On Sat, 21 Jun 1997, Vladimir Volovich wrote:

> > Inputenc currently enforces all characters to be declared as text or
> > math, but not both. (Allowing use of both in general is very expensive
> > as normally in text mode it uses \chardef, but \chardef'ed tokens
> > don't work in math mode, so it would require twice as many csnames, so
> > that some math-mode version could be declared as well (cf
> > \textsterling and \mathsterling, and \pounds which uses one or the other)
> > However it may be possible to do something in this case.
>
> \chardef\newCYRA=128
> \chardef\newCYRB=129
> \chardef\newCYRV=130
>
> $\cyrmathit{\newCYRA\newCYRB\newCYRV}$  % this works well
>
> $\cyrmathit{\CYRA\CYRB\CYRV}$           % this produces warnings
>                                         % and does not work
>
> Probably, either the definition of \DeclareTextSymbol should be changed,
> or the new \Declare* command should be added to avoid this `bug'.

I second this. What about the following:

  \DeclareMathTextSymbol{<name>}
                        {<encoding>}
                        {<math character class>}
                        {<slot>}

e.g.

  \DeclareMathTextSymbol{\CYRA}{T2}{cyrletters}{"0C0}

the math category would always be \mathalpha (is this reasonable?)

\CYRA should then work in text and math mode...this is more like an alias
for a letter than a symbolic name IMHO.

LaTeX does this implicitly: you have a \DeclareMathSymbol{a}... but no
corresponding sequence for text mode!


    Werner