From: Charles Giefer (cgiefer@cs.washington.edu)
Date: Mon Apr 26 2004 - 00:51:40 PDT
Of Objects and Databases: A Decade of Turmoil
Michael Carey and David DeWitt
The main topic of this paper was to discuss the relationship between objects
and databases. It is intended to track the evolution of object related
database research over the course of a decade and to predict future trends.
One of the better ideas that have followed through since 1986 is the use of
ADT (abstract data types) as a way to extend relational databases to include
some object support. Other ideas, such as persistent programming languages
and unique tool-kits to build and customize domain-specific applications,
failed because of their complexity, vagueness, and lack of interest.
One topic on which this paper comments that is of significance importance
today is the support of legacy systems. This is true even for today's
semi-structured data, XML. Seamless interaction between old systems and
standards, and leverage of established systems is very important to today's
research on topics such as XML processing.
I think this paper underestimated the impact and pervasiveness of the
Internet in the year 2006. I don't specifically see where objects would
play a role, but it is clear that web based processing and distributed
systems are not as much of a priority in this paper.
Much of the paper is beyond my ability to provide meaningful comment. The
actual term "object" is not thoroughly defined in this paper; intentionally
because it means different things in different circumstances. That is why
it was difficult to find a common thread in the areas of research. Also,
this paper lacks specific examples so that someone not familiar with the
purpose of using objects in databases, nor familiar with many of the
database terms and unique concepts, can get very confused.
It is clear that relational databases are still the preferred systems. This
may be because of the simplicity in their model. Objects seem to add a new
level of complexity and mess that has made their implementation a bit
burdensome.
Our course project will make use of a specific tool mentioned in this paper.
We are using an object-oriented language (Java) to provide and interface to
a relational database. Therefore, we can manipulate and encapsulate the
data in objects, but use the relational database backend.
This archive was generated by hypermail 2.1.6 : Mon Apr 26 2004 - 00:49:52 PDT