solitary, but there's a little sideways chit-chat. More than in anything else I do, at least. And I enjoy working with my hands. I enjoy cooking, for that matter! This has been a year of muck: clay, food and bike grease all over my hands. Muck and tools feel more real than keyboards and mice, pens and paper.

I have many other things I could note here -- it's been a busy few weeks -- but I'm sort of tired and would prefer to use the next 30 minutes meditating before bed. FYI: it helps turn ok sleep into fantastic sleep.