Why archive packets?

[Roy Williams]

Thank you Rob for starting this discussion.
>> Perhaps the place to start is therefore, why are we archiving VOEvent 
>> messages in the first place? I'm not arguing we shouldn't, but I 
>> think we need to be clear about the reasons.

I would like to add correlation of event streams as a driver for 
archiving events. It is like cross-match except in time. Perhaps the 
query interface should look like other cross-match tools. Some snip 
below about event correlation.....

New science will come from the archival correlation of multiple event 
streams. Because of the significance of correlation, we should encourage 
authors to publish even marginal events (e.g., “two-sigma” detections 
that would otherwise go unreported), since these may gain great value in 
correlation studies: often heralded in this respect is the detection of 
an otherwise marginal gravity wave event coincident in time with a 
gamma-ray burst. Another example: a flare at the center of a distant 
galaxy might be classified in an optical survey as normal active nucleus 
activity but instead, with concurrent X-ray and UV data from space, 
could be recognized as rare tidal disruption of a star, giving direct 
information about black hole growth (GALEX, in conjunction with a 
supernova survey, recently found such an event). A marginal detection of 
a neutrino event from Super-Kamiokande could be bolstered by a 
concurrent (yet unplanned) observation of a transient at another 
wavelength. Likewise, understanding the origin and nature of 
gravitational wave detections from LIGO-II and LISA will benefit from a 
federation of transient event data.