W3C home > Mailing lists > Public > public-rdf-wg@w3.org > December 2013

Re: Proposed resolution needed: ISSUE-148: IRIs do *not* always denote the same resource

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Wed, 18 Dec 2013 07:36:20 -0500
Message-ID: <52B196C4.6040402@openlinksw.com>
To: public-rdf-wg@w3.org
On 12/18/13 1:52 AM, Pat Hayes wrote:
> On Dec 17, 2013, at 5:00 AM, Kingsley Idehen <kidehen@openlinksw.com> wrote:
>> On 12/17/13 7:06 AM, Markus Lanthaler wrote:
>>> I don't care much whether we use denote or identify. According to Pat,
>>> "identify" is technically more correct whereas Richard points out that
>>> "denote" is more consistent with the rest of the section. I personally
>>> prefer "identify" in this case because I believe that it is the term that's
>>> best aligned with RFC3986/RC3987 and WEBARCH.
>> Are you sure that Pat preferred "identify" over "denote" as you've presented above?
> Yes, I did.
>> RDF is about IRIs denoting entities. I believe it's been long established that every entity isn't a Web accessible resource.
> But a lot of Web specifications and foundational writings use "identify" as a close synonym for "denote" (or "refer to" or "name"). I used to rail against this, but its too late to change it and so instead I am proposing to use it. The issue we are dealing with, recall, is that David Booth is pointing out that speaking very strictly, we can't say that an IRI always denotes a unique thing, because it might denote one thing in one interpretation and something else in a another interpretation. And David is strictly correct, even when the IRI has been given a clear unambiguous sense by some specification or authoritative source. (Take any dbpedia resource IRI as an example.) This does not have the semantic consequences that David thinks it has, but that is a different discussion: the fact is, that this statement is not *strictly* correct; but there is a correct, and important, principle which is very close to this, but which would be very hard to explain in full detail in a document of this kind. If we use the slightly technically looser term "identify", we can escape the trap of being undercut by the strict model-theoretic reading, and have a statement that is reasonably close to being the actual architectural principle that is being appealed to here.
>> A sign provides signification [1][2]. For instance, I can have an Identity Card that "Identifies" me. In that identity card I could have an sign (e.g., an IRI) that denotes (signifies) me.
> The IRI is identifying you in the sense in which that word is used throughout the documents that Markus cites above, which are the ones that matter most in this context.
>> Basically, the Identity card is a collection of claims expressed as attribute=value pairs that coalesce around the sign that denotes me [3][4] i.e., a collection of statements represented as RDF triples.
> In this sense of "identify", it would be an entire specification document that "identifies" what xsd:number denotes. Whereas the usage on the Web is that xsd:number identifies a datatype, and on the Semantic Web that, in addition, it denotes what it identifies. I am going with the latter convention here.
> Pat

I am okay with the paragraph above. Thus,  we simply need to reflect 
that same sentiment in the primer i.e., the subtle (but important) 
difference between the Web and the Semantic Web.




Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen

Received on Wednesday, 18 December 2013 12:36:46 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:04:37 UTC