home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cis.ohio-state.edu!rutgers!faatcrl!iecc!compilers-sender
- From: masticol@cadenza.rutgers.edu (Steve Masticola)
- Newsgroups: comp.compilers
- Subject: Re: Abstract Interpretation Query
- Keywords: functional, logic
- Message-ID: <92-09-044@comp.compilers>
- Date: 6 Sep 92 14:22:09 GMT
- References: <92-09-038@comp.compilers> <92-09-041@comp.compilers>
- Sender: compilers-sender@iecc.cambridge.ma.us
- Reply-To: masticol@cadenza.rutgers.edu (Steve Masticola)
- Organization: Rutgers Univ., New Brunswick, N.J.
- Lines: 22
- Approved: compilers@iecc.cambridge.ma.us
-
- vugranam@pike.ee.mcgill.ca (Vugranam Chakravarthy Sreedhar) writes:
- >AI, although nice, is expensive in terms of time and space. I doubt if the
- >concept is being used in any production compilers.
-
- I don't believe that this generalization holds. As I understand it,
- abstract interpretation (please forgive me if I don't use the prejudicial
- term "AI" :-) is a theoretical framework for connecting a program to an
- abstraction of that program, or for generating the abstraction given some
- particular information you'd like to estimate. The abstraction is
- supposed to have properties that make it nice for data flow analysis
- (monotonic edge functions, etc.)
-
- Standard problems like reaching definitions can be cast in an abstract
- interpretation framework, and it costs nothing in terms of analysis time
- to do so.
-
- If anyone knows differently, please jump in...
-
- - Steve Masticola (masticol@cs.rutgers.edu).
- --
- Send compilers articles to compilers@iecc.cambridge.ma.us or
- {ima | spdcc | world}!iecc!compilers. Meta-mail to compilers-request.
-