Thanks Sandra for the corrections --- glad I broadcast my reply rather
than mailing directly.
I knew compilation was the more common route --- just didn't emphsise it
enough. The point I totally failed to make was that having an interpreter
can be pretty nice. (The work I'm doing at the moment requires me to link
to Motif so I can't prototype in Gofer - I really miss that!)
Do you really have structures with 20-30 components? That seems an
unmanageable number. (Though more manageable for being accessible via
field names (or whatever) I'm sure. Presumably these are mostly manipulated
via destructive update??
What I meant when I said "no type classes" was that those nice overloading
tricks we play (user extensible Text class, defining Monad classes, etc.)
aren't available in LISP. Are you saying you can?
(I expect something like Text's show function is built in but what of
defining a new class from scratch? Do you mess around with examining
the type of an object directly? Is this easy/ reliable to use? Can this
distinguish between different abstract data types with the same
representation?)
Alastair
