VOEvent priorities

[Andrew Drake]

Hi Rob,

On Wed, 3 Feb 2010, Rob Seaman wrote:
>> While I certainly agree there is a need for more work on time series 
>> representation within the VO, I am still unsure why "VOEvent needs time 
>> series".
>
> The general response is because astronomers often include time series in 
> their "telegrams" and VOEvent seeks the capability of representing diverse 
> sorts of such messages.

Yes, CBAT certainly encourages a reference magnitude or two. However, I
highly doubt they would publish the hundreds of points I could send them. 
Likewise, time series data is not very common in ATels since the number of
characters one can include is limited. In most cases people tend to refer 
to external data if people want to include time series or additional data.

> I welcome specific examples from Berkeley and others of how SimpleTimeSeries 
> has been and will be used.

So do I.

> However, as chair I'm a little unsure how to respond to a request to revisit 
> a group decision made more than two years ago at HTU-1.  Has something 
> changed to cause us to view the question differently?

I don't remember this either. Perhaps somebody else will.

>> Secondly, is there any size limit on the time series that one can include 
>> within a VOEvent?
>
> Is there a size limit on what you can handle?
> If a size limit is needed on a packet, we can legislate this - although I
> don't know if XML schema supports such a concept.

I do enforce VOEvent packet size limit, as do others.
However, this is a broader question which I thought we may want to 
consider in the context of large surveys with hundreds of data points
per event, thousands of events per night, and hundreds of customers
(many of which will likely have slow and unreliable connections).
Perhaps we could just set some guidelines.

cheers,
        Andrew
--