Frederic V. Hessman
Hessman at Astro.physik.Uni-Goettingen.DE
Mon Mar 21 03:44:31 PDT 2011
Dear Norman et al.,
On second thought, put back the corks.....
> On reflection, I think that @sciencetype (for the meaning of the
> information) and @mimetype (for the format) might be better
> attribute names.
The "mimetype" designation has the great beauty of being simple and
compliant with the rest of the world, but I imagine that there might
be cases where non-standard content may need to be shipped.
"sciencetype" also sounds ok, but maybe we can do even better. I
remember the initial discussion at CalTech when we came up with
<What>, <WhereWhen>, <Why> and how wonderful it was to simply say what
one means. If the "type" attribute is supposed to say what the format
of the dataq are in and the semantic content of the reference is
supposed to say what the reference "means", then may I suggest we
considering the following brainstorming examples which will NOT work
swift#SwiftByteFormatStream" meaning="ivoat:telemetry" />
skos#CelestiaObjectDescriptionFile" meaning="ivoat:simulation" />
" meaning="ivoat:survey" />
Better to have a general-purpose means of expressing any kind of
format (which is also just a vocabulary) and a basic set of examples
pre-defined by the IVOA for convenience use by those less-inclined to
invent their own.
Thus, in the true VOEvent tradition of "make it simple" and "say what
you mean", I vote for
<Reference format="format-vocabulary-term" meaning="semantic-
vocabulary-term">Put your ignored text here.</Reference>
P.S. Oh yes, one IS allowed to include several <Reference>'s anywhere
one needs them, isn't one.....? E.g. finder chart reference plus
light curve reference. I still don't like having just one semantic
meaning (how do I say "spectrum of a galaxy"?), but don't want to open
that can of attribute versus child-tag worms.
More information about the semantics