All posts made the week commencing Sunday 26 Oct., 2003:

George Lakoff tells how conservatives use language to dominate politics: "Language always comes with what is called 'framing.' Every word is defined relative to a conceptual framework. If you have something like 'revolt,' that implies a population that is being ruled unfairly, or assumes it is being ruled unfairly, and that they are throwing off their rulers, which would be considered a good thing. That's a frame. If you then add the word 'voter' in front of 'revolt,' you get a metaphorical meaning saying that the voters are the oppressed people, the governor is the oppressive ruler, that they have ousted him and this is a good thing and all things are good now." The interview goes into the differences between the conservative and progressive worldviews, and why one puts effort into defining the language frameworks and the other doesn't. Good read.

Nasa's Goddard Space Flight Center has some incredible satellilite photography of the fires in southern California [via scribot], at impressively high-resolution too. Those photos are from October 29th; there are more for the 24th, 27th and 28th.

netObjects is a "collection of everyday objects for the home that present real time information from the web." (They all look fantastic, but my favourites are -Gossip and -Umbrella, for the interface design, and -Peep cos it's funny.)

I went to buy some new trainers the other day, and did, in fact [or in actual fact: I didn't], and wearing these trainers I walked over the anti-slip pattern by the traffic-lights crossing Oxford Street from the north to south -- and actually I can't remember exactly where it was, possibly near where C&A used to be, but that's not important, what is, is this: the combination of the bumps of the anti-slip overlaid (underlaid!) on the modern pseudo-random bumps that are the tread of my shoes spelt out a message in braille on the bottom of my feet: the message directed me to such-and-such portion of the night sky at such-and-such a time and date, with blah magnification.

And so you know what, I did that, thanks to my new handheld GPS and some free binoculars I picked up at a server farm launch party at the tail end of the dot-com goldrush, and there was another message, also in braille: The Lord's Prayer, written nearby M31, in nebula and stars and the glints of satellites, and actually for one point, the taillight of a 'plane. You won't be able to see it now, a miracle!

Actually, that reminds me of another story, of before Transport for London when the tubes lived underground like moles, before we sunk the escalators and elevator-terriers, when the tracks were fresh, and the trains ran on time, and they painted beautiful mandelbrots with chalk and clay, and made love in the dark, and if you stood under where Centre Point is now - in that hub that we're digging up and trashing - and faced east, you could feel the sea air on your face, the wind from Southend-on-Sea.

The trains used to call to one another down the foot tunnels, their clacking voices whispering poetry as they passed, seducing, romancing in the 30 seconds before their doors would have to close, sometimes dwelling for up to a minute and - if they were lucky - a signal would get stuck on red and they would enjoy sometimes a quarter of an hour more in muttered sweet nothings, voices zigzagging down the corridors like light down fibre optics, or transmitted in the changing arrangement of the wall-tiles, every second more beautiful than the last thanks to the growing intimacy, but more painful because the end must be getting closer.

At Green Park once, long before it was called Green Park, the love to end all loves came to be between a train on the Piccadilly Line and a train on Jubilee. This was in the middle of an age of irony, and shone like a diamond, even brighter for being surrounded by aging tracks and scuffed seats. A chance broken rail meant these two lovers were granted hours together, and as they finally broke through the lack of earnestness that characterised those years, opened up to one another, they admitted with full frankness and not caring whether it was reciprocated, because this was so true, the truest most pure thing there had ever been, etc, they sung down the foot tunnel, simultaneously

I love you

and these two messages met in the middle of the tunnel-conduit and set up a standing wave, and in a catastrophe this wave crystallised and became fixed, lodged in the walls, and whereas before the decoration of the walkways acted as a substrate for encoded signals to zoom along at a much greater group velocity, these messages become solid, like concrete: and this fixedness spread through the tunnels and the trains and the tracks and the interconnects and the ticket barriers like a virus, and it killed them all.

So that's where we find ourselves now, travelling in the fossil shells of an ancient civilisation, trotting down lava tunnels where the lava has gone, and that's why when we tapped our first boreholes down through London we didn't hear singing.

But if you walk at Green Park down that walkway, between the Jubilee and Piccadilly lines, look at the intricate tile pattern on the wall, blues and whites and turquoise, and know what it really says, in the language of the London Underground.

And, as it happens, if you decode it into ASCII instead, into English, the tunnel walls spell out the whole of Beowulf too. But that's just coincidence. It doesn't mean anything.

Actually, Dunbar's Co-evolution of neocortex size, group size and language in humans is worth a good read -- there's more in it than the magic number 150 I referred to yesterday. The argument goes roughly like this: the maximum cohesive social group of 150 comes from the size of the primary network - intense friends - which is itself correlated with the neocortex size (in relation to the brain) and the proportion of time spent on social grooming. These three figures seem well matched across different primates.

However, if you look at humans we maintain a group size (of about 150) which matches our neocortex size but which is way over the time budget we allow for it (we'd need to spend about 42% of our time on one-on-one grooming, but it doesn't seem we do).

Dunbar contends that humans evolved vocal grooming (language) as a more efficient form of bonding. Assuming that our closest ancester, the chimpanzee, has hit the time budget limiting factor, and that our extra efficiency has all come about with the transition to vocal grooming, this means language is 2.8 times more efficient for bonding than the mechanism nonhuman primates use. That is, "a speaker should be able to interact with 2.8 times as many other individuals as a groomer can. Since the number of grooming partners is necessarily limited to one, this means that the limit on the number of listeners should be about 2.8. In other words, human conversation group sizes should be limited to about 3.8 in size (one speaker plus 2.8 listeners)." (Which makes sense if you think about the different qualities of a conversation with three versus four participants. A study is quoted to back this up.)

And for supporting evidence, showing that disparate human characteristics are coherent with this: "It turns out that there is, in fact, a psycho-physical limit on the size of conversation groups. Due to the rate at which speech attenuates with the distance between speaker and hearer under normal ambient noise levels, there is a physical limit on the number of individuals that can effectively take part in a conversation. Sommer (1961), for example, found that a nose-to-nose distance of 1.7m was the upper limit for comfortable conversation in dyadic groups; this would yield a maximum conversation group size of five individuals with a shoulder-to-shoulder spacing of 0.5m between adjacent individuals standing around the circumference of a circle."

All extremely amusing.

Two things:

  1. In the paper Co-evolution of neocortex size, group size and language in humans, Robin Dunbar predicts that the maximum group size that humans can maintain as a cohesive social unit, based on the ratio of neocortex volume to brain volume, is 147.8 (100.2-231.1 at 95% confidence). Consulting the literature, he finds that there's a trimodal distribution of group sizes: bands at 30-50 people, tribes at 1000-2000, and an intermediate one. The mean size of the intermediate level group societies is 148.4.
  2. The AOL Instant Messenger servers impose a hard limit on the number of people you're allowed to put in your buddylist: 150.

(For more, and a better summary of Dunbar's paper, read The Magic of 150. Malcolm Gladwell also refers to the number 150 in his book The Tipping Point.)

Some Mac OS X 10.3 ("Panther!!!1!") notes:

  • The MySQL 4.0 package for Mac OS X 10.2 installs fine, as does the startup item for it.
  • For Perl interfaces to MySQL, DBI installs fine but there's a bug in the system version of Perl that prevents going any further. Make the simple (ish) fix, then use cpan to install Bundle::Msql and choose the MySQL items and not the Msql ones. Or at least: you could do otherwise, but for some reason (despite loads of errors printed on screen) that was the only way I got it all to work.
  • The Fink package management system has already updated for 10.3, but I'm not using that for a while: I'm trying DarwinPorts (instructions also here) which is smaller but comes with a graphical Ports Manager (and is nicely BSDish). It built expat fine (but don't forget to add "export PATH=$PATH:/opt/local/bin" to your .profile).
  • I haven't had a play yet, but the Python bindings to CoreGraphics look really interesting. In the Mac OS X Unix Users technology brief [pdf], it mentions command line scripts to render html to bitmap (which otherwise is hard).
  • I was going to say iChat was finally AppleScriptable, but it looks like although all the stubs are there, the return values aren't useful (but I may be doing something wrong).

This feels like a good release, even if there are a whole load of scrappy UI decisions. It's close enough to 10.2 that most of the weird hacks still work and there's not much more to learn. And it runs well on my aging iBook, which is a pleasant surprise.