some remarks on VOEvent

[Rob Seaman]

Roy says:

> More properly we should *not* say VOEvent represents an  
> astrophysical event, but rather it represents an *observation* of  
> an event.

And "observation" is the default role to distinguish actual VOEvent  
packets from tests and predictions.  It's the word "event" that is  
shaky.  An astronomical event can extend vast distances and persist  
over immense periods of time.  The observations that we're tracing  
together as followup VOEvents may well involve detections of distinct  
physical effects arising from disjoint regions of space-time.  We can  
trust that causality ultimately ties it all together - but causality  
is inferred, not observed.

One could argue that the semantic structure of a VOEvent is on a more  
stable footing than the astronomical ontology we would hope to lean  
it upon.  My point?  My point is that a single static astronomical  
ontology is a vain hope - rather, we need a flexible mechanism for  
building and extending a rich network of context dependent ontologies  
as time and tide require.

That's actually a good example that may illuminate these issues.  It  
would be straightforward to describe a specific ocean tide  
measurement in a VOEvent packet.  These are certainly observations of  
astronomical events, after all.  Conjectures on tidal expressions in  
astronomical ontologies?  Will these support generalized central  
force representations?  And will they be expressed as spatial  
derivatives of Kepler's or Newton's laws?  Or, rather, directly as an  
inverse cube?

Rob