corollary

I would have written an entry on Wednesday this week, but I found that I couldn't connect to the Cambridge computers that host the pages. Thwarted in this way, I felt that my brother’s offer to host my log on the machine where he has his own log, Archæology, was particularly timely. Jamie also said that I could have my own one-word name for my log, and thus be located at myname.artefact.org.nz. Cool! With his domain called artefact.org.nz, I discerned a definite trend in his naming scheme and briefly considered artifice (as well as artichoke). I think I’ve settled on corollary however. I may even go so far as to capitalise it.

I’ll need to convert all of my old log entries to the format expected by the code that Jamie runs, so I don’t know when exactly it will happen, but I hope to get it done pretty soon. I will have to make sure that I can write appropriately standards-conforming HTML. It will be nice to have slick feedback mechanisms available (either mailto forms just don’t work for anyone, or I have particularly unadventurous readers; go on, type something into the box!)

In the meantime, all (or at least some) of the links fit to print:

• A cute story about English ex-patriates bringing up small children in Paris.
• Last week, you may recall that I had Friday off. This is because it was ANZAC day, the equivalent of Veterans' Day in the US, and Remembrance Sunday in the UK. This was the first ANZAC day I'd had in 9 years. One thing I didn't know is that the traditional Australian gambling game Two Up becomes generally legal (as opposed to only allowed in casinos) on ANZAC Day here. How odd. The Sydney Morning Herald explains all; a general explanation of how to play.

I am pleased to report that the log has followed me to the Southern Hemisphere, and is now residing (in a closet) in my beautiful home-town, Wellington, New Zealand. Many thanks to Jamie for helping me set it up.

I kicked the bed-leg this morning, and my toe came out worse. In fact, the bed-leg wasn't even scratched. I even dropped my original plan of walking to work, and cycled instead. So, here I am, feeling a little sorry for myself, but at least I didn't have to saw my own arm off.

My toe is rather an ugly purple colour on top, but my gait is less of a hobble than it was on Monday, so I think things are on the mend.

Here's a marker of cultural difference between the UK and Australia (other than: one can play cricket, and the other can't): in the UK you can buy orange and grapefruit juice blends in cartons. This is typically known as ‘breakfast juice’. In Australia, you can't (I have seen ‘breakfast juices’, but these are typically orange with mango and/or pineapple). But Australia distinguishes itself by providing apple and blackcurrant juice, which I don't remember seeing on the shelves at Sainsbury's. Australia's better because you can mimic the UK by buying the orange and grapefruit juices separately, and mixing them yourself, but you can't buy blackcurrant juice separately in either country. Isn't that fascinating?

There's a guy out there called Paul Graham who is definitely something of a guru. I earlier linked to a piece of his about using Bayesian analyses to detect spams (given samples of non-spam and spam messages). For whatever reason, his proposal hit a nerve, and lots of people rushed to implement this (good) idea. I am using Xavier Leroy's SpamOracle implementation of this idea, and it works wonderfully.

On the other hand, Graham's latest piece is all about hacking and how it's really just like painting, and it gets up my nose in all sorts of ways. Perhaps this is inevitable. He says

I've never liked the term "computer science." The main reason I don't like it is that there's no such thing. Computer science is a grab bag of tenuously related areas thrown together by an accident of history, like Yugoslavia.

and I just don't buy this. Yes, there's a spectrum of interests in a Computer Science department, but this is true of any discipline too. Graham's also got a real hang-up about what he calls static typing, and he proffers Lisp's dynamic typing as a real win over languages where you have to declare the type of every variable (from another essay of his). SML and Haskell are both statically typed languages, and neither requires you to declare the types of variables.

Of course, both of these languages are academic languages (i.e., the product of academic research), and academia is generally getting a bum rap here so perhaps it's not surprising that I should feel so antsy. It's all symptomatic of a typical mentality that says that hacking is cool, and where the real work is done, while stuff done in ivory towers is obviously not useful, or applicable to the practice of real programming. In order to demonstrate that they absolutely don't suffer from maths envy, people of this view find it too easy to lurch to the opposite extreme and dismiss academic stuff out of hand.

Seethe, seethe.

Jamie recently bemoaned the demise of the generalist, saying that

The world seems to have moved on from this era when it was possible for an individual to have great knowledge in many disciplines. Specialisation is now the accepted approach, ostensibly because there is so much to know that only through singleminded devotion to a single area of knowledge may any advance be made. I am not convinced that this is true.

It boils down to what you think qualifies as an advance. If you want to push back the frontiers of human knowledge, then you have to get to the frontiers first, and there really is too much distance to cover between blissful ignorance and the frontier to be able to do much more than travel quite a narrow path. What humanity knows is increasing in leaps and bounds all the time, and as more and more people come to work in research too, this problem is going to get worse. It is a problem, something we can regret, because, though I love my ‘specialist subject’, I would like to also know lots of stuff in other areas. Not for nothing did I do a BA in history as an undergraduate as well as my BSc in computer science. And now the route to great knowledge in many disciplines is to devote my life to a great deal of reading.

On the other hand, if making an advance means advancing human society some other way, perhaps through applying all of this knowledge that researchers are busy creating, then Jamie's right; specialisation is not the only route to being useful. A generalist might be just the right person to draw several strands of specialist knowledge together and see how to apply them. (To be provocative, people like this tend to be given titles such as ‘manager’ and ‘venture capitalist’.)

Finally, it's worth noting the related area of specialist skills. When my toilet breaks, do I ring a generalist builder or a plumber to fix it? When I have go to hospital for a heart-bypass, do I get my GP to perform the operation? To end with a homily, everyone has their place, and specialists are important too. Awww.

On Friday, I lived up to my web-log title by actually proving a corollary: number 11.4.8 from Barendregt's The lambda calculus, which states that M has a head normal form iff M's head reduction path terminates. Ironically, I reckon this result isn't a corollary of the standardisation theorem at all, as Barendregt claims: my proof only needed lemmas 11.4.3 (i) and (ii), and lemma 11.4.6. Therein lies the dubious virtue of replaying all this stuff in the mechanical theorem-prover.

In other news, I have recently been impressed by a series of web-log entries at Public Address, a site that hosts six web-loggers, including Russell Brown, who writes cynical political analysis that I almost always agree with. These links are all by others though:

• Damian Christie on the NZ government's awful decision to throw money at the America's Cup campaign.
• Jolisa Gracewood on finding somewhere nice to live in America that isn't New York city.
• Chad Taylor on X-Men 2, being entertained without thinking that it was the greatest thing since sliced bread. (I'm off to see the Matrix sequel this evening; it's hard to know which is more off-putting, this negative review, or this off-puttingly gushing one from Salon.)
• Debra Daley on her top tip for dealing with teenage children.