X-FeedAbuse: http://nntpfeed.proxad.net/abuse.pl feeded by 88.191.16.109
Path: csiph.com!eeepc.pasdenom.info!news.pasdenom.info!news.dougwise.org!nntpfeed.proxad.net!nospam.fr.eu.org!usenet-fr.net!de-l.enfer-du-nord.net!feeder1.enfer-du-nord.net!news.glorb.com!news2.glorb.com!news.glorb.com!transit3.readnews.com!textspool1.readnews.com!news-out.readnews.com!news-xxxfer.readnews.com!news.misty.com!news.iecc.com!nerds-end
From: Jack Smith <ilikequoting@katamail.com>
Newsgroups: comp.compilers
Subject: Re: Abstract Interpretation vs DFA in Clang and GCC
Date: Wed, 9 Feb 2011 00:31:54 -0800 (PST)
Organization: Compilers Central
Lines: 27
Sender: news@iecc.com
Approved: comp.compilers@iecc.com
Message-ID: <11-02-010@comp.compilers>
References: <11-01-034@comp.compilers> <11-01-069@comp.compilers>
NNTP-Posting-Host: news.iecc.com
X-Trace: gal.iecc.com 1297263048 83026 64.57.183.58 (9 Feb 2011 14:50:48 GMT)
X-Complaints-To: abuse@iecc.com
NNTP-Posting-Date: Wed, 9 Feb 2011 14:50:48 +0000 (UTC)
Keywords: analysis
Posted-Date: 09 Feb 2011 09:50:48 EST
X-submission-address: compilers@iecc.com
X-moderator-address: compilers-request@iecc.com
X-FAQ-and-archives: http://compilers.iecc.com
Xref: csiph.com comp.compilers:972

On 17 Gen, 12:14, torb...@diku.dk (Torben Fgidius Mogensen) wrote:
> Jack Smith <ilikequot...@katamail.com> writes:
> > since someone says that nowadays it's better abstract interpretation
> > than DFA can anyone tell me what's the difference among them?
>
> Abstract interpretation (AI) has a strong tie to the semantics of the
> language: Each value in AI corresponds to a set of values in the
> semantics, which makes it relatively easy to prove the correctness of
> an anlysis.  Values in data-flow analysis (DFA) do not have such a
> clear relation to the semantics, so they are more difficult to prove
> correct.  On the other hand, DFA can do analyses that are difficult to
> express as AI, such as liveness.  To do liveness analysis with AI you
> need a continuation-passing semantics.
thank you for your reply,

while studying for my exam i  figured out what you wanted to say :)
we could say also that abstract interpretation can do many things,
among these there's DFA too, could we?
for example, the polyhedra library (PPL) can do a value-range analysis
for imperative programs that is a sort of DFA. right?

could you go more in deep about liveness analysis?
why would i need a continuation-passing style semantic? have you got
some paper about that?

thank you