VOEvent, unique id's, and IVORNs

[Alasdair Allan]

All,

Unless I'm mistaken we've all been generating VOEvent messages with 
invalid IVORNs, the stuff we're sticking into the <VOEvent id="ivo:/*"> 
attribute doesn't seem to conform to what's actually supposed to be 
there.

For example lets take the 3 existing feeds. The Caltech feed seems to 
have something along the lines of,

   <VOEvent id="ivo://gcn.gsfc/HETE_SC_Alert_Source/4006-1"

the RAPTOR/TALONS feed has,

   <VOEvent id="ivo://gcn.gsfc/swift/17766663a"

while the eSTAR feed has something like,

   <VOEvent id="ivo://estar.ex/test/7046.81"

I'm at the AstroGrid Technical Workshop this week, and according to the 
great and the good we should be shipping messages where the id="" 
attribute contains something like this...

   <VOEvent id="ivo://AUTHORITY ID/PUBLISHER#UNIQUE ID"

So the equivalent (valid) eSTAR id (equivalent the the example above) 
would have to be something like this,

   <VOEvent id="ivo://uk.org.estar/estar.ex#test/7046.81"

and the Caltech & RAPTOR ones would be something like,

   <VOEvent id="ivo://edu.caltech/gcn.gsfc#HETE_SC_Alert_Source/4006-1"
   <VOEvent id="ivo://gov.lanl/gcn.gsfc#swift/17766663a"

The important point here is that the authority ID

   ivo://uk.org.estar

(for example) should be registered as an authority resource in the 
registry, and is therefore a namespace, and the corresponding publisher 
ID

   ivo://uk.org.estar/estar.ex

should be in the registry as an organisation resource.

In other words our IVORNs are all malformed (forget about the fact that 
we haven't registered anything with a registry, that's not relevant, 
right now we can't even when we get round to that because we're not 
doing the authority IDs correctly)...

We should probably fix this before we get much further. I'll try and 
move eSTAR to the "proper" way of doing things when I get back to 
Exeter next week.

Cheers,
Al.