VOEvent v2.0, some thoughts

[Frederic V. Hessman]

On 21 Jan 2008, at 8:52 pm, Roy Williams wrote:

>>     2) support for IVO vocabularies (for <Why> and <What>)
>
> How can I say "supernova Ia" using one of the vocabularies?

How about any of the following, to be found in already existing SKOS  
vocabulary files:

	iau93:SUPERNOVAES						(1993 IAU thesaurus - available now)
	aoim:StarEvolutionaryStageSupernova			(AOIM working group thesaurus  
- available now)
	ivoat:typeIaSupernova						(proposed IVOA thesaurus - available now)
	ivoat:snIa									(Ibid., just an alternate label - available now)

  or

	voe:supernovaIa								(if Roy wants a native VOEvent token - TBD)

For example:

	iau93 : http://www.astro.physik.uni-goettingen.de/~hessman/rdf/IAU93/ 
IAU93.rdf
	ivoat:  http://www.astro.physik.uni-goettingen.de/~hessman/rdf/IVOAT/ 
IVOAT.rdf

and official copies will be available soon at a central IVOA URI as  
soon as the official draft of semantics working group is submitted.

Note that a human-readable short descriptive text (e.g. "Supernova of  
type Ia") is available for all of these tokens, so one can already  
use these terms in a real-live GUI.

> Do I have to import a schema?

No, SKOS/XML is basically pretty simple XML and you're welcome to  
parse the even simpler N3- or Turtle-formatted equivalent  
representations.

However, if you want to be exact, you'll need, e.g.

	<!ENTITY rdf "http://www.w3.org/1999/02/22-rdf-syntax-ns#">
	<!ENTITY rdfs "http://www.w3.org/2000/01/rdf-schema#">
	<!ENTITY dc "http://purl.org/dc/elements/1.1/">
	<!ENTITY dct "http://purl.org/dc/terms/">
	<!ENTITY skos "http://www.w3.org/2004/02/skos/core#">

(the "dc" things are Dublic blah-blah-blah doc stuff, if you also  
want to parse various other doc information).

> What would the <Why> section look like?
> In the <What> section, we are having trouble with those parameters  
> which are links to further information. We are mashing up VOEvent  
> feeds for Google Sky (voeventnet.org), and each placemark on the  
> sky may have links to finding chart, event portfolio, photometry  
> file, etc, which should be linked from that placemark. But each  
> event stream uses different types of Params or References for these.

Good question.   Given that we are pushing the current limits of  
standard semantic-web technology (lots of ideas out there, but few  
standards), I'm afraid that there aren't too many tools for actually  
USING the vocabularies in the sense of being able to say "the event  
is either an X or a Y plus Z but definitely not a W" in a standard  
syntax, much less Roy's ultimate test case, "the eclipse occured when  
a G3V star went in front of the K3 giant".

Within the current VOEvent <param> paradigm, we could easily and  
quickly use something like (I'm brainstorming here for didactic  
purposes only...)

	<group name="object">
		<param name="object-type" rdf:resource="iau93:SUPERNOVAS"/>
		...
	</group>
	<group name="telescope" rdf:resource="ivoat:telescope">
		<param name="focal-length" rdf:resource="ucd:instr.tel.focallength"  
value="8770" units="millimeters"/>
		...
	</group>
	...

The use of rdf:resource attributes makes the contents semantic-web- 
compatible, since each of the entries can be identified by someone  
with little knowledge of what the document actually means or is  
intended with a global vocabulary, so that the contents of different  
documents can be compared/linked just on this basis.  Note that I  
slipped in a reference to UCD expressed in terms of RDF.  Note also  
that one can mix-and-match, according to your needs, i.e. no more  
monolithic "UCD and that's it".

If the VOEvent working group can figure out what terms are
	1) needed, and
	2) not yet present in an existant vocabulary,
then we can produce a special VOEvent vocabulary with nearly zero  
effort (after the terms have been identified).

Try http://www.astro.physik.uni-goettingen.de/~hessman/rdf/

Rick