Jeni's Musings

I do sometimes wonder if we’re the same person:-)

most (or some, depending on how you count) of the data in UnicodeData.txt has been available as “unicode.xml” for years (er, decades nearly:-) First in Sebastian’s jadetex support for dsssl then in the MathML sources, and most recently in the sources for the entity set draft at http://www.w3.org/2003/entities/xml/unicode.xml

As part of the build up to the the MathML 3 drafts I recently needed to update those to Unicode 5 (which has some new characters specifically to support the entity sets) and so I got fed up of only having “most” of UnicodeData.txt and so I “put together an XSLT 2.0 stylesheet” which added all the missing info.

The draft is currently w3c-member only but only really as that’s a convenient cvs archive until ready to update the public version. If you (or your readers) have w3c access they may like to compare with http://www.w3.org/Math/Group/spec/xml/unicode.xml (the http-view of that is a client side stylesheet which doesn’t show all the information in the file it does have all of unicodedata.txt plus a pile of other stuff about entity names and names in mathematica, affii glyph register and TeX etc. I really should get that updated to a public part of the site…

Although actually while having unicodedata as xml is clearly a good thing, isn’t the approved way of solving the stated example problem to use a collation to do the comparison, which probably needs to be written in some less interesting language than XSLT, and may not benefit so much from an xml version?

David