Fourteenth Scientific and Statistical Database Management Conference
I recently attended the Fourteenth Scientific and Statistical Database
Management Conference (14SSDBM), which was held in the NeSC building
in central Edinburgh during 24-26 July 2002. This note is a brief report
of this conference. Different disciplines have different traditions, I
suppose, and there were a couple of aspects of the conference which
were slightly different from what would be the norm for an astronomy
conference. For example, the authors of poster papers gave short oral
presentations. More significantly, the conference proceedings were
prepared before the conference and distributed with the conference
information: a commendable innovation. I'll include the full reference
for the proceedings at the end of this note. There seems to be no
mention of an on-line version, which is a pity.
The emphasis of the conference was on bioinformatics and statistical
metadata. I'm still not entirely sure how much astronomy really has in
common with bioinformatics. However, our problems do seem to be
quite similar to some of those encountered in the statistics community:
they have large, distributed, heterogeneous tables that they want to
query.
Areas in which papers were presented that might be relevant to
AstroGrid included: ontologies and user interfaces, provenance tracking
and managing distributed resources. I'll discuss these topics briefly
below. Note that this list does not exhaust the topics covered at the
conference.
Ontologies and User Interfaces
There were several talks describing systems where a domain-specific
ontology was used by the user-interface to assist the user in creating
complex queries. One speaker had a hierarchical (ie. tree-based)
ontology which was used to drive a user-interface with hierarchical (ie.
tree-based) display. This user-interface was used to query a database
of different types of trees (ie. real trees - big plants that grow out of
the ground). All of which left me somewhat confused.
Raguenaud gave a review of multiple overlapping classification
systems. This talk didn't mention `ontology' once, but I suspect that
much of it would apply to any astronomical ontology that we adopt.
Provenance Tracking
There were a couple of papers describing papers for tracking
provenance. One was by Bose and the other by Foster et al. described
their `Chimera' system. This sort of work seems to be at a relatively
early stage.
Managing Distributed Resources
Unsurprisingly, there were several papers on managing distributed
resources. The most relevant ones seemed to come from the statistics
community. Ryssevik described the NESSTAR system and associated
Data Documentation Initiative. NESSTAR seems to be a distributed
system with many similarities to AstroGrid. Unfortunately (and
unusually) the proceedings contain only an abstract of this paper.
Nelson described a metadata registry for statistical data built using
SOAP, WSDL and the OASIS/ebXML Registry standard. Again there is a
close correspondence with the resource registry that AstroGrid will
need. Both these projects (which are related) seem sufficiently similar
to AstroGrid's work that I think that we should investigate them
further.
Miscellaneous
Sattler presented an interesting paper on annotating scientific images.
I suspect that in astronomy we would adopt a different approach (the
annotations were stored separately from the images), but it was
interesting work. There were also several talks on efficient ways of
executing queries, but none seemed immediately relevant.
References
`Proceedings of the Fourteenth International Conference on Scientific
and Statistical Database Management', ed. J. Kennedy, 2002 (IEEE
Computer Society: Los Alamitos, California). IEEE Computer Society
order no. PR01632.
Conference home page:
http://www.ssdbm2002.ed.ac.uk/
NESSTAR home page:
http://www.nesstar.org/
OASIS/ebXML Registry:
http://www.oasis-open.org/committees/regrep/
Clive Davenhall,
6/8/02.