🚀 go-pugleaf

In article <1993Jan8.081830.15294@eecs.nwu.edu> hpa@eecs.nwu.edu (H. Peter Anvin N9ITP) writes:

>I'd suggest some form of multiple locales, i.e. 
>setenv LANG ukranian-swedish-ipa-greek-farsi-japanese

>In this model, each of the sorting algorithms would assign sorting
>values (integers, that may or may not be identical) to each character
>recognized in that language.  Any character undefined (e.g. KANJI HITO
>for Ukranian) is assigned MAXINT and thus is sorted last.  The sort
>proceeds with the first language as the primary key, second language
>as the secondary etc.  That means not only that the text in the
>different languages will show up in approximately the order listed,
>but also that multiscript languages can be easily accommodated.

I like this idea.  Can anybody see any serious problems with it?
(Of course the description above cuts some corners, like multiple
letters sorting as one, but I can't think of any such problem
that couldn't be handled in this scheme.)
--
Tapani Tarvainen  (tt@math.jyu.fi, tarvainen@finjyu.bitnet)