Querying Params

[Roy Williams]

First I would like to thank Matthew for his excellent work in pushing 
the SEAP document!

Just to follow up in my own words too. What surprised me about the spec 
is that the query is not just for events where the Params satisfy a 
criterion (eg Magnitude < 18), but also for events that simple *have* a 
given Param (eg those with Param called "Magnitude"). I do not 
understand the use-case for asking if there is a specific Param name in 
the event. It is like a SQL query which asks for records with X>3 and 
also for records from any table that has a column called Y. But somebody 
must have requested this at some time, so it must stay I suppose.

The units thing seems even more extreme, as in "Give me all the events 
that have a Param whose unit is centimeter". I cannot imagine why such a 
thing would be wanted, even given the confusion over unit naming -- are 
those with "centimeter" the same as those with "cm"?

The same kind of difficulty comes in searching against UCD: if I ask for 
events which have a Param whose UCD is "phot.flux" (photon flux) should 
it also let through a variation like "phot.flux.sb" (surface brightness) 
or "phot.flux.density.sb" (Flux density surface brightness)?

Roy


Matthew Graham wrote:
> Hi,
>
> Roy has passed some comments about the current SEAP spec onto me and 
> there is one issue that needs discussion/resolution at the WG level.
>
> In the current (1.1) VOEvent model, a Param can have a name, a value, 
> a unit and a UCD. There have been previous requests to be able to 
> query against all of these (although UCD got omitted due an oversight) 
> but Roy would rather that we just limit to name and value and drop 
> unit and UCD.
>
> This would mean that you could not search for events with a parameter 
> whose values are expressed in a particular unit or where the parameter 
> has a particular UCD.
>
> The decision therefore is:
>
> (1) Maintain the current functionality which allows querying against 
> name, value, unit and UCD
>
> (2) Drop the ability to query against unit and UCD
>
> Please comment, discuss or just express which option you prefer.
>
>     Cheers,
>
>     Matthew
>

-- 

California Institute of Technology
626 395 3670