VOEvent priorities

[Joshua Bloom]

I agree with Andrew and Andy here, that large numbers of parameters might be taxing.

The current version of version of SimpleTimeseries allows for each time entry to be a pointer to another VOEvent, which Andrew notes could contain a bunch of info relevant for that specific "event".

So...at what point (datarate-wise or N parameters-wise) is pushing around full sets of timeseries from LSST or ALMA (or any data for that matter) starting to get silly?

Josh


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



On Feb 4, 2010, at 4:17 PM, Andrew Drake wrote:

> 
> Hi Andy,
> 
> I think that we can deal with small static values of N as
> we both have in past VOEvents (with groups and params).
> It seems that we need to decide is whether there a range where
> N (params, measurements, characters, bytes, etc) that is better
> contained within the VOEvents than referenced from them. For
> event sources (or follow-up) without any secure and robust
> repository it certainly seems like a good idea to include such
> info. For very large N this would seem to create additional
> difficulties. There is also the question of whether we should
> actually be thinking beyond the SimpleTimeSeries data model for
> all the other kinds of tabulated data that are related to events
> (perhaps not full sodsets). I expect that VOEvent v3.0 could
> well address issues relating to additional types of tables.
> 
> cheers,
>       Andrew
> -------------
> On Thu, 4 Feb 2010, Andy Becker wrote:
> 
>> Hi - At least for LSST, I could envision us publishing in the event the photometry from our 2 back-to-back exposures.  If you ascribe to the zero-one-infinity rule of thumb, this calls for some sort of time series to represent these N>1 blocks of information.
>> 
>> Andy
>> 
>> 
>> Andrew Drake wrote:
>>> Hi Rob,
>>> On Thu, 4 Feb 2010, Rob Seaman wrote:
>>>> On Feb 3, 2010, at 8:11 PM, Andrew Drake wrote:
>>>>> Your comment above seemed clear, "astronomers often include time series in their "telegrams"".
>>>> That also includes people on this group.  LSST is only one of the upcoming projects that - in effect - will be taking movies of the sky.  Those movies will often be expressed as simple time series of some object or another.
>>> I wasn't aware that LSST had agreed to use this particular representation of a simple time series. If they do want to include their full time series
>>> within events, rather than simply link to it, we should certainly ponder how this amount data can be handled within the VOEvent network.
>>>>> Based on Roy's limited results I would think that this is not often the case. Rather, even
>>>>> though it is possible to include time series in ATels, it is uncommon.
>>>> There are many kinds of telegrams: http://bit.ly/a7tpoL
>>> There certainly are still sources of VOEvents to tap. Even though
>>> most of the sources on this slide do not issue telegrams or VOEvents,
>>> and a couple have long since closed shop, including time series would
>>> certainly be useful if attracts a few more groups to use VOEvents.
>>> Even though most groups create links to their time series data, the
>>> amateur astronomers often do include small pieces of time series
>>> data in their emails to VSNET and AAVSO.
>>> cheers,
>>>       Andrew
>> 
>> 
> 
> -- 
>