Re: Properties not predicates (was Re: PRIMER: draft data model section)

Brian--

Well, I *thought* I was clear as to what was being renamed:  the part of 
the statement that identified the property part of the statement (and 
since there's only one of them, I don't really have a problem with a 
certain amount of overloading, if that's what it is)!  If we want the 
terminology to make this distinction clearer, I see some trickiness on 
the horizon.  The M&S also refers to "property name" (especially in the 
context of the XML serialization), but, as we now know, the property 
isn't identified by a simple "name", but by a URI (even though that's a 
name, in a slightly difference sense).  That also means that the 
property is a resource, so you could have equally written:

 > The subject must (debatably) be a resource
 >
 > The predicate must be a resource
 >
 > The object may be a resource or a literal.

What we want to say is something like "the property component of a 
statement specifies the resource acting in the role of property for this 
statement".  We certainly need to distinguish that from situations in 
which a resource *is* a property, but we're referring to it in the 
subject or object positions of a statement (which are also 
possibilities).  In any event, we probably want to change "predicate" to 
something else, as Pat suggests, since it may be a bit misleading in the 
context.

--Frank

Brian McBride wrote:

> Just checking we are clear what is being renamed.  M&S uses the term 
> predicate for a component of a statement.  Thus a statement has three 
> components:
> 
>    a subject
>    a predicate
>    an object
> 
> The subject must (debatably) be a resource
> 
> The predicate must be a property
> 
> The object may be a resource or a literal.
> 
> The terms predicate and property in M&S mean different things.  Are we 
> losing the distinction, or is this distinction just no longer applicable.
> 
> Brian
> 
> 
> Pat Hayes wrote:
> 
>>> Frank:
>>> I called it a "predicate" because that's the term used in the M&S.  If
>>> we've changed it officially, I must have missed it (where would this be
>>> documented?).
>>
>>
>>
>> OK, let me make this a formal suggestion for an item.
>>
>> PROPOSED that:
>> The things called 'predicates' in the M&S shall henceforth be called 
>> 'properties' (preferred) or 'relations'.
>> This is in line with the standard usage in the description logic, 
>> database and formal logic communities, and is used in the DAML+OIL 
>> documentation. The term 'predicate' is deprecated as potentially 
>> misleading, since properties are not predicates in the sense that word 
>> is usually used in formal logic.
>>
>> Pat
> 
> 
> 


-- 
Frank Manola                   The MITRE Corporation
202 Burlington Road, MS A345   Bedford, MA 01730-1420
mailto:fmanola@mitre.org       voice: 781-271-8147   FAX: 781-271-875

Received on Monday, 22 October 2001 11:00:43 UTC