I've written something coming up on a million lines of C code now, and I find the syntax great for writing straight-forward basic archival support code.
Furthermore, about 70% of today's programmers were imprinted like geese on C notation and will never be comfortable using anything radically different.
So a C-ish compiler and shell for Muq are strategically very important.
My current notion would be to write add some C syntax lexical scanner
primitives to the Muq virtual machine, plus an LALR(1) parser inner
loop primitive, and then write some Perl (say) scripts to massage
the output from (probably) byacc into form suitable for loading
into the Muq db. (2)
The output from the parser might well be Lisp S-expressions, which could be fed into the Lisp compiler after a little massaging by suitable Lisp macros. Naturally, we'd never tell lisp-phobic C devotees what was going on under the hood. :)
C-style pointer arithmetic doesn't make much sense on the context of the Muq virtual machine, so the actual language is likely to wind up chopped down much like Java on that front. Java might actually be a good model, in some ways. (Python might be another model worth looking at.)
I'll do this if need be, but it may take me awhile to get to it. This would be a helpful project for someone else to take on.
Go to the first, previous, next, last section, table of contents.