Compiling list of transformation/synthesis projects

baxter@SLCS.SLB.COM (Ira Baxter)
Thu, 15 Aug 91 09:57:41 CDT

          From comp.compilers

Related articles
Compiling list of transformation/synthesis projects baxter@SLCS.SLB.COM (1991-08-15)
| List of all articles for this month |

Newsgroups: comp.compilers
From: baxter@SLCS.SLB.COM (Ira Baxter)
Keywords: question, experiment, prototype, tools
Organization: Compilers Central
Date: Thu, 15 Aug 91 09:57:41 CDT

I would like to compile a relatively complete database of active
research projects, anywhere in the world, on "real" program synthesis
and/or transformation systems. The intent is to foster better
communication, between such projects. I will eventually post a
summary of responses, both to the net and to respondants.

The criteria I consider useful for deciding which projects match are
the following:

        a) A specification language (anything that can be interpreted as
              a predicate over acceptable generated programs, including
              abstract programs)
        b) A transformation or synthesis engine, applying rewrite rules
              explicitly or procedurally. Any control regime is acceptable.
        c) The intent that the rule base be augmented over time,
              by others than the maintainers of the system.
        d) Possession of a running implementation, or significant
              resources committed to obtaining an implementation
        e) Expectation that something larger than a toy program will
              be constructed, often in a language different
              of the specification langauge.

I view conventional compilers and application generators as outside this
realm. However, I would rather err on the side of including too much than
too little, so if you are not sure, please respond; also tell me why the
criteria should be loosened or changed. As an example, I don't view VLSI
synthesis as being excluded; ignoring geometry, the problems seem the same.

Ideally, responders will supply the following information, just to provide
initial comparative information. I have filled it in for our in-house

System Name: Sinapse
Specification Language: Schema slot fillers + various performance goals
Problem Domain: Numerical modeling codes
Target languages: CM Fortran, Fortran77, C
Transformation process: Symbolic and procedural refinement of program schemas
Overview Publication: "Scientific Programming by Automated Synthesis",
                                                E. Kant, F. Daube, W. MacGregor, J. Wald,
                                                in Automating Software Design, M Lowry and R.
McCartney, eds., AAAI Press, 1991
Research Group: Modeling and Simulation
Contact: Ira Baxter
phone: 1-512-331-3714
Organization: Schlumberger Laboratory for Computer Science
Address: PO Box 200015
                                                Austin, Texas, 78720

Pointers to other projects (and their principals) would also be useful.

Post a followup to this message

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