I generally recommend against using non-ASCII characters in Racket
source code.
Besides the human readability question (and the searchability question),
non-ASCII still presents a significant risk of your characters getting
corrupted "in transmission". Especially with all the Web sites and
other network services through which code might be presented and
copy&pasted. It's still not uncommon for a site/service to send
non-ASCII characters under the wrong encoding labeling, and you'll even
see services that map problematic Unicode characters (e.g., special
spaces, dashes, quotes) to ASCIII.
One forgivable exception to ASCII-only is if you want to be cute in code
examples in an academic paper. Although you can still run into
transmission problems, when people copy&paste code, and when your PDFs
are translated to a different format.
Then there is the cute lambda character, which has become a special
case, since it's been adopted by some Racket developers. I think some
of the appeal of that might go away if there were a more terse lambda
syntax. Maybe just make `fn` an alias for `lambda`. (nd, as an old
programmer friend (who was generally impressed when I explained Scheme's
first-class closures to him) remarked, "They actually type out the word
'lambda'? That's *so precious*." I've actually grown to find typing
out `lambda` to be tolerable, but if people are going to use a more
terse alias, `fn` is not problematic like the lambda character.
Neil V.
____________________
Racket Users list:
http://lists.racket-lang.org/users
