Foundation Matters
C.J. Date
Independent Consultant
PO Box 1000
Healdsburg, CA 95448
U.S.A.
Abstract
This talk is meant as a wake-up call ... The foundation of
the database field is, of course, the relational model. Sad
to say, however, there are some in the database
communityMcertainly in industry, and to some extent in
academia also--who do not seem to be as familiar with
that model as they ought to be; there are others who seem
to think it is not very interesting or relevant to the day-to-
day business of earning a living; and there are still others
who seem to think all of the foundation-level problems
have been solved. Indeed, there seems to be a widespread
feeling that "the world has moved on," so to speak, and
the relational model as such is somehow
passe.
In my
opinion, nothing could be further from the truth! In this
talk, I want to sketch the results of some of my own
investigations into database foundations over the past
twenty years or so; my aim is to convey some of the
excitement and abiding interest that is still to be found in
those investigations, with a view~I hope~to inspiring
others in the field to become involved in such activities.
First of all, almost all of the ideas I will be covering
either are part of, or else build on top of,
The Third
Manifesto [ 1 ]. The Third Manifesto
is a detailed proposal
for the future direction of data and DBMSs. Like Codd's
original papers on the relational model, it can be seen as
an abstract blueprint for the design of a DBMS and the
language interface to such a DBMS. Among many other
things:
It shows that the relational model~and I do mean the
relational model, not SQL~is a necessary and
sufficient foundation on which to build
"object/relational" DBMSs (sometimes called
universal servers).
Permission to copy without fee all or part of this material is granted
provided that the copies are not made or distributed for direct
commercial advantage, the VLDB copyright notice and the title of the
publication and its date appear, and notice is given that copying is by
permission of the Very Large Data Base Endowment. To copy otherwise,
or to republish, requires a fee and~or special permission from the
Endowment
Proceedings of the
28 th
VLDB Conference,
Hong Kong, China, 2002
It also points out certain blunders that can
unfortunately be observed in some of today's
products (not to mention the SQL: 1999 standard).
And it explores in depth the idea that a relational
database, along with the relational operators, is really
a logical system
and shows how that idea leads to a
solution to the view updating problem, among other
things.
Note"
The foregoing interpretationmi.e., of what a
database really is--is directly relevant to the process of
logical database design
(and I will mention some recent
results in this connection). It is also directly relevant to
what the commercial world calls
business rules
[2].
Reference [1] also complements the relational model
by introducing a detailed proposal for a theory of types. In
particular, that theory includes a novel approach to the
vexing issue of type inheritance, an approach in which the
answer to the famous (or infamous) question "Is a circle
an ellipse?"
ismpace
much of the object literature on the
subjectMa resounding
yes.
In fact, I will explain why I
believe objects and a "good" approach to type inheritance
are fundamentally incompatible.
More recently, Hugh Darwen and I, along with Nikos
Lorentzos, have been building on Lorentzos's original
work and the ideas presented in reference [1]mincluding
the type inheritance ideasmto investigate the question of
support for
temporal data
[3]. Again, it is our belief that
the relational model is a necessary and sufficient
foundation on which to build such support. It is true that
we have defined a large number of new relational
operators (with a view to raising the level of abstraction
and simplifying implementation), but all of those
operators are, in the final analysis, nothing but shorthand.
We have also, among other things, defined a new
("sixth") normal form and proposed a temporal database
design methodology.
Acknowledgments:
Most of the work I will be
reporting on was done in conjunction with my friend and
colleague Hugh Darwen of IBM in the UK. Other
collaborators include David McGoveran of Alternative
Technologies in California and Nikos Lorentzos of the
Agricultural University in Athens, Greece.

Get Proceedings 2002 VLDB Conference now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.