On 23/7/09 11:07, Peter Mika wrote:
peter - would you share those publicly, please?
Sure, here is my cost/benefit analysis on tel as a resource:
Benefits:
-- Slightly easier data integration, e.g. using SPARQL queries. However,
how many people are doing data integration using SPARQL alone?
-- We would like to be compatible with the ontology... (or should the
ontology be changed?)
Costs:
-- Gives the illusion of a resource that you can dereference. Tom Heath
these days is on the road with an excellent Linked Data presentation
that explicitly advises against using non-http URIs.
-- There is not much anyone would ever want to say about a phone number,
which would be the most common reason for making something a resource.
-- Sites owner are expected to read an RFC on how to write down a
telephone number, and then figure out the transformation from their
internal representation to the scheme. Not likely to happen...
-- Search engines index URIs differently than literals or not at all. In
this case, this behaves as a literal in that I want it to be indexed.
Also consider recent changes to vCard underway at IETF: see
http://danbri.org/words/2008/06/25/348 for a summary.
Latest seems to be
http://www.ietf.org/id/draft-ietf-vcarddav-vcardrev-08.txt
"""7.4. Communications Properties
These properties are concerned with information associated with the
way communications with the object the vCard represents are carried
out.
7.4.1. TEL
Purpose: To specify the telephone number for telephony communication
with the object the vCard represents.
Value type: A single URI value. It is expected that the URI scheme
will be "tel", as specified in [RFC3966], but other schemes MAY be
used.
"""
Mention is also made of the mailto: URI scheme (surely this is still ok
to use, privacy issues aside), and a "geo" URI scheme
[I-D.mayrhofer-geo-uri] that I don't know much about.
If the goal of this vocabulary is to reflect the IETF vCard vocab,
keeping close to trends in vCard-land might be prudent...
cheers,
Dan
