VOEvent priorities

[Joshua Bloom]

Roy-

 The idea of timeseries is that it has a time axis. We should be able to represent gravity chirp signals, sn light curves, pulsar pulse shapes, etc. But it's certainly not a kitchen sink on purpose. Those examples you propose don't pass the smell test of being a time series.

 Recall that we are also hoping (eventually) for a concept of a "source" which could be a bunch of those things you mentioned that are not a timeseries per se. I think this is like a flat single-file notion of your portfolio idea (as a directory). The difference is between keynote '08 and ppt/keynote '09 (presentation as a directory and presentation as a big file).

J

****************************************
Joshua Bloom
Associate Professor
UC Berkeley, Astronomy
510-643-4621 (Lab)
510-643-3839 (Office)
*****************************************



On Feb 2, 2010, at 10:18 AM, Roy Williams wrote:

> 
> Rob Seaman wrote:
>> V2.0 is defined by its support for time series ... this support will be provided by ... the Berkeley SimpleTimeSeries
> 
> Are we restricting ourselves here to light curves in optical wavelengths? Hope not! So I am assuming that a "time series" is a set of records labeled by time.  Each record can be strings and numbers and everything, all with semantics and units, right?
> 
> With the Berkeley model, can we express, for example:
> -- An ephemeris or other set of astrometric observations
> -- A set of time-stamped URLs that link to images of past observations
> -- Global banana production over the last 10 years ....
> 
> What if I want to express a set of records that are *not* labeled by time. Can we just set the time to zero or null? I am thinking of:
> -- A frequency series or SED of the object
> -- A set of objects near to the transient, i.e. a finding chart
> -- A set of people who are interested/observing/commented on the transient
> 
> Roy
> 
> -- 
> 
> California Institute of Technology
> 626 395 3670
>