RE: [Axiom-developer] Axiom BNF

[Martin Rubey]

Dear Bill,

your analysis is of course correct, my only point being, that *I* just don't
*see* the correct amount of indentation. And yes, the first part is correct and
as I made up the second part, I mis-indented the if, so given this snippet, the
compiler would complain (yes, it would). On the other hand, the reference to g
was not there when I discovered the bug, so both snippets were completely
legal.

An emacs mode for editing piled code would be great and should go on the
WishList.

Since we are most unlikely to agree on a "good" style of coding (there are
simply too many very different mathematician-programmers around, some of them
come from a C backgound, some of them like Lisp, others Phython, I know some
who do their stuff in ML) I maintain that it's a good thing to have two possible
syntaxes to choose from.

Personally I'd favour compatibility with Aldor usage, in order to enlarge the
community.

Regarding the original (BNF) discussion: Note that Aldor has some new concepts,
like extend. Furthermore, it seems to me, that in Aldor you can define domains,
categories and functions in places where the Axiom compiler would complain.

Martin