[Crm-sig] disjointness (E39.Actor vs E78.Collection vs E53_Place) and saving on nodes

Vladimir Alexiev vladimir.alexiev at ontotext.com
Thu Nov 24 12:04:39 EET 2011


> From: Vladimir Ivanov [mailto:nomemm at gmail.com]

Thanks for your feedback! 
(My middle name is "Ivanov" so I was like "am I speaking to myself??" :-)

Another heretical thought: I'm thinking of also making it E53_Place that P89_falls_within rkd-place:TheHague,
so that the "thing from place" search in our system can also search by collection (for free).

With Exhibit faceting it could look like this 
(Exhibit doesn't have hierarchical facets but maybe we can fool it to indent):
  10 Netherlands
     5 Amsterdam
       3 Rijkmuseum
     5 The Hague
       5 Mauritshuis
  3 France
    3 Paris
      3 Napoleon Museum
Nice, eh? 

How come the 3 paintings in Rijkmuseum increase to 5 paintings in Amsterdam?
Well, maybe 2 of the paintings that are *currently* in The Hague were *made in* Amsterdam,
and our definition of "thing from place" includes both of these relations.

Note: this stuff is based on M.Doerr's New Framework for Querying Semantic Networks (FORTH TR419 2011)

> scope notes of the mentioned classes (E39.Actor, E78.Collection)
> may contain contradicting statements,

You're right. So we'd need 3 nodes for each museum:
- the legal organization
- the collection 
- the place

But looking pragmatically, my user doesn't care about these semantic distinctions.
To an art researcher a museum is a museum. I think that he'd rather see:
- the has_type label (e.g. "public museum", "private collection") rather than the CRM class(es)
- all properties of the museum bunched together, rather than an extra level of nesting/navigation

M.Doerr himself is a bit of a heretic :-) 
Tuesday, November 22, 2011 7:21 PM:
>Other combinations might in practice be used in a non-disjoint manner, depending on the
>context, because of a 1-1 relationship... 
>one may like to get rid of Time-Spans. Why not. Saves a lot of identifiers and joins

Right on! 
We can't tolerate blank nodes in our system, since that'd hurt Annotation of statements under implementation Alternatives 1&2
(see http://personal.sirma.bg/vladimir/crm/art/PropertyTypesAndAnnotations.html#sec-2_2_ and my email Mon 10/24/2011 11:25 PM).
So I've had to introduce measly nodes like this (/person, /person/name, /date and /id), and I sort of hate it:

<obj/2926/part/2/production> crm:P14_carried_out_by <obj/2926/part/2/production/person>.
<obj/2926/part/2/production/person> a crm:E21_Person; crm:P131_is_identified_by <obj/2926/part/2/production/person/name>.
<obj/2926/part/2/production/person/name> crm:P3_has_note "Willem de Vries".

<obj/2926/part/1/production> crm:P4_has_time-span <obj/2926/part/1/production/date>.
<obj/2926/part/1/production/date> crm:P82_at_some_time_within "1636"^^xsd:gYear.

<obj/2926/file/6> crm:P48_has_preferred_identifier <obj/2926/file/6/id>.
<obj/2926/file/6/id> crm:P3_has_note "549".





More information about the Crm-sig mailing list