On Nov 14, 2006, at 2:23 PM, John Cowan wrote:
> Unicode does provide rules for case-folding strings (which is not
> quite the same as converting them to lower case). The table is at
> http://www.unicode.org/Public/5.0.0/ucd/CaseFolding.txt .
Excellent. Here's an excerpt:
# This file is a supplement to the UnicodeData file.
# It provides a case folding mapping generated from the Unicode
Character Database.
# If all characters are mapped according to the full mapping
below, then
# case differences (according to UnicodeData.txt and
SpecialCasing.txt)
# are eliminated.
#
# The data supports both implementations that require simple case
foldings
# (where string lengths don't change), and implementations that
allow full case folding
# (where string lengths may grow). Note that where they can be
supported, the
# full case foldings are superior: for example, they allow "MASSE"
and "Ma?e" to match.
If I read this correctly, Scheme could easily support the "simple
case foldings" described there. Supporting "full case folding," on
the other hand, would not be possible without locale information
specifying how to interpret the file being read. For example, the
conversion listed for "MASSE" above is correct for German, but
incorrect for English.
I'd love it if we supported just simple case folding as specified in
that document. It would be easy to implement that and the case-
preserving reader macro mentioned earlier, thus supporting special
uses for case-sensitive symbols, e.g. XML and foreign-function
interfaces.
Editors, please do consider this.
Received on Tue Nov 14 2006 - 17:34:59 UTC