Friday, 18 October 2002

Unicode will eat your brain

Listening to:

Vaughan-Williams, symphony no. 8 in D minor.

Once upon a time, deciding that this log shouldn't just be me wittering on about trivialities, interspersed with book reviews, I tried to start a discussion of something fairly technical and work related. (Others do this quite successfully. I've recently become a reader of Raph Levien's online diary, where Raph talks all about proof systems and other stuff near and dear to my heart.) Sadly, I never got beyond Part II of my discussion on higher order matching. But why? Well, now the truth can be told.

It's all the fault of standards zealots in the land of fonts. In order to discuss higher order matching, I needed to be able to write the Greek symbol lambda, and I couldn't get this working under Netscape 4 (and it doesn't work under Netscape 6 either). Yes, my maths would look OK under Internet Explorer, but I didn't (and don't) have such easy access to that browser. For all that 90% of my readers would see my lambda's OK, I wanted to see them too. But it seems that the device whereby the mathematical symbols are made available for Explorer is not standard, and Netscape doesn't support it. We should be using Unicode instead apparently. (A comprehensible explanation of the problem, a cute illustrative page.)

It seems to me that people who get involved in the worlds of computer fonts quickly lose all of their previous usefulness in other fields. Computer fonts suck them in, chew them up and they're never the same again. My attitude is one of wilful ignorance; I just don't want to have to know about three levels of abstraction, and vocabulary like glyph and charset.