Re: RM and abstract syntax trees
Date: Tue, 13 Nov 2007 16:33:27 -0800
Message-ID: <1195000407.415228.291170_at_t8g2000prg.googlegroups.com>
On Nov 13, 9:26 pm, "David Cressey" <cresse..._at_verizon.net> wrote:
> Now maybe we can get back to RM and abstract syntax trees, the main topic
> of discussion.
>
> Is there anything about syntax parsing that makes the RM an awkward tool for
> modeling it?
Consider that a string is represented relationally (eg as a mapping between index position and character) and an AST is represented relationally, I have doubts whether RM/RA itself could offer any particular advantages (with its set level operations) for the problem of parsing the string into the AST for some given grammar, even if teamed up with a functional, imperative or logic programming language, or direct support for transitive closure. Either a counterexample or some formal justification for that conjecture would be interesting.
I also doubt whether RM/RA offers any particular advantages for subsequent processing of a relational representation of an AST. For example, does it help with unification?
Does anyone have any insights on these questions? Received on Wed Nov 14 2007 - 01:33:27 CET