3D logo

We recently got a MakerBot, a 3D printer, at work. We test printed flexible snakes. Now one of the York CS research students, Gary Plumbridge, has made a neat video of it printing a 3D logo for the York Doctoral Symposium.

Philosophy, the great attractor

Wednesday's xkcd had the intriguing hovertext:

Wikipedia trivia: if you take any article, click on the first link in the article text not in parentheses or italics, and then repeat, you will eventually end up at "Philosophy".

So I tried it (as, apparently, did many other people at work that day, judging by the emails buzzing around).

Starting with "Spark plug" (the content of the xkcd cartoon), I got to philosophy. Starting with "Coffee" (which I was drinking at the time), I got to Philosophy. Starting with "Philosophy", I got (back) to "Philosophy". I also started with a couple of random pages, and got to "Philosophy".

Given the number of wikipedia pages, that's not much data in support of the Munroe Conjecture, but enough to blog about...

So I quickly hacked together a graphviz picture of the paths taken by my explorations, where each node represents a wikipedia page, and each edge represents the first link in that page.

First observation: If the Munroe Conjecture is true, the whole of wikipedia forms a single connected graph -- has a single "basin of attraction" -- like this picture, but with many many many routes, all eventually leading to "Philosophy".

Second observation: The "attractor cycle" at the base of all these paths contains "Philosophy", but also contains other pages, including "Science" and "Mathematics". So the conjecture could equally well mention any of these other pages.

Third observation: I understand that the number of hits on the wikipedia Philosophy page has increased dramatically since Wednesday :-)

---

UPDATE (12 Nov 2011): Others have put a bit more work into illustrating this.

mangling language

I'm now inured to things like having to press the "start" button to stop Windows. But even so I boggled slightly at the message I got when I last upgraded Firefox:

"Click Finish to continue starting Firefox"

Wait, what?

owl head stability

Yesterday I was at an interesting seminar given by Dr Tom Stafford (University of Sheffield) on "Inferring cognitive architectures from high-resolution behavioural data" -- about some interesting new results from psychological experiments on learning motor skills. Apparently it pays to be a bit rubbish at the start of learning, maybe because you're exploring more of the space.

Tom also recounted a great little anecdote about how Edward Thorndike, a late 19th century experimenter, thought the best psychological test subject was "a hungry cat": he repeatedly put the cat in a small box that had a foot pedal to release the door, and timed how long it took the cat to escape. Unsurprisingly, the cat got faster. More surprisingly, when the setup was changed to remove the foot pedal, the cat still pressed where it had been -- because it had learned where to press (or maybe it was going, "where's that fsking foot pedal gone?")

During the post-presentation conversation, the discussion turned to visual perception being hard wired for upright heads, to animals turning their heads upright even when they were upside down, to chickens keeping their heads stationary when their bodies are moved, to owls doing the same, to (inevitably) "there's a YouTube video of that!" A few minuted googling showed that indeed, there is. The bored-looking owl even keeps its head stationary when blindfolded. (The guy doing the waving about looks pretty grim, too.)

visualising algorithms through dance

I'm nowhere near the first to post about this, but I have to advertise this. It is a wonderful set of folk dances, from AlgoRythmics, performed by a dance troupe in association with Sapientia University, Romania, that animate six different sorting algorithms. Sheer genius.

They have produced YouTube videos showing: bubble-sort via Hungarian folk dance, insert sort via Romanian folk dance, select sort via Gypsy folk dance, shell sort via another Hungarian folk dance, merge sort via German folk dance, and, almost unbelievably, quick sort with yet another Hungarian Folk dance (using hats as index pointers!)

not shopping

I hate shopping for clothes. (Books are a very different matter.)

I hate it so much that my current clothes have to be literally falling to pieces before I will rouse myself to go hunter-gathering for more. Not only that, I'm picky. I only like certain colours (nothing bright, nothing pastel -- only browns, green, yellows, and other earth colours) and plain styles.

This Christmas vacation, I realised all my jumpers had holes in them, so I needed a new one. I grumbled out into the New Year sales, and found that Cambridge's Edinburgh Woollen Mill had a sale on. Not only plain, earth-coloured jumpers, but at half price! I used a technique my other half taught me (he's even pickier than me: he'll only buy in blue or grey): buy two. This means you have to shop only half as often. I've broadened the approach: buy four, and you have to shop only a quarter as often (and it's a great incentive to stay the same size). Four jumpers later, I sigh in relief -- no more jumper shopping for ages.

Now, this multiple purchase techniques doesn't always work. Shops often have only one or maybe two items of a given style and size. But, with the advent of the Web, this is no longer a problem. Buy the one successfully hunted down, go home, log onto the shop's website, type in the product code, buy several more, get them delivered to the door. I've used this technique successfully, too.

Surprisingly, though, even this technique does not always work. When next my shoes fell apart (hole in the sole, one of the two standard failure modes), I needed another pair. Off to Clarks, where I found a comfy pair, eventually. So, I thought I'd use the same technique. (I find buying shoes even worse than buying other clothes, for some reason.) Off to the Clarks website: try to order three more pairs. It won't let me! I can order pairs in other colours or styles: I don't want other colours or styles. I can't order three identical pairs. I fire off an email to the company, pointing this out. "Yes", they say, "the site doesn't allow you to buy three identical pairs." Yes, I know; that's the problem. But why won't it let me? The only suggestion I have heard is that companies don't want you ordering a whole range of styles or colours, trying them all, then sending back all but the one you like. Okay, that might make a sort of sense. Except that this site would have let me buy three different styles; it's buying identical things that's the problem!

Oh well.