> Using XML to communicate algebraic expressions would be a
> very poor choice. Reworking Maxima to use XML would be
> exceedingly time consuming, have wide ranging effects,
> and yield no additional functionality.
I don't aggree! There is no need at all to rework Maximas internals! It's only
the input/output parser to frontend applications and you named the advantages
yourself:
(1) context: Is the new input prompt belonging to some question the program
asked in the last output belonging to some computation demanded from the last
input? (blocks of blocks)
(2)
> very strongly constrained.
You don't have to tinker with maximas internals just to write a GUI for it.
and mentioned by others:
(3) There are ready to use tools to handle XML so there is no need to reinvent
the wheel.
(4) If maxima defines a DTD for itself we could ensure interoperability
between various frontends.
> Lisp lists have equal expressive power, are not
> verbose and redundant, and fit comfortably with the
> computational structure already implemented in Maxima.
Of course, maybe I wasn't clear enough about my intentions: Define a strongly
constrained protocol for communication with the maxima kernel. Ensureing a
clear distinction between controller (maxima kernel), document (notebook),
and views. The core project should only worry about the controller, e.g.
there is no need to provide load and save functionality by the kernel itself.
Kind regards,
Michael