As is (hopefully) obvious, I've been digging in to the parser combinator idea. I'm looking at it as a DSL.
The thing about an in-language DSL extension is that it separates the input syntax from the semantics. Once an input syntax for a DSL is defined, it can be processed in multiple ways. At one extreme, we can imagine an implementation of the DSL that results in a parser. At the other, we can imagine a *parser* AST that can be analyzed by various sorts of algorithms. The point is that these can be done from the same input. Thing is: given that an input syntax can be assigned different meanings, it raises the possibility that the same input might be included/loaded in different operator binding contexts for different purposes. Is this a good thing? A bad thing? An excessive complexity? shap
_______________________________________________ bitc-dev mailing list [email protected] http://www.coyotos.org/mailman/listinfo/bitc-dev
