W3C home > Mailing lists > Public > www-rdf-interest@w3.org > December 2003

Re: Trust, Context, Justification and Quintuples

From: Thomas B. Passin <tpassin@comcast.net>
Date: Thu, 18 Dec 2003 19:07:08 -0500
Message-ID: <3FE2412C.9030305@comcast.net>
To: www-rdf-interest@w3.org

Dmitry Borodaenko wrote:

> On Thu, Dec 18, 2003 at 09:16:40AM -0500, Thomas B. Passin wrote:
>>>we did some brainstorming about trust, context and the justification of
>>>query results and ended up with:
>>>- an extended RDF data model based on quintuples (a triple plus two
>>>additional elements: context and statement ID).
>>It seems to me that it would be possible to make statements about a
>>statement by creating a predicate that is an instance of a predicate
>>class.  That instance would be using in only one statement.
>>Therefore, any statement about that predicate could be taken to be a
>>statement about the triple it is part of.
>>For example, instead of
>>{#Bob, #friendOf, #Mary}
>>We could have
>>{#friendOf_1, rdfs:type, #friendOf}
>>{#Bob, #friendOf_1, #Mary}
>>{#friendOf_1, #veracityAssessment, #Dubious}
>>This would work as long as the instance of the predicate were used in
>>only one statement.
> Why not just reify the statement?

Well, of course one can do that.  But the reason people keep coming up 
with suggestions to have a quad or a statement identifier instead of 
using reification has, I think, to do with weaknesses in RDF 
reification.  There are several issues that I know about -

1) It is complex - you end up with four triples where all you want to do 
is to reference a statement.

2) The interpretation of a reified statement is not well defined.  For 
example, it is NOT a representation for any actual triple in the data 
store, and it is NOT considered "asserted"... So what is a reified 
statement and how should it relate to the other triples?

3) It is contorted -  if a statement had its own resource identifier, it 
would be easy and natural to refer to it as the object of an RDF 
statement.  And, of course, practical RDF processing systems are likely 
to have some internal identifier, so why to make one externally available?

I think it is issues like this that cause people to dislike reified 
statements and to desire an alternative.


Tom P
Received on Thursday, 18 December 2003 19:09:50 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:07:48 UTC