[Crm-sig] Colourful problem

Vladimir Alexiev vladimir.alexiev at ontotext.com
Wed Aug 13 20:48:37 EEST 2014


Hi Dan! Interesting question.

> Documents with applied seals... Beside the material of the seal, I want to specify its colour,

To my mind things like Seals and Ink Stamps are E25 Man-Made Features.
Since Man-Made Feature is a subclass of Man-Made Thing, you can still attach Material.
But there's a strong correlation between type and material: you can't make a relief seal from ink
So I think it's better to incorporate type and material into an enumeration, using a thesauris like:
  <thesaurus/marks/wax_seal>
  <thesaurus/marks/golden_seal>
  <thesaurus/marks/leaden_seal>
  <thesaurus/marks/leaden_seal_with_machined_stamp>
  <thesaurus/marks/ink_stamp>
  <thesaurus/marks/customs_banderol_sticker>
  <thesaurus/marks/brand_burned_on_cow> (I'm sure there's a proper term for that).
NOTE: E37 Mark is an information object so I should have used something like "seal" in the URLs above, but I'm too lazy to change that now..

In fact AAT defines some such hierarchy:
http://www.getty.edu/vow/AATHierarchy?find=&logic=AND&note=&subjectid=300200896
And  here is the sub-hierarchy of seals:
http://www.getty.edu/vow/AATHierarchy?find=&logic=AND&note=&subjectid=300028877 

> <mySeal [E84]> <P56_bears_feature> <myColour [E26_Physical_Feature]>

I don't think a Feature is something that covers the entirety of a Physical thing.
But I consulted the scope note of Feature. Relevant excerpts:
- "*may* have a one-, two- or three-dimensional geometric *extent*", i.e. not the whole thing. But not necessarily
- "*portions* of particular objects with partially imaginary borders"
- "can be .. scratches, holes, *reliefs*: so Seals are properly modeled as features, even if they have relief.
- "*surface colours*, reflection zones in an opal crystal": so colors *are* included. But I think "colored spots/zones" is meant here

> What's the right predicate (domain E26, range E55) to use, in order to say "myColour IS blood red" ?

I think an extension prop: P2X_has_color, subprop of P2_has_type

> using a suitable concept from the AAT Colour facet
> NB. I'm converting AAT too.

Well don't. Use http://vocab.getty.edu/. E.g. the above is http://vocab.getty.edu/aat/300310722 
- The old doc is here: http://www.getty.edu/research/tools/vocabularies/lod/aat_semantic_representation.pdf
- the new doc (coming in a week) is here: http://vladimiralexiev.github.io/aat/index.htm 
- also coming in a week is TGN

--

In summary, I would model it like this
  <document> P56_bears_feature <document/seal>.
  <document/seal> a E25_Man-Made_Feature;
    P2_has_type <thesaurus/marks/wax_seal>;
    P2X_has_color aat:300310722. # blood red

Or if you already have a Materials thesaurus, split it like this:
  <document> P56_bears_feature <document/seal>.
  <document/seal> a E25_Man-Made_Feature;
    P2_has_type <thesaurus/marks/seal>;
    P45_consists_of <thesaurus/materials/wax>;
    P2X_has_color aat:300310722. # blood red




More information about the Crm-sig mailing list