John Cowan wrote: > Since "large" is not quantified, I can't call this false, but I > believe it to be misleading. Without the (rnrs unicode (6)) library, > no tables whatsoever are required.
That's not quite true, because most programs are likely to import at least one of the (rnrs io simple) and (rnrs io ports) libraries, which export read and get-datum, which are required to implement the lexical syntax, which is defined in terms of Unicode general categories. At the very least, the io libraries need a small table for Unicode general categories. In the reference implementation of (rnrs unicode) that Mike Sperber and I wrote, those tables occupy a little over 5000 bytes. > The ICU tables that would be > required to implement R6RS as of the current version of Unicode (5.1), > namely casing + normalization + general properties (which also provides > additional information not required by the standard, such as a mapping > from non-European digits to their numeric values) are 213KB in total. In the reference implementation, all of the Unicode tables add up to a little over 85000 bytes (on a 32-bit machine). Will _______________________________________________ r6rs-discuss mailing list [email protected] http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
