The VLDB Journal (2000) 9: 1
This special issue brings together a set of papers that ad-
dress some of the areas of intersection between database
technology and the World Wide Web. This intersection is
fast becoming a key focus of database research, with mul-
tiple facets that touch many different aspects of database
systems. As semistructured data and XML become preva-
lent, data models and query languages for them must be
proposed, studied, and implemented. As database systems
become key components in the construction of large web
sites, new sets of requirements must be incorporated. Query
optimization in the Web environment poses new challenges
and requires new techniques.
The ﬁve papers in this issue were selected from among
the fourteen that were submitted in response to our call for
papers. Each paper was reviewed, and for each them two
rounds of reviews were made possible by the enthusiastic
cooperation provided by our reviewers.
The ﬁrst paper in this issue, “Semantic caching of web
queries”, by Boris Chidlovskii and Uwe M. Borghoff tackles
performance problems of “web queries” (intended as queries
asked for by means of meta-searchers, which return the tu-
ples that satisfy a given Boolean condition). More precisely,
the paper considers conjunctive queries and studies how “se-
mantic caching” can be useful. The cache maintains regions,
and when a query is issued, part of the result is found in the
cache and only the “remainder” is retrieved over the net.
The paper “Learning response time for WebSources us-
ing query feedback and application in query optimization” by
Jean-Robert Gruser, Louiqa Raschid, Vladimir Zadorozhny
and Tao Zhan also considers performance issues, but from
a different point of view. The goal is to predict, by means
of learning techniques, the response time of a Web source.
It also shows how the predictions can be used by a scram-
bling optimizer, that is, an optimizer that changes a query
Mary Fernandez, Daniela Florescu, Alon Levy and Dan
Suciu in the paper “Declarative specifcation of web sites
Strudel”, provide a comprehensive description of Strudel,
a system for implementing “data-intensive” web sites. Ma-
jor features of the system are the distinction of three levels
(data, structure, and presentation) and the use of a declarative
query language for the speciﬁcation of the structure of sites.
The paper also contains an a posteriori discussion of some
experiences and brief indications about three “successors”
In the paper “Analysing navigation behaviour in web
sites integrating multiple information systems”, Bettina
Berendt and Myra Spiliopoulou discuss how the navigation
behavior of web users can be analyzed to evaluate the qual-
ity of a site (with speciﬁc reference to sites that dynamically
generate pages in response to queries speciﬁed by ﬁlling
forms). The analysis is based on the WUM environment,
proposed by one of the authors in a previous paper, and is
demonstrated with an experiment on a large real site.
In the ﬁnal paper of the issue “UnQL: a query language
and algebra for semistructured data based on structural re-
cursion”, Peter Buneman, Mary Fernandez and Dan Suciu
present the full details of a language for semistructured data
and XML. The paper is theoretical in nature, but the lan-
guage has a strong motivation regarding the current interest
in semistructured data and XML. Various formal results are
shown, some of which could ﬁnd application in the imple-
mentation of query languages for XML.
We thank all the authors who submitted papers to this
special issue, as well as the anonymous referees for their
disinterested and essential contribution.
Paolo Atzeni and Alberto O. Mendelzon