W3C home > Mailing lists > Public > public-rif-wg@w3.org > March 2006

RE: [UCR] RIF needs different reasoning methods

From: Gerd Wagner <wagnerg@tu-cottbus.de>
Date: Mon, 13 Mar 2006 10:33:38 +0100
To: "'Sandro Hawke'" <sandro@w3.org>
Cc: <public-rif-wg@w3.org>
Message-ID: <000201c64681$360e7340$55032b8d@TMGWAGNER>

> My understanding is that derivation rules and integrity 
> constraints are
> just a bifurcation of Horn rules (or sometimes the two halves of
> something beyond Horn).  As such they fit easily into the same
> semantics.

Not really, there are both syntactic and semantic 

- derivation rules, in general, correspond to Gentzen 
  sequents; in logics having a reasonable implication 
  connective (satisfying the deduction theorem), a 
  derivation rule may be viewed as an implication 
  formula; derivation rules are mainly used to express
  definitions of derived predicates (or derived tables
  in the style of virtual views); this means they are
  applied at query time 
- integrity constraints, in general, correspond to
  sentences of some logic (which may be classical FOL,
  temporal FOL or some modal FOL); as emphasized by
  SBVR, they come with a modality (either alethic or
  deontic) that determines their semantics in terms 
  of handling violations; integrity constraints are 
  used, at the business logic level, for expressing
  all kinds of laws (includig structural and natural 
  laws); at the IT system level, they are mainly used 
  to catch all kinds of faulty states of the system 
  (such as referential integrity violations); this
  means they are applied at state change time

Given all these differences, how can you say "they 
fit easily into the same semantics"?

The fact that, in certain circumstances, they may both
take the form of a Horn formula does not mean they 
have the same semantics (like a question Q and an 
assertion A do not have the same semantics even if
Q and A are expressed with the same logical sentence).  

Received on Monday, 13 March 2006 09:36:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:47:37 UTC