Re: symbolic interpretation

"Ira D. Baxter" <>
18 Jan 2001 01:05:04 -0500

          From comp.compilers

Related articles
symbolic interpretation (Lorenzo Bettini) (2001-01-11)
Re: symbolic interpretation (2001-01-18)
Re: symbolic interpretation (Ira D. Baxter) (2001-01-18)
Re: symbolic interpretation (Steven Carroll) (2001-01-18)
Re: symbolic interpretation (Basile STARYNKEVITCH) (2001-01-19)
Re: symbolic interpretation (2001-01-26)
| List of all articles for this month |

From: "Ira D. Baxter" <>
Newsgroups: comp.compilers
Date: 18 Jan 2001 01:05:04 -0500
Organization: Posted via Supernews,
References: 01-01-072
Keywords: interpreter, optimize
Posted-Date: 18 Jan 2001 01:05:04 EST

I don't about specific papers, but you might be interested in the
DMS Reengineering Toolkit, which is generalized compiler technology
for automating customized analysis and modifications to sources
for large scale software systems.

DMS parses files to ASTs, and then can apply associative-commutative
rewriting using the surface syntax of the target langauge.
(It can do procedural rewriting too, but everybody does that).
This can be used, for example, to simplify C++ preprocessor conditionals
when a configuration variable is known to be forever false,
and to drop away the obviously dead branches of the preprocessor
There's lot of other practical applications; we are doing
a factory controller code generator that makes heavy use
of boolean simplifications implemented by rewriting.

I can send you some slides showing roughly how DMS works,
and a number of applications that use (directly or indirectly)
the rewriting.

Ira D. Baxter, Ph.D., CTO email:
Semantic Designs, Inc. web:
12636 Research Blvd. C-214 voice: (512) 250-1018 x140
Austin, TX 78759-2200 fax: (512) 250-1191

"Lorenzo Bettini" <> wrote in message
> Do you have any pointer and/or reference to symbolic interpretation in
> compiler writing, please? I would need a specific one, not just a
> generic one.

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.