W3C home > Mailing lists > Public > public-rdf-wg@w3.org > May 2012

Re: RDF Concepts - What is a Node?

From: Nathan <nathan@webr3.org>
Date: Tue, 29 May 2012 19:03:54 +0100
Message-ID: <4FC50F8A.60206@webr3.org>
To: Richard Cyganiak <richard@cyganiak.de>
CC: Dan Brickley <danbri@danbri.org>, "Peter F. Patel-Schneider" <pfpschneider@gmail.com>, RDF WG <public-rdf-wg@w3.org>, David Wood <david@3roundstones.com>
Richard Cyganiak wrote:
> On 29 May 2012, at 08:30, Nathan wrote:
>> What do we call the thing which is the conjunction of the set of statements/assertions for a node within an RDF Graph then?
> 
> I tend to use phrases like this (obviously not formally defined, but reasonably accurate):
> 
> If talking about the graph level:
> 
> “the triples having (the node/IRI/blank node) as their subject”
> (if I only talk about “outgoing” arcs)
> 
> “the triples involving (the node/IRI/blank node)”
> (if talking about outgoing and incoming arcs)
> 
> If talking about the domain level:
> 
> “the statements about (the resource/thing)”
> (if I only talk about the meaning of the “outgoing” arcs)
> 
> “the statements involving (the resource/thing)”
> (if talking about the meaning of outgoing and incoming arcs)
> 
> “a description of (the resource/thing)”
> (very informally; could include additional triples beyond those immediately connected to the resource)
> 
>> the CBD?
> 
> The CBD is different depending on blank nodes/IRIs and is actually very rarely used in practice.
> 
>> It seems like a rather large concept which goes strangely undefined / unnamed in the RDF specs - and the original reason for the mail. I wrongly termed it the "node" in my original mail which led to some confusion, and if I'm getting confused / don't know what to call this concept which is pretty pivotal to RDF usage, then I'd suggest it may be useful to define / provide for others.
> 
> This concept doesn't play a significant role in any spec I'm aware of.
> 
> It plays a role in some implementations, but often inappropriately (e.g., incoming arcs are omitted for no good reason, crippling the implementation).
> 
> If I needed to make up a term for it, I'd probably call it a “node description” and carefully define it. A node description would essentially be isomorphic to a set of triples that all share the same subject.

Yes - this appears to be the same as the "entity description" Kingsley 
often mentions.

Also trying to think of a time when we use names in our machines / code 
to refer to anything other than these "descriptions", as anchors to 
them. It appears to me that the interpretation as to what 
entity/resource a human then associates the name with, is open to 
interpretation, changes over time, and may be one of those things we'd 
normally separate out as being of no concern to tech specs.

Best,

Nathan
Received on Tuesday, 29 May 2012 18:04:34 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:49 GMT