Transcript The Database Striptease and Informative Query Answers

The Database Striptease &
Informative Query Answers
Martin Kersten
CWI, Amsterdam
May 4, 2003 Lowell, USA
© 2003 M.Kersten
The Database Striptease
• My Volvo S80 has
– A integrated phone database
– A 6-CD GIS database
– A 14-CD Audio database
– A 3 Channel audio database
– A streaming database with traffic info
– A hardware control database
© 2003 M.Kersten
The Database Striptease
• At home we have
– A distributed telephone database
– Several electronic agenda databases
– MP-3 databases of the boys
– A large DVD database
– Tens of photo databases
– Insurance …
– ….
© 2003 M.Kersten
The Database Striptease
• My life depends on ca 50 incomplete, inconsistent, replicated smalllarge databases.
• As technology developer, I have not been able to locate even a basic
SQL interface to start managing them.
• As a person with finite resources
to enjoy life, let the DBMS
handle my data management
problems
© 2003 M.Kersten
The Database Striptease
• GRAND CHALLENGE:
Develop an organic database software system which
can be embedded in a wide collection of hardware
appliances and provides an autonomous self-descriptive,
self-organizing, self-repairable, self-aware and stable
data store-recall functionality to its environment
© 2003 M.Kersten
GONG-2
© 2003 M.Kersten
Informative Query Answers
• Setting the stage
– At //www.cwi.nl/~mk/onderwijs/db2/voc.tar.gz you
will find a database of the VOC (East Indian Trading
Company) stored as an XML file.
– You can load this file into a relational DBMS to prepare
for querying. Any of the standard products will do.
– You have a form-based interface in front of you that
permits presentation of a few records of the database
© 2003 M.Kersten
Informative Query Answers
• Next step: fill the missing part ???
select * from VOC where ????????????
Boat
Master
© 2003 M.Kersten
Kind
Tonnage
Built
Yard
Chamber
Departure
CallatCape
Informative query answers
• HIDDEN AD-HOQ QUERY PROCESSING
ASSUMPTION
– A user can only pose a precise query
• If he has a clear search objective
• If he understands the data distribution
• If he knows the semantic relationships
© 2003 M.Kersten
Informative Query Answers
<voyage>
<number>4408.3</number>
<boatname>BRESLAU </boatname>
<master>Jan Kornelis Roos </master>
<tonnage>1150 </tonnage>
<hired>1774 </hired>
<yard>A </yard>
<chamber>Z </chamber>
<departure>15-02-1783 </departure>
<harbour>Rammekens</harbour>
<callatcape>
<arrival>27-05-1783</arrival>
<departure> 12-06-1783</departure>
</callatcape>
<destination>
<arrival>06-08-1783</arrival>
<harbour> Batavia</harbour>
</destination>
</voyage>
© 2003 M.Kersten
Informative Query Answers
• GRAND CHALLENGE
Develop an informative query processor which takes
into account the user interest, ignorance, and lack of
knowledge on what to ask from the database and provides
the most informative answers in real time.
© 2003 M.Kersten