The metadata communities spend a lot of time talking about whether
identifiers should be de-referenceable as a network locatable resource.
Can, should, should be allowed to, should be prohibited from, all are
argued. A URI does not need to be a de-referenceable HTTP URL. Even when
people use http:// url's as _identifiers_, some argue that they ought
_not_ to be actually de-referenceable, they should just serve as a
unique ID that happens to _look_ like a network locatable resource. I
can see the arguments both ways for whether an _identifier_ (URI) ought
to be purely a dumb unique ID, or ought instead to be a resolveable
location on the network. We who are interested should figure out how to
join the community of practicioners who have been discussing these
issues for several years now, rather than trying to re-invent the wheel
ourselves.
So yeah, Casey said using URI as a persistent identifier, which is not
quite the same thing as Conal (switching to 'URL') suggests.
Jonathan
Conal Tuohy wrote:
> This idea of using a URL as a public identifier in authority work is a
> very good one. In the Topic Map community the concept has long been
> known as a "Published [or 'Public'] Subject Indicator"; PSI.
>
> The idea is that if a URL-addressable information resource is about
> ('indicates') some subject (a person, for instance), then the URL of
> that resource can be used as an identifier for the person. This is
> simply another way of treating that URL. Of course the URL is an
> identifier for an information resource, but when read in a particular
> context (such as inside a <subjectIndicatorRef> in an XML Topic Map), it
> can be interpreted instead as an identifier for the thing which the
> resource is about (a person, for instance).
>
> For those interested, an OASIS committee did some valuable work on PSIs
> a few years ago:
> http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=tm-pubsubj
>
> In particular, they produced a short paper which introduces the concept,
> and provides practical guidelines:
> http://www.oasis-open.org/committees/download.php/1217/wd-pubsubj-introduction-01.htm
>
> Con
>
>
> On Mon, 2007-05-14 at 22:03 -0400, Ross Singer wrote:
>
>> On 5/14/07, Jason Griffey <Jason-Griffey_at_utc.edu> wrote:
>>
>>
>>> This is a really interesting idea...something like OpenID for authority information.
>>>
>> Actually, it's more like RDF. A name is assigned a URI (that may or
>> may not resolve as a URL) that no other name can occupy. OpenID
>> assumes that somebody can log in and use that identity on the web
>> (which is quite a different use case).
>>
>> If the URI is a URL, the authority record could be returned with a
>> stylesheet that produces to a browser something like a Worldcat
>> Identities page (http://orlabs.oclc.org/Identities/) or a marcxml
>> authority file to a machine.
>>
>> Given how our data can be (and should be!) spread among various
>> formats and services, RDF seems a very logical delivery mechanism for
>> our metadata.
>>
>> -Ross.
>>
>
>
--
Jonathan Rochkind
Sr. Programmer/Analyst
The Sheridan Libraries
Johns Hopkins University
410.516.8886
rochkind (at) jhu.edu
Received on Tue May 15 2007 - 07:50:38 EDT