Fwd: Light curves in the IVOA

[Roy Williams]

On 6/22/2010 2:42 PM, Doug Tody wrote:
> Another point - The problem here is really time series, of which a
> light curve is only one example.  Of course it is the most common
> one, but in general time series analysis and time series data in an
> archive can have anything on the Y axis.  The UCD can say what this
> quantity is.  The same is true for spectra, and the model already
> supports this.  (Hence for example SLiCAP is not the right name for
> the general time series interface).
>
> In some specific context such as a VOEvent it might be appropriate
> to restrict the data to a simple light curve, however this is just
> a matter of subsetting the model.

Doug

Of course I agree that a light curve inherits from time series, and that 
a general solution has more applications. However, a directed, specific 
solution might have a better chance of survival and propagating in the 
astronomical community, something small and agile. I suggest that first 
we make a light-curve specific protocol based on cones in the sky. This 
is what will get the fire burning, and people interoperating.

Of course the more general protocol is in our minds behind the curtain, 
to be standardized later, once there is an appetite for it. We know the 
general shape of it, and how the light-curve protocol will fit as a 
subset. I believe that a initial focus on optical light curves will get 
a lot of astronomy done with a simple protocol. There is a pent-up 
demand for this already.

In other words -- we make the subset a Rec before the whole enchilada 
goes through.

Roy



-- 
California Institute of Technology
626 395 3670