A Few Older Research Projects
Data integration: We have several projects going on in
the field of data integration.
- Our
Tukwila
data integration system processes streaming XML data from multiple
networked data sources, and embodies novels techniques for adaptive
query processing. In developing Tukwila we developed techniques for
efficient processing of XML data, as well as a language for updating
XML.
- We are developing the BioMediator System that integrates data from
multiple genomic data sources.
- the
MiniCon Algorithm, the first algorithm for answering queries using
views that scales up to hundreds (and sometimes thousands) of views.
(check out my
survey on answering queries using views).
We have recently investigated the problem of selecting
views that should be materialized in a database or data
warehouse (see a paper in VLDB-01).
Web-site management: We developed the
Tiramisu
web-site management system. Tiramisu is a successor to the Strudel
system I co-developed at AT&T Labs. Unlike Strudel, which couples the
design and implementation of a web-site into a single tool, Tiramisu
enables a site designer to design a web-site, and then for the implementation to
be carried out by a collection of cooperating tools.
Model Management: In a collaboration with Phil Bernstein at Microsoft Research, we
have started considering the problem of Model Management. The
following
paper gives our vision for the field. In a nutshell, the problem
is to create tools for easily manipulating models of data, such as
relational schemas, DTDs, web-site structures, E/R diagrams, UML
models, etc. In the paper we argue that with a small set of operations
we can greatly simplify a lot of the work that programmers need to
perform today.