> Actually, your suggestion is very nice IMHO. It's dependent on too many assumptions. And the colors are assigned only on basis of the font used, not on the semantics of the formula. [Not that we would really need this all of a sudden. In most of history, mathematicians have been able to make do with simple black-and-white equations. Now, on the other hand, colorizing *parts* of equations (or using colored boxes, see attached PDF) for didactic purposes could indeed be useful, but that would not mean an all-out attack but rather a selective insertion of \m-escapes by hand, which has been possible ever since \m was introduced (and before that, via device specials).]
> Why do you need \R? It's `eqn' which interprets $foobar$ -- > why can't it be translated to > > .nr eqn-font \n[.f] > .nr eqn-size \n[.s] > foobar > .ft \n[eqn-font] > .ps \n[eqn-size]u It's just that I had noticed that eqn inserts stuff like the following: .ds 0sfont \R'0sifont \\n[.f]' etc. (save font) .ds 0rfont \f[\\n[0sifont]] etc. (restore font) .as1 10 \&\E*[0sfont] -- lots of eqn stuff -- \E*[0rfont] \*(10 and had assumed that this would allow doing things like .ds xx "This is an equation: $x = sqrt ( 2 )$. .ps +10 \*(xx but this doesn't work as I wished it to.
coloreqn-2.pdf
Description: Adobe PDF document
