new syntax - a provocation

Richard A. O'Keefe <>
Tue Sep 23 10:11:15 CEST 2003


I suggested "-import_literal(Module, Functions)".

"Vlad Dumitrescu" <> replied:
	A good idea!  Why not let the calls that can't be resolved at
	compile time expand into a run-time constructor call?  Well, at
	least where this is meaningful.
	
That's precisely what I proposed:  replace the calls if you can (and,
implicitly, but, I thought, obviously, _don't_ replace the calls otherwise)
and give an error message if unreplaced calls are left where they are _not_
meaningful.

	But I still think it might be a good thing to mark in some way
	that these constructs are special ;-)

Well, prefix ? isn't being used for anything _useful_ (:-) (:-) (:-) ...

	> Lisp doesn't get its power from its *concrete* syntax but from its
	> *abstract* syntax.
	
	True, but in Lisp's case the two are the same :)
	
Would that it were true.  When you write a Common Lisp macro, a
substantial amount of rewriting may already have been done, and this
rewriting may involve forms that were not in the source code and quite
possibly could not have be, certainly undocumented stuff.  And in
Scheme, the "identifiers" that hygienic macros have to worry about are
not at all the same as "symbols" that you find in the source code.



More information about the erlang-questions mailing list