Frank Mittelbach <[log in to unmask]> writes:
> but isn't this ambiguity already present anyway in TeX? and those interfaces
> have to deal with it?
> so from that perspective I see no argument against having the possibility to
> use keys generating 8bit characters to be used in text and math producing
> different glyphs.
I didn't explain clearly. What we have now is that
(a) For people with GUI interface like Scientific Word, the software has
to deal with the ambiguity as best it can.
(b) For people with non-GUI interface like vi, writing a Greek letter
(for example) outside of math leads to a TeX error at run-time.
If the 8-bit latin-1 characters that look like \times and \mu are made
to work equally well in text as in math, then doesn't it seem likely
that the people in case (b) are going to make more markup errors
(writing \mu etc as a text symbol when it should be wrapped in a math
formula) because there is nothing to bring it to their attention (other
than reading the documentation :-) ?
In other words currently it is easy for people in case (a) to make those
markup errors, and the proposal under discussion would extend this so
that it is easier for all users. It's not clear that this is progress.
I guess I don't have a solution, I am just reflecting on the idea that
the restriction of mathchardef'd symbols to math mode was intended (I
think) by Knuth to bring markup mistakes to the user's attention.
Instead of an error, he could have arranged that \gamma triggers
start-math, just as letter X triggers a switch from vmode to hmode.
To be sure, because we have \ensuremath it is already possible for
people who read the documentation to inconsistently omit math markup all
over the place anyway. But it has always seemed to me that it might be
better to tell the editor about the "ensuremath" objects and have the
editor insert $ $ automatically when necessary rather than have LaTeX do
it by using \ensuremath.
If anyone is not sufficiently confused enough yet I invite them to
consider how LaTeX might better handle "<" and ">" when someone uses
them in text. In OT1 encoding LaTeX silently gives the wrong symbol, in
T1 you get the math less and greater (when what the user wanted was
probably langle rangle or else guillemets).