W3C home > Mailing lists > Public > public-rif-wg@w3.org > January 2008

Re: [BLD] Frame without slot/value pair?

From: Michael Kifer <kifer@cs.sunysb.edu>
Date: Thu, 10 Jan 2008 14:41:30 -0500
To: Christian de Sainte Marie <csma@ilog.fr>
Cc: "Boley, Harold" <harold.boley@nrc-cnrc.gc.ca>, RIF WG <public-rif-wg@w3.org>
Message-ID: <29020.1199994090@cs.sunysb.edu>

> Michael Kifer wrote:
> > 
> > This is precisely what I explained:
> > 
> > t[] == t[?S -> ?V] or ?O[t->?V] or ?O[?S->t].
> > 
> > It is a useful shortcut, which comes naturally syntactically and
> > semantically.
> Oh, ok! This is not how I understood what you said initially:
> > Formulas like t[] are also useful. If they are allowed, their
> > semantics is that the object t exists (without testing any of its
> > properties).
> So, you propose that we could have t[] as syntactic sugar to say that:
> - object t has some properties and values,
> - or there exists some object with some value for property t,
> - or there exists some object with value t for some property.
> Is that correct?

In simple terms, it just means that the object is known to the KB.

But t[] is not a syntactic sugar. It is a natural degenerated case of a
particular syntactic form -- just like f(). If we have f() then there is no
reason to not have t[]. The fact that t[] is equivalent to something else is
just a tautology, and tautologies always exist in logic. You would not say that
a /\ (b\/c) is a syntactic sugar for (a/\b) \/ (a/\c), right?

> If yes, then I would rather not allow that syntax: it seems a really 
> counter-intuitive shortcut to me; and rather cumbersome to implement for 
> target languages that do not have it, for a benefit that seems fairly 
> limited.

It is neither counter-intuitive nor cumbersome to implement.
It is trivial to implement in logic-based rule languages that are
targets for BLD.

Now, I do not know which target languages you are talking about, but if you
are talking about PRD then it is a different can of worms. We have no PRD
right now, and it is unclear when we'll have a usable version. It is not
clear what will go into PRD either. For instance, how about equality in
PRD? It is part of BLD, but I am not aware of any PR language where
equality is allowed (not as a test, but as facts or rule heads).  Ditto
with function symbols. So, if you have PRD in mind, please do not use it as
a design principle for BLD, because these two are quite different.


> Christian
Received on Thursday, 10 January 2008 19:41:39 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:47:49 UTC