At 10.32 +0100 2003-02-03, David Carlisle wrote:
>> Ideally I think that I'd like the latex field to consistently have
>> a command that could be used as latex's internal encoding independent
>> command, together with some latex packages to define any additional
>> commands needed, so you could switch at the latex level between
>> displaying the glyph, or faking it with TeX constructs or making a
>> missing-glyph marker, depending on the fonts available.
>That's just another way of saying the latex field should have LICR
>commands. For Latex use only it would be just as easy to define the
>mappings directly in the latex package files but having them in that XML
>would I think help translators from XML (especially MathML) to TeX to
>come up with consistent mappings, as all the MathML entities and
>character documentation is derived from there.
IMO some definition of the LICR commands which is distinct from the code in
the LaTeX packages is needed, since the current state of things is rather
implicit. It is possible to see that something works in all cases one can
think of, but it is not possible to see whether it should work in general.
Take for example \&: is this LICR, or is it a category 12 `&' token that is
in LICR, or is the LICR missing a \textampersand command? I don't know, but
I suspect the matter is still undecided.
Some documentation for the file under discussion might be a good place as
any to start with a more precise definition of the LICR.