RE: ungetable http URIs

If the only information that we record for contributions is the proxy for the person who contributed to the course, then it doesn't really matter that there was a person behind that proxy.  There simply isn't enough data in the records to link back to the person who created the contribution; that linkage is inferred and imprecise.

Take the example of a CVS log.  In that log you might find information that 'ks' made a change dated 12/3/2003 to a file named canon.pl.  Did 'ks' actually make that change?  No, there was a real person who made the change, but the only metadata that is recorded is an account-name, not a canonical person reference.

Likewise with OCW, the data that is recorded doesn't include canonical person references, nor could it since it doesn't have the features required of a naming authority.  The only data that is available is a proxy identifier; practically, an entry into the list of contributors maintained by OCW.  These are non-canonical references and cannot be canonicalized without reference to some authority, and the mapping to authorities can only be weak.

>> I propose that the databases we have in hand always reference proxies, 
>> not real things.
>
> I disagree.  I would say that in databases what we have in hand are simply references to real things.

A nice thought, but inconsistent with the actual data.

-----Original Message-----
From: www-rdf-dspace-request@w3.org [mailto:www-rdf-dspace-request@w3.org] On Behalf Of Nick Matsakis
Sent: Thursday, December 04, 2003 10:59 AM
To: Kevin Smathers
Cc: Stefano Mazzocchi; SIMILE public list
Subject: Re: ungetable http URIs




On Thu, 4 Dec 2003, Kevin Smathers wrote:

> An OCW course lists the contributors to that course; gettable or 
> ungettable.  Your argument doesn't provide any advice -- it depends on 
> what you mean by a contributor, a real person, or a proxy for that 
> person.

Records/proxys don't contribute to courses, people do.  If we have a statement that says "X dc:contributor Y", then "Y" is a URI that _names_ that person.  Statments like "Y vc:FN 'Nick'" then are easily
interpretable: 'that which has the URI name Y has the Vcard first name 'Nick'"

If we want to create a name for the name, we can do this, but I think this is really only useful for the database administrator. The typical researcher wants to find information about the materials, not about the database.  The database administrator might want to make statements like "that record has an incorrect title" or "these two records refer to the same person", but I don't see why the administrator would want to enter that information in the database, rather than just fix it.

> I propose that the databases we have in hand always reference proxies, 
> not real things.

I disagree.  I would say that in databases what we have in hand are simply references to real things.

Regards,

Nick

Received on Thursday, 4 December 2003 14:32:47 UTC