[Ecls-list] Latest changes

Juan Jose Garcia-Ripoll juanjose.garciaripoll at googlemail.com
Sun Jan 11 11:41:42 UTC 2009

- ECL now uses a compacted Unicode database, compiled using SBCL's
tools, for upcasing, downcasing and querying basic properties.

- An important fix in the binary file loaded. Due to a typo, ECL was
unable to load a recompiled file.

Regarding the first change, the support for Unicode is very minimal.
Having the database in-memory (barely 92k) will allow other libraries
to benefit from the information, but tasks such as uppercasing,
lowercasing, etc, are not Unicode conformant and require additional

Indeed, any ANSI implementation CANNOT be Unicode conformant because
both standards contradict. The most important point is that the
character case is defined as an invertible transformation, which is
not the case in Unicode. There are characters that when uppercasing
become two, and viceversa, and Unicode understands the notion of
titlecase. String collation also differs. I presume a library like
cl-unicode should provide most of this, and it would be in our
interest to make this seamlessly integrate with ECL.


Instituto de Física Fundamental, CSIC
c/ Serrano, 113b, Madrid 28009 (Spain)

More information about the ecl-devel mailing list