Impossible Queries

[Clive Page]

On Tue, 30 Sep 2003, Martin Hill wrote:

> > I _would_ like to explore the general principle of what to do when a
> > resource can't answer a query exactly as phrased.

> My feeling on this is that a query that cannot be properly processed should
> be rejected at submission.

I support the general idea that a misleading answer should not be
returned.  In many cases it should be possible to make use of nulls, which
are very important in astronomical data.  To digress for a moment: upper
(or lower) limits are almost equally important, but existing software can
sometimes manage to handle nulls (e.g. common-or-garden DBMS) but support
for upper/lower limits is contained in only a few packages specially
written by astronomers.

Here's a trivial example: if you do a query asking for proper motions
alone of a service which doesn't have them for the objects specified, then
an exception might be the right result, but if you ask for postions and
magnitudes and proper motions, than returning a table with just proper
motions set to NULL would be better.


-- 
Clive Page
Dept of Physics & Astronomy,
University of Leicester,    Tel +44 116 252 3551
Leicester, LE1 7RH,  U.K.   Fax +44 116 252 3311