uscholdm
commented
6 years ago See inline for a few quick reactions…
From: Dan Carey [mailto:notifications@github.com] Sent: Thursday, March 15, 2018 10:23 AM To: semanticarts/gist gist@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: [semanticarts/gist] Tighten definition of :Landmark (#26)
Currently, :Landmark is defined as "Something permanently attached to the Earth." Is a mountain a landmark or a geophysical region? MU: I would say ontologically, a landmark geoOccupies a GeoRegion. The mountain is a landmark, in topo ontologies, they call them physical features, and they include rivers, mountains cliffs, etc. This may or may not be important for the business.
I am inclined to put it in the latter category because similar physical features (rivers, canyons, bays, etc) are so extensive as to not fit readily into the former, nor are they "permanently attached" as anyone can attest who has seen a river shift its banks.
MU: well, buildings and fire hydrants don’t fare well with sinkholes or earthquakes, I think you could ignore that edge case.
We can draw a more distinct line between landmarks and geophysical regions by introducing the notion of something man-made.
MU: except that mountains (and big trees) do serve well as landmarks. You can always subclass to be only human-made ones.
I propose the :Landmark definition be changed to "A human-made construct permanently attached to the Earth."
Semantically, that does exclude, say, a giant redwood tree from from the class. But there's no underlying restriction that would prevent someone asserting such a tree was a landmark.
MU: true, but that tree would be inferred to be human-made due to the necessary restriction. Might not be a good thing.
And off the top of my head, I couldn't think of a good way of limiting the scope of a natural, living "landmark". Some coral reefs might be small enough to be considered landmarks; but some are so large as to more properly be viewed as geophysical regions. And in a business-oriented ontology, such edge cases are unlikely to arise.
MU: after all this, my gut says you can more or less leave it as it is. I don’t see a problem.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/semanticarts/gist/issues/26, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AKRzyRdNQFc8s_ziEZ6vFYdiQZzO7HVSks5teqODgaJpZM4SskLB.
mkumba
Currently, :Landmark is defined as "Something permanently attached to the Earth."
Is a mountain a landmark or a geophysical region?
I am inclined to put it in the latter category because similar physical features (rivers, canyons, bays, etc) are so extensive as to not fit readily into the former, nor are they "permanently attached" as anyone can attest who has seen a river shift its banks.
We can draw a more distinct line between landmarks and geophysical regions by introducing the notion of something man-made. I propose the :Landmark definition be changed to "A human-made construct permanently attached to the Earth." Thus, an airport could be classed as a landmark (and/or a georegion).
Semantically, that does exclude, say, a giant redwood tree from from the class. But there's no underlying restriction that would prevent someone asserting such a tree was a landmark. And off the top of my head, I couldn't think of a good way of limiting the scope of a natural, living "landmark". Some coral reefs might be small enough to be considered landmarks; but some are so large as to more properly be viewed as geophysical regions. And in a business-oriented ontology, such edge cases are unlikely to arise.