Impossible Queries

[Martin Hill]

> I _would_ like to explore the general principle of what to do when a
> resource can't answer a query exactly as phrased.  Is it acceptable to
> return an approximate answer (e.g. assume zero proper motion if no proper
> motions are known)?  If so, how would we inform the "user" of the
> approximation, and how might the user limit the approximations made in the
> processing?

My feeling on this is that a query that cannot be properly processed should 
be rejected at submission. That way the astronomer (or calling service) can 
redraft it so that the results are known properly.  I feel that data services 
at least should be very explicit and straightforward; any 'vaguenesses' and 
'best guesses' should be made at some higher workflow level, so that 
astronomers always have direct access to predictable results.

Over to you all...


-- 
Martin Hill
Astrogrid/AVO, ROE
Tel: 07901 55 24 66