Hrvoje> char= is all nice, but then we'd also need char<, char>, Hrvoje> char<= etc. That sounds ugly. How does CL collate Hrvoje> characters? ecm> A character in Common Lisp has an associated code (which in ecm> most implementations is an ASCII value), and the #'char< ecm> ordering is consistent with #'< on characters' ecm> codes. Alphanumeric characters are guaranteed to respect a ecm> sensible partial ordering. True. And useless in a Mule world. CL is just not specific enough for the purposes of a multilingual text processor. Unless there are conventions beyond what's available in the hyperspec, we're on our own with this stuff. We can have an arbitrary default collation sequence that satisfies the minimal restrictions cltl2 puts on characters easily, just use the Unicode encoding or the Mule encoding. (IIRC with MINIMAL_TAGBITS you can just compare the internal representations as unsigneds.) The Unicode standard (2.1) specifies the Unicode order as default, so that's a reasonable way to go. But in the long run we want character collation to be dependent on language environment (IIRC the Scandinavian languages do not agree on the collation sequence for Latin-1, or maybe they agree on Latin-1 but disagree about whether they are interleaved with ASCII or all come after ASCII in ISO-8859-1), which means POSIX-like LC_COLLATE tables are the most straightforward way to implement. -- University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091 _________________ _________________ _________________ _________________ What are those straight lines for? "XEmacs rules."