Skip to content

Conversation

@EOltmanns
Copy link

Even though the domain of the property hasEvent includes items,
manifestations and work/variants, the Cataloguing Manual provides a
number of strict guidelines associating certain subclasses of Event
with instances of Manifestation only, others with instances of Item
only, and so on. Similarly for the property hasActivity.

This pull request tries to make the ontology reflect those guidelines
more accurately. The Cataloguing Manual appears to be far less
specific in other respects. For instance, the same values for colour
or sound characteristic may be recorded for manifestations or items.
Accordingly, the ontology cannot impose further restrictions on the
hasColourCharacteristic property. In fact, I have reviewed the list of
properties whose domain consists of a union of classes, and I was
unable to identify any property beside hasEvent and hasActivity that
comes with further "obvious" restrictions according to the Cataloguing
Manual.

On the other hand, the domain of the hasEventLocation property is
defined as Event even though the cited section of the manual (D.4.1)
seems to relate locations to publication events only. That is why I
have restricted hasEventLocation to PublicationEvent. If you have
discussed this before and deliberately chosen to define the domain as
Event rather than PublicationEvent, I can remove this change from the
pull request again. Generally speaking, there may be more properties
whose domain consists of a single class that may or even should be
replaced by a more specific subclass; naturally, they are hard to find
and therefore out of scope for this pull request.

EOltmanns added 5 commits May 31, 2023 22:20
State explicitly whether a given event type relates to Item,
Manifestation, or WorkVariant.
State explicitly whether a given activity type relates to Item,
Manifestation, or WorkVariant.
Section D.4.1 of the Cataloguing Manual seems to associate locations
with publication events only. Broadening the domain may not do much
harm or even be desirable but the decision (and reasoning) should
perhaps be documented somewhere. At any rate, broadening the domain in
a later version of the ontology would preserve backwards
compatibility, whereas restricting it from Event to PublicationEvent
later would break compatibility.
@EOltmanns
Copy link
Author

I forgot to mention that I declared WorkVariant, Manifestation and
Item as disjoint classes. It seems to be sensible to me but is rather
complementary to the rest and certainly up for debate. I can easily
drop that part if you would rather have it that way (see the commit
log).

@EOltmanns
Copy link
Author

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant