W3C home > Mailing lists > Public > public-sw-meaning@w3.org > November 2003

Re: SWSL declarative semantics

From: Graham Klyne <GK@ninebynine.org>
Date: Fri, 21 Nov 2003 15:37:37 +0000
Message-Id: <5.1.0.14.2.20031121153530.030d4958@127.0.0.1>
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
Cc: public-sw-meaning@w3.org

At 08:49 21/11/03 -0500, Peter F. Patel-Schneider wrote:
>Agreed.  However, this is not really a model-theoretic semantics at all.
>Instead it is a translation to some other formalism that may or may not
>have a model-theoretic semantics.  If the translation is intuitive (and
>simple and ...) then, sometimes, a model-theoretic semantics for the other
>formalism may induce a decent model-theoretic semantics for the initial
>formalism.  (But then why not take the easy step of writing down this
>induced model-thoeretic semantics directly?) However, such translations are
>usually not very simple and thus any benefits of a model-theoretic
>semantics of the other formalism do not accrue to the original formalism.
>
>The translation for DAML+OIL, for example, does NOT provide this benefit.
>Even the simple translation from SKIF to FOL is rather suspect in this 
>regards.

I *think* I understand the distinction between proof-theoretic and 
model-theoretic semantics.  But I'm not clear what are these advantages of 
MT semantics to which you refer.  Can you point me at a not-too-heavy essay 
or paper that provides some elucidation of this point?

Thanks.

#g



------------
Graham Klyne
For email:
http://www.ninebynine.org/#Contact
Received on Friday, 21 November 2003 11:09:16 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:42:15 GMT