[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Cyrillic in math

> use of "native languages" and therefore native glyphs in math formulas
> is common especially in textbooks of physics and engineering.
> (example: feynman lectures in physics) 
> one could write for example $\int_{\rm all\ space} ...$ or
> $\rho=\rho_{\rm free}+\rho_{\rm bound}$ etc. clearly when such a
> book gets translated to other languages, these words will also
> be translated and hence the use of native glyphs in math will be necessary.
> this issue also underscores the inefficiencies arising from the poor
> design considerations that underlie the new math font encodings
> ---necessitated by the use of Cork encoding for text (missing greeks!).
> hence, one would need to assign a family for *text* roman, in addition
> to the math roman family, not to mention the family that will hold the
> default math characters, among them numbers from the \rm font!

No, what your example shows is the poor design of the plain TeX way of
writing the bits of text `free' and `bound' directly in math mode.
They should better be done in text mode, not math mode, by putting
them into an \hbox (and a \mathchoice, as in the \text macro of
AMSTeX). Consider what happens if you write {\rm quasi-free} in math
mode, as opposed to \text{\rm quasi-free}. The surprise to the user of
little problems like this seems to me about the same as the \bf\gamma
surprise that you wrote about earlier.

Writing text in text mode, no extra families are needed for
text-encoded fonts, and in fact more families are freed up, because it
is no longer necessary to allocate math families for \sl, \it, \tt,
\sc and the other nonmath fonts.

Michael Downes                              mjd@math.ams.org (Internet)