Format of tokens
Frederic V. Hessman
hessman at Astro.physik.Uni-Goettingen.de
Mon Nov 5 03:06:29 PST 2007
On Nov 2, 2007, Rob Seaman wrote:
> In fact, even if we replaced #astronomicaltoken with #1743, the new
> token would float to its own natural level of recognizability.
> Individuals (whether end-users or programmers) handling the token
> on a frequent basis would learn that #1743 meant "Supernova Ia" or
> "Gamma-ray Burst" or "Impact Crater".
We're not really expecting ordinary users to deal with any of these
dirty details - they get the preferred labels and maybe the
alternative labels, basta!
The vanilla flavored VO-programmers won't deal with any tokens
individually - they'll simply have to insure that the apps can skos,
re-skos, inter-skos, and de-skos and do all the other wonderful
semantic webish stuff, utterly independent of what the vocabularies
are and mean.
The only ones who really have to worry are the astronomers who have
to decide what concepts are needed for their own contexts (e.g.
VOEvent), whether it's easier to simply re-use the IVOA's thesaurus
or simply define their own (or both). They will not be using the
tokens but the prefLabels unless they are totally batty, since the
prefLabels - hopefully by definition - contain the fundamental
astronomical content. Since there are a few batty colleagues out
there, maybe minimially readable tokens are a nice gesture of
> The main issue is reaching a consensus quickly.
Amen. I believe we're just waiting for the completion of a simple
SKOS proposal which the IVOA is bound to accept (what else is there)?
> VOEvent could have used an officially blessed list of concepts two
> years ago. The alpha (perhaps not the omega) of the semantic web is
> in labeling objects according to a widely known vocabulary (or -
> ies). The only way to reach the far end of the Greek alphabet is to
> start at the beginning. Let's bless the SKOSified IAU Thesaurus,
Ack! Say you meant the new and improved IVOAT thesaurus - the old
IAU thesaurus is old-fashioned, highly incomplete and down-right
wrong in places. There is not a single advantage of using the IAU
over the IVOAT thesauri since the IVOA represents the IAU, so we can
quickly make IVOAT as official as the original. All we have to say
is that the old list has been modernized and made SKOS-compatible -
which is perfectly correct.
> add additional application specific word lists, and start tagging
> our varied and sundry data objects. Give it a couple of years and
> then, perhaps, we'll all find renewed interest in seeing what
> ontologies can bring to the IVOA Bar Mitzvah after a few guests
> have gathered.
I have suggested to the VOEvent community on repeated occations that
we quickly glance through the current list (I've already sent around
a good selection), pick out what one needs (a relatively simple
task), think about what's missing (a more painful task, but easy
relative to starting from scratch) and get started. As soon as
we've officially asked the IVOA to adopt SKOS as the minimally
conforming format, there's no reason why groups like VOEvent can't be
already past the starting-block stage. The selection of tokens/
labels needed is initially utterly independent of the vocabulary format.
For the benefit of our ontology colleagues, I've really been trying
to clean up the IAU ontology entries as I go along so that we can
have the best of both worlds: the starting vocabulary for Rob et al.
and basic ontological info for those who might be able to use it.
Thus, we don't have to wait for the Bar Mitzvah to see if our
colleagues can make a good ontological case - it would be good enough
to see some hot pastrami made out of the present vocabulary.
More information about the semantics