W3C home > Mailing lists > Public > www-rdf-logic@w3.org > April 2001

Re: Question: DAML cardinality restrictions

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Tue, 03 Apr 2001 10:03:41 -0400
To: dallsopp@signal.dera.gov.uk
Cc: www-rdf-logic@w3.org
Message-Id: <20010403100341E.pfps@research.bell-labs.com>
From: David Allsopp <dallsopp@signal.dera.gov.uk>
Subject: Re: Question: DAML cardinality restrictions
Date: Tue, 03 Apr 2001 14:20:54 +0100


> "Peter F. Patel-Schneider" wrote:
> 
> > > Let's say I have data for a person 'A', but not for their father. Does
> > > this mean my data do not conform to the ontology/schema (since A can't
> > > point to a father instance in my small model of the world)?
> > 
> > Not at all.  There is nothing in DAML+OIL that requires the data to make
> > everything explicit.  It is perfectly OK to require that people have
> > exactly one father, and also have object that belong to that class and have
> > no known specific father.
> 
> OK. What if the data say that someone has two fathers? (for the sake of
> argument, I receive an RDF message where the person's Resource has two
> Father properties).  Presumably we must reject this message as logically
> inconsistent? (Or invoke some process to decide which is the correct
> father).

Yes, this would be inconsistent.  Any attempt to remedy this inconsistency
would be outside the scope of DAML+OIL.

> I take your point about representation, but at some point, to make
> use of our data, we will often have to leave the logical
> representations;
> most software does not use FOL or anything like it, and probably never
> will.  That's my angle - I am trying to introduce semantic web
> technology to allow _existing_ applications/agents to communicate and
> interoperate more effectively, without complete re-writes.  These
> applications, on the whole, cannot
> deal with logic, only concrete data. (If the family-tree plotting
> software can't find out _who_ Joe's father is then it isn't
> interested...).  
> 
> So, I am trying to understand how I can use semantic web technology to
> allow some inference, some translation of data, some
> merging/collation/fusion of data, whilst eventually allowing me to
> extract data into more rigid forms to be imported into
> applications, displayed on screen, used to manipulate hardware, or
> whatever.
> 
> Sorry to sully the pure logic with primitive unenlightened applications,
> but I can't just throw everything away and start from scratch 8-)). We
> have to be able to bridge between 'ordinary' applications and logic, if
> only because most technology users don't know any formal logic and won't
> be prepared to learn any; this group, of course, is atypical!  For this
> reason I worry when I see query languages couched in terms of FOL; I
> think most people just won't use them without a 'user-friendly' layer on
> top, however powerful they are.

There is nothing to prevent (some, limited) applications from working with
only the ground portion of DAML+OIL (or RDFS) pages.  This, however, has to
be done with great care.   

Perhaps the biggest change is from the closed-world of databases to the
open world of the semantic web.  This has little to do with RDFS or
DAML+OIL, of course.  For example, either the family-tree plotting
software has to understand the open world that it works in, or, perhaps
more likely, users of the software have to understand that its results only
represent the currently-known information, and may not be complete.

As far as logical sophistication of users goes, I don't any qualitative
difference between databases and knowledge bases.  To really understand
either one, and, in particular, to understand querying of either one,
requires a certain level of understanding of their underlying formalisms.
The only qualitative advantage that databases may have is one of
familiarity.

peter
Received on Tuesday, 3 April 2001 10:04:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:38 GMT