Vocab AND Ontology?
thomas at astro.umd.edu
Tue Oct 9 09:13:39 PDT 2007
On Tuesday 09 October 2007, Frederic V. Hessman wrote:
> By and large a simple exercise, particularly since many already exist, but one sees how this process has to
> asymptotically converge BY DEFINITION, not by time or effort. May I suggest that the state of the Vocabulary/Thesaurus
> be fixed by the end of this week - last chance for last-minute additions! Whatever you can't find after that must not
> have been very important.
Not sure I agree that we have just a week to finish additions/corrections!
I would think that we should simply version the Thesaurus. Your call for
and end to the process should be for a given version (whatever that may be,
v1.0, 0.1, 0.001?). Expect and plan for an ongoing process of improvement
> Rob's suggestion of splitting things into the original thesaurus (or at least a corrected version of the old thesaurus -
> mixed cases, underbars, and lots of tokens shifted into ALTs) and new additions would be a pain, given the amount
> of cleaning up we've done, but if ya'll want it that way and as long as it means that the WG makes a decision about
> how to handle multiple vocabularies: ivorn formats, normative file formats, suggested translation infrastructure (e.g.
> references to math:ellipses). If this will take too long, I vote for one giant starting vocabulary based upon the IAU
> but with all the present corrections/additions.
It may be a pain, but its worth doing IMO. At the least, I continue to vote for breaking the thesaurus up into
sub-parts by category. I peeked into Ed's office this morning and was happy to see that he was doing
this with a 'shadow' ontology based on the Thesaurus.
> Wait a minute - a great idea: wouldn't the IAU/IVOA thesaurus be a GREAT place to put all the STC shortcuts?!
> How else are we to know where to find them? Or create a STC shortcut vocabulary as a new, practical, and
> exemplary example.
I'm not so sure, and at this point I would 'vote NO'. The thesaurus is quite large as it is, and really should
focus on an 'updated IAU' set of terms. STC encompasses a number of structures which are (IMO) outside
of time/space (it has hooks which are for integration with VOTable, and its own form of a Quantity + Units
to name a few).
STC should have its own, separate ontology, although I would hope that it would re-use terms from the
More information about the semantics