limitations of MathML



> You can only convey --at best-- the intersection of information content
> between CAS-1 and CAS-2 and MathML by using MathML.

But isn't this the best we can do without doing anything actively with an 
import/export filter?

> MathML is currently a way of defining a display.
> For simple things this works.  If everyone's browser
> supports MathML display  (like current Netscape), then
> MathML display is ok, but for small expressions only.

Sorry, but XML - and therefore MathML - has nothing to do with display!
XML describes structure and content and MathML is supposed to be a common way 
to do this for mathematical content.

> MathML may be useful, but it is not likely to be a really
> good total solution.

An XML interface to the Maxima engine really is the best solution! The fact 
that, at the moment, most web browsers can't display a certain document 
doesn't mean the way of describing the document is not a good concept. MathML 
provides a ready to use DTD and we can easily extend this DTD to our needs 
without having to reinvent the wheel at all places: display, input, file 
handling. Also, it should be more than easy to convert the lisp tree 
representation of the current output to an XML tree. And by letting the input 
already be in a tree structure parsing should become really simple. There is 
no need at all to convert the maxima internal representation, we just need 
input and output filters!

Greetings,
	Michael