...

It may seem so to western europeans, but others may differ. A case could be made for UTF-8.

I tend to agree. Further, the choice of defaulting to Latin-1 seems a strange one when much of the rest of the world (well, the networking world) seems to be moving towards more universal character set encodings.

We didn't choose Latin-1 over UTF-8: the current situation just reflects the fact that we haven't implemented UTF-8 yet. It's not trivial to expand GHC's current Alex specification to handle the whole Unicode character set. See this message for some musings: http://www.haskell.org/pipermail/glasgow-haskell-users/2003-December/006 094.html Cheers, Simon