The VLDB Journal (1998) 7: 129
The 23rd International Conference on Very Large Data Bases
was held in a beautiful peninsula near Athens, Greece, Au-
gust 26–29, 1997.
From 335 submissions in the research paper track, 53
were selected for presentation at the conference. Within the
team of program sub-chairs, we selected some of the most
highly ranked papers whose authors were invited to sub-
mit expanded and updated versions of their paper for pub-
lication in this Special Issue of the VLDB Journal. After
re-reviewing and some further revisions, the following ﬁve
papers were accepted.
A major theme of the conference were the challenges
posed by multimedia to database technology. Three papers
address this theme.
The ﬁrst paper, “Enhanced abstract data types in ob-
ject relational database” by Praveen Seshadri, criticizes that
current ADT concepts in object-relational databases are too
closed to allow efﬁcient support for multi-media content.
His extension allows an ADT to expose some of its method
semantics at the interface in order to facilitate query opti-
In a complementary approach, “Integrated document
caching and prefetching in storage hierarchies based on
Markov chain predictions” by Achim Kraiss and Gerhard
Weikum develops an integrated vertical data migration strat-
egy between primary, secondary, and tertiary storage to im-
prove the trade-off between storage costs and accessibility
of multimedia databases.
The third paper, “Scalable feature selection, classiﬁ-
cation and signature generation for organizing large text
databases into hierarchical topic taxonomies” by Soumen
Chakrabarthi, Byron Dom, Rakesh Agrawal and Prabhakar
Raghavan, looks at a special case of multimedia databases,
text databases, from the viewpoint of query interfaces. They
offer contextualized and annotated query responses based on
subject domain taxonomies and analysis of the present user
situation in a text network.
The efﬁcient implementation of large-scale object-ori-
ented databases still presents many problems. One of them
is the issue of garbage collection, i.e. the reclaiming of un-
referenced memory. The fourth paper, “Garbage collection
in object-oriented databases using transactional cyclic refer-
ence counting” by P. Roy, S. Seshadri, A. Silberschatz, S.
Sudarshan and S. Ashwin, provides a promising new solu-
The last paper, by Wee Teck Ng and Peter Chen, “Inte-
grating reliable memory in databases”, is a bit more futuris-
tic with respect to database architectures. Even though this
integration would require major changes in present-day com-
mercial databases and is thus certainly not around the corner
in practice, the authors show signiﬁcant potential for fu-
ture DBMS in terms of reliability handling and performance
improvement. This paper received the best paper award at
Thanks are due ﬁrst of all to the authors and program
committee members whose diligence in revising and review-
ing made it possible to publish this Special Issue within
a year. The attractive conference environment, made even
more attractive through the efforts of a perfect local organi-
sation under the chairpersons Yannis Vassiliou, Timos Sellis,
John Mylopoulos, and Nick Roussopoulos, certainly helped
to attract so many excellent submissions. The program com-
mittee sub-chairs, Mike Carey, Klaus Dittrich, Fred Lo-
chovsky, and Peri Loucopoulos coordinated not only the pa-
per selection for the conference but also the pre-selection
for this special issue.
Aachen, June 1998
VLDB 97 Program Chair