W3C home > Mailing lists > Public > www-rdf-interest@w3.org > December 2002

RE: Re: why arcs never be merged?

From: Graham Klyne <GK@NineByNine.org>
Date: Fri, 29 Nov 2002 17:15:12 +0000
Message-Id: <5.1.0.14.2.20021129171420.03a25ac0@127.0.0.1>
To: "Danny Ayers" <danny666@virgilio.it>
Cc: <jeffzhang726@yahoo.com.cn>, <www-rdf-interest@w3.org>

Yes.

I think many applications would eliminate the duplicates, but strictly 
speaking they're not wrong (according to RDF) if they don't.

#g
--

At 04:13 PM 11/29/02 +0100, Danny Ayers wrote:

> >Dear Graham,
> > I mean the first example
> >    ex:foo ex:property ex:bar .
> >and
> >    ex:foo ex:property ex:bar .
> >    ex:foo ex:property ex:bar .
> >
> >Our team is developping an rdf engine which has an API as its part.
> >When we discuss about the operation of adding a statement to a
> >graph, we are puzzled
> >about whether a statement should be added to the graph if there is
> >already a same
> >statement in the graph. As to the rule of merging graphs, it seems
> >the same statement should be added to the graph. But we can not
> >find any real world example in which multiple copies of statements
> >make different meanings with single statement. Is there any example?
>
>(not Graham, but still)
>
>let me try...I hope I've got this right...
>
>if we assert a couple of statements
>
>http://example/x dc:creator Jeff
>http://example/x dc:creator Jeff
>
>then ask the question,
>
>http://example/x dc:creator ?
>
>then we should get the answer
>
>http://example/x dc:creator Jeff
>
>or
>
>http://example/x dc:creator Jeff
>http://example/x dc:creator Jeff
>
>or even
>
>http://example/x dc:creator Jeff
>http://example/x dc:creator Jeff
>http://example/x dc:creator Jeff
>
>- they're all true
>
>if had asserted instead
>
>http://example/x dc:creator Jeff
>http://example/x dc:creator Ora
>
>the answer could have been
>
>http://example/x dc:creator Jeff
>http://example/x dc:creator Ora
>
>or
>
>http://example/x dc:creator Ora
>http://example/x dc:creator Jeff
>http://example/x dc:creator Jeff
>
>or other variations, depending on the app. In practice I suppose we'd avoid
>having the query just return a 'null' statement, even though that would in
>effect be true too. So in an app I would be tempted to go for the simplest
>solution, i.e. merge two identical statements into one.
>
>Cheers,
>Danny.

-------------------
Graham Klyne
<GK@NineByNine.org>
Received on Sunday, 1 December 2002 05:08:48 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:57 GMT