On Jul 30, 2009, at 2:34 PM, Richard Light wrote:
In message <[email protected]>, Sean
Gillies <[email protected]> writes
There's much to like about CRM -- not the least of which is the
active and helpful community -- but I've concluded that it's overly
normalized for what I'm trying to do, which is to link data about
ancient inscriptions to data about places of the classical world.
Inscriptions are found or observed at places, but in the CRM this
relationship is always mediated by an event: an inscription is
discovered during a "finding event", which occurred at some place.
We are not ready to mint resources for all these events, most of
which will never be reused, and so we're bypassing and using non-
CRM properties to relate inscriptions and places.
Well, this is sort of where we came in - wanting an event ontology.
It's worth noting that within the CRM itself there is a practice of
declaring "short-cuts" to finesse over-complex relationships,
replacing them by a custom property (e.g. property "P130 shows
features of" is a short-cut for "P15 was influenced by" in the
context of an event).
Yes. "P62 depicts" is another example.
In the specific case you cite, the problem may be not so much to do
with the need to record events as to do with objects vs. places:
there is the property "P128 carries (is carried by)" which relates a
physical man-made thing to an information object. So inscriptions
on objects are catered for, but then you have to bring in additional
machinery to associate the object with the place where it was found,
which is what you are really interested in.
Richard
Well, I'm working with historians and epigraphers, so what we're
really interested in is where the inscription's ancient audience found
(and read), and to which ancient places an inscription attests.
To try to get back on topic: I'm very interested in ontologies for
ancient events, and even for modern events, to the extent that a dig
might be thought of as a long-duration event.
HEML is the other event model I know. Its author, Bruce Robertson,
writes about it at http://digitalhumanities.org/dhq/vol/
3/1/000026.html, including future plans for RDF. It's not unlike
Raimond's event model, but adds an "evidence" element.
Sean
