> it still needs to be `realworld tested'.
Clearly it fails that test today, but who knows what tomorrow will
bring.
> I am not trying to knock it
> MathML and Sebastian's ideas of semantic markup cater very well for
> the ideal of what Physicists and Computer Scientists (ie people who
> designed Mathematica and Maple) think maths and maths notation is.
True. The phrase that you are supposed to use in this context (for some
strange reason) is ``K through 12'' I am not sure if that automatically
means anything to American readers, it didn't mean anything to me until
it was explained that this meant basically starting school until first
year University. Ie almost no 20th century mathematics. Not
surprisingly, 19th Century Mathematics is by now reasonably stable,
rather like LaTeX 2.09 I suppose.
> The kind of maths I indulge in (when not indulging in chat on serious
> email lists, etc etc:) is not like this,
But (Content) MathML is not designed for you. OpenMath is. OpenMath is
what I am supposed to be doing (when not indulging in chat on serious
email lists, etc etc:). The import of MathMl for OpenMath is that
it potentially supplies a rendering mechanism.
> Don called a `subformula' but which is very badly handled (both
> syntactically and semantically in `standard TeX/LaTeX').
Since Chris and I appear to be pretending to reply to each other's
messages while in fact speaking to a wider audience, I could point out
in particular the fact that not everyone seems to realise, that if you
put something inside {....} in TeX in math mode the effect is rather
strange, all spacing is frozen, and all line breaking is prevented, just
as if that had been an \mbox construction if you actually used an
explicit box, it would also fix other things, like \textstyle, but in
all other respects groups in math mode behave like boxes (except for
\begingroup groups which just behave really strangely:)
A typical MathML renderer is recommended to progressively alter its line
breaking behaviour depending on the depth of mrow nesting, which is
rather different from the TeX all or nothing behaviour. (As far as I
know none of the currently available MathMl renderers actually do this,
or get anywhere near TeX quality generally, but it is early days yet).
David
