Rick wrote:
> I would have thought that a "TopConcept" is one which is at the top
> of a connection-hierarchy
...unless we just want to declare all entries to be narrower than "AstroThingy". Surely, beyond that it's turtles all the way up.
> If there is a concept "gummi bears" but no "BT candy" then the
> authors of the vocabulary have obviously left "candy" out for some
> reason, making "gummi bears" pretty top-level
Do we have a sense for whether there are any (significant) missing ur- concepts we should introduce to complete the hierarchy?
Let me just say here that a few of us from NVO had a very cordial meeting last week with the CBAT (literally the entire IAU Bureau). I mention this because pertinent to this discussion, the question came up of what to call the "Minor Planet Center" in an era sans "Minor Planets". I won't say I have an answer for you - but clearly VO personnel weren't the first to inquire...
Just to clarify my own understanding of the activity we're engaged in, a concept "MPC" would be neither NT or BT the concept of "Minor Planet", right? (Unless, I suppose, Massachusetts were an asteroid rather than a commonwealth.)
> Dropping the TopConcept links to entries with no NT's is trivial -
> is this the general consensus?
If I understand the suggestion, I concur. I wouldn't be surprised to find my understanding still to be devoid of "substance", however.
Alasdair says:
> [AG] In principle I agree that top level concepts should not have a
> broader term. However, we are trying to accurately model the IAU 1993
> thesaurus not correct it. Thus, the top level concepts should be those
> that appear in the top level in their hierarchy list, not what we feel
> should be a top level concept. We can correct this for the IVOAT :)
Not just accurately model, but coherently model. If the original has errors at the level of "mappings", not just concepts, we should consider correcting it. What this means to me would be for this WG to form a consensus of the best new-tech representation of the 90's thesaurus (where "best" is understood in a multi-parametric way), and then to present this to the original authors and the broader Comm. 5 for comment.