On Tue, May 15, 2001 at 08:24:55AM +0200, Andre Poenitz wrote:
> > Alejandro, this is what we had talked some years ago. How is the support
> > for the [mitic ;-)] new mathed?
>
> As far as I understand MathML it covers two aspects of formulae: some
> typesetting part and some semantical part.
There are two "styles", the presentational and the content based.
The best analogy is the relation between the <bold> and <strong> tags in
html, <bold> is the presentational style, and <strong> is the content based.
More up to the point, suppose you have two functions, f and g, then the
content based style will know that the resulting output:
f(g(x))
x is the argument of the g function, and that g(x) is the argument of
f...
> I am afraid we won't get the full blown semantical part for starters, since
> this does not match nicely the internal representation.
That was the reason why Alejandro choose the presentational style to
implement (at least for a first stage), this is much close to the present internal
structure.
> I plan, however, to improve this later. One of my personal goals is
> interoperability with computer algebra systems, and we need the semantical
> correctness for these anyway.
The content based "style" is really easy for those systems since they know
the real meanings of the different objects.
It will be fun if you succeed. :-)
> I think the way to achive this is to built more complex insets with a
> complete meaning both for typesetting and semantics (i.e. \int_a^b xdx
> currently is \int + _a + ^b + xdx which is pretty hard to parse correctly.
> If we had a special 'integral with bounds' insets, it would be trivial.
>
> Andre'
>
> --
> André Pönitz ............................................. [EMAIL PROTECTED]
--
José
