The Napkin Representation

[Rob Seaman]

Note that I've limited the distribution back to DM.  The VOEvent  
folks who are interested should be listening over here now.

On Jul 3, 2007, at 5:52 AM, Silvia Dalla wrote:

> 1. Is this representation aimed at Time/Flux time series only,
> or should it be catering for time series of other quantities too?

Other quantities, too, but I don't see any problem as long as we  
stick to scalars.

> (orbits of objects have been mentioned by others earlier).

That's a separate discussion.  I'd say we're 90% of the way to a  
consensus of supporting orbital elements using STC.  Individual  
observations of asteroids often accompany the orbital elements,  
however, and the former would still be a target for time series  
expressions.

> 2. The simplest time series table consists of 2 columns only: one
> giving the time and one giving the observable that is varying in time.

The simplest is one column, with a time grid expressed parametrically  
as a header attribute.  A theoretical time series might even  
parameterize the dependent variable, too.

> Is the new representation aimed only at this very simple time series?
> Or can the representation describe more complicated time series tables
> that have a time column and several columns for different observables?

Most observational data will have one observable per time value, but  
may collect several observables into a single time series, which is  
to say that few time steps may correspond to two or more measurements.

> For those that were not at the HTN meeting it would be useful to know
> what are the objections against using a VOtable representation.

Mostly that we don't want to make our schema dependent on the varying  
quirks of all other VO schema.  This issue may be neatly finessed by  
Doug's suggestion of referring to a separate object in the same  
bundle.  Such is not currently supported.

> there are several VOtable parsers that work and applications that  
> know how to deal with VOTables.

VOEvent is a publish/subscribe paradigm.  The "publish" part of that  
is quite literal - we reach out of the VO into the larger community  
as a permanent record of assertions made by astronomers (through  
various instrumentalities).  A particular project or institution may  
choose not to rely on standard parsers or applications.  This doesn't  
mean that we want to avoid interesting VO technologies - but we do  
want to minimize their complexity for our purposes.

Rob