VOEvent References

[Mike Fitzpatrick]

Finally, an admission that VOEvent is a completely ad-hoc collection of
developers agreeing on what to do in the absence of a standard 8-)

Where your description of utility breaks down is in 5-10 years when the
satellite making the observations is dead and the people on the project
have moved on to derivatives trading in the new world economy.  All we
have at that point is the XML event packet itself sitting in a repository
and some poor grad student trying to write a thesis, or some app processing
a current stream and mining old repository events to create Nobel science.
If we have to rely on what the humans who created the packet *meant to say*
we may be well screwed.  In the absence of the original Mayan High Priest,
how do we know the original voevent carvings they left really mean we're
all
dead next year?

Packets have two clear audiences:  humans like the Keck telescope scientist
that might appreciate seeing a PNG of a possible GRB on their cell phone to
decide whether to slew the telescope, and software that needs other forms
of data to make automated decisions and can make use of that data.  For
presentation purposes, clearly only *some* information in a packet will be
useful to humans, and some will be needed by machines only. The XML doc
standard should accommodate both uses (and others), not simply declare
itself to be pragmatic and machine-readable when there are examples that
still require humans for events to be usable.

Item 1 on the Naples agenda clearly needs to be a way to keep the
opinionated
trouble-makers off the mailing list.

Cheers,
-Mike



On Wed, Mar 16, 2011 at 2:58 PM, Roy Williams <roy.williams at ligo.org> wrote:

>
>
> On 03/16/2011 2:04 PM, Matthew Graham wrote:
>
>> Get the humans out of the loop as much as possible - they can go "ah"
>> and "ooh" at pretty pictures. The future of astronomy is machine
>> astronomy.
>>
>
> Matthew
>
> Yes of course. The events are created and processed by machines. But the
> *streams* are built and connected up by humans who know what they are doing,
> like I have been doing with many event authors in the last 2 years.
>
> Building the stream is like setting the thermometer in the house (a human
> activity), followed by the heating system going on and off 1000s of times
> (automated activity), analogous to the event delivery and processing.
>
> The human understands that there is a light curve, and arranges for
> automated analysis of it, with the *knowledge* that the analysis is
> scientifically suitable for the data.
>
> Are you suggesting that more formalism in the VOEvent structure could
> somehow replace this human understanding?
>
> Roy
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/voevent/attachments/20110316/75d401fc/attachment.html>