On Tue, Apr 27, 2010 at 06:56:04PM +0200, Bernard Vatant wrote:
> We are currently revisiting the Geonames ontology (long overdue) including
> feature codes, on which I would like to add some Dublin-Core like
> annotations re. concepts. In particular some concepts are going to be
> deprecated using dcterms:isReplacedBy, some will be deprecated altogether
> without replace because they've never be used, and quite a lot will be put
> in a status of quasi-obsolescence or quasi-extinction. Those are feature
> codes used less than one per million features, which means less than 7 times
> considering Geonames has over 7 millions features in store ...
> I was wondering which vocabulary is the best for expressing that.
> http://www.w3.org/2003/06/sw-vocab-status/ns is used by FOAF, Bibontology
> and others. But it defines status as a data property, I would prefer an
> object property with an enumerated range of individual values.
Yes, I see advantages to that.
> Of course I
> can define those in geonames data space, but since this is not particular to
> geographical codes, I was wondering if there was some vocabulary on some
> shelves I miss. BTW I was amazed not to find any "status" property in Dublin
> Core.
Historically, DCMI Metadata Terms began as a series of
documents (an RFC [1], then a DCMI Web page), and those
documents had status as documents - e.g., "informational"
as defined by IETF [1], "DCMI Recommendation" for [2]).
This information was not expressed in the corresponding RDF
schemas (e.g., [3]).
Properties and classes officially came into being
and were assigned URIs when they were published in a "DCMI
Recommendation" document -- i.e., they effectively derived
their "status" from the status of the Web document in which
they were published. The DCMI Namespace Policy namespace
policy [5], then, made persistence guarantees for URIs, once
declared, but without providing for lesser statuses, such as
"deprecated".
DCMI Metadata Terms never followed the style of FOAF (and
other vocabularies) of coining experimental terms, giving
them progressively higher status as they were implemented,
and eventually possibly deprecating them -- the terms either
existed (and were "recommended") or not.
DCMI invented its process year by year, and if emerging good
practice now dictates that it is good to assign status to
individual terms, I should think we'd want to help standardize
such a status vocabulary, whether in a DCMI namespace or
elsewhere.
Tom
[1] http://www.ietf.org/rfc/rfc2413.txt
[2] http://dublincore.org/documents/2000/07/11/dcmes-qualifiers/
[3] http://dublincore.org/2001/08/14/dces
[4] http://dublincore.org/documents/dcmi-namespace/
[5] http://dublincore.org/documents/2001/10/26/dcmi-namespace/
--
Thomas Baker <[log in to unmask]>
|