> Hmm. I should have raised this before, but what about the IAU
> Thesaurus? Even if its not supported since mid-90's its pretty
> comprehensive. It even has some files available online from which
> the project may be jump-started.
> So.. with these machine parse-able text files available, which
> seem to supply
> both multilingual aliases, and hierarchy of terms, and the bonus
> of this being
> as 'official' a source as one might hope to encounter, I would
> find it a difficult
> argument to NOT use this.
The main objection to avid adoption to date has been (at least my understanding)
> Perhaps there are too many terms? In this case then, why not
> simply cherry
> pick terms? Clearly additional terms which have come into use since
> the last publication are going to be needed, alongside some
> bridging terms,
> but you can hardly go wrong with using this source, and I would urge
> not (re-)inventing stuff that is already in there. Perhaps the
> work in the
> recent draft is already based on the Thesaurus (mea culpa: I
> haven't had
> time to read it), kudos if this is the case.
Then a few (but not too many) kudos are due :-)
> I also want to add that while machine-understandability is the
> definite
> goal of such a dictionary, it should also have human-readable
> definitions
> for the terms embedded in the document. Without this kind of
> 'self-documentation' you eventually get into problems of what the
> original meaning of the term was supposed to be (some terms
> will be ambiguous). Again, the IAU thesaurus is invaluable here,
> as it
> already supplies these definitions, and it might be enough to put in
> an annotation pointing to that term in the published work.
There are no definitions, just translations of the labels. Free multi-linguality is cute, but not really the point right now.
The amount of NT's and BT's in the thesaurus is a reasonable model of the level of detail I'd like to see in the vocabularies - just enough to make things useful but so very little that one isn't tempted to declare it a real ontology. Rather less would be just right.
Rick
Dr. Frederic V. Hessman Hessman-at-Astro.physik.Uni-Goettingen.DE Institut für Astrophysik Tel. +49-551-39-5052 Friedrich-Hund-Platz 1 Fax +49-551-39-5043 37077 Goettingen Room F04-133http://www.Astro.physik.Uni-Goettingen.de/~hessman
http://monet.Uni-Goettingen.de ------------------------------------------------------------------------ -------------------------Received on 2007-09-19Z19:05:28