> No, I know you can apply them `recursively', I mean something more > like an expression in a CFG or yacc. > > > > > can you outline somehow what you're thinking of? > > Basically, if you could take a bracketed expression in sam and then > name it, and then call it recursively. > > All the problems with CFGs (shift/reduce problems, ambiguities, etc) > would possibly apply.
could you give an example in your proposed language? i don't see how greedy regular expressions wouldn't kill you. example. let's say you have a SRE that breaks text into lines, you couldn't apply that recursively. in fact i think you'd have the same problem with any SRE, since they are greedy and can't count. maybe i'm just confused because 'x' goes from a blob of text to a stream of tokens. where as grammars go from a stream of tokens to productions. maybe you mean to replace the traditional tokenizer with named SREs? - erik
