Database Reference
In-Depth Information
tially form, we now dispose of technology that is normally interconnected to support many terabytes or
petabytes of data, consequently we need systems that are able to efficiently manipulate this data.
The problem resides in the fact that human reasoning is very different from the logic involved in
software. In order to narrow this gap, there exists a branch in the scientific community which studies
different techniques to model the human mind. This branch is known as soft computing (Zadeh, 1994).
The aim of this work is to describe, in a general form, how to render the constraints of the queries on a
database more flexible, therefore making a Database Management System (RDBMS) flexible without
affecting in a considerable way its performance to manipulate large quantities of data.
Natural language is imprecise, vague and inexact; for example, when we say
Juan is a young person
or
Milton is very tall
, we are expressing imprecise terms like
young
and
very tall
; how does the computer
interpret these terms? Normally, we have to employ precise terms: a person is young if she/he is 30 years
old or less; but, what about the person who is 31 years old? One day I am young (I was 30 years old),
and the day after my birthday, Am I no longer young?
Furthermore, Juan is 50 years old and he considers himself to be young. That is true for him but false
for others. If Milton is 1.80 meters, do you think that he is very tall? A lot of persons will say no, but
if Milton is being selected to ride horses as a jockey? Then he is very tall; truthfulness or falseness is
relative, and depends on many factors. People have different points of view (preferences) and the role
of context is also important.
In a computer system, objects are clearly divided into white and black, but in the real world there is
an infinite grey area between the two and most things normally fall in the grey area. Morin (1999) states
that “Reality is not easily legible. Ideas and theories are not a reflection of reality they are translations,
and sometimes mistranslations. Our reality is nothing more than our idea of reality” (p. 44). Morin, a
French philosopher and sociologist, recognizes ambiguity and uncertainty as the hallmark of science
and of human experience.
Morin´s approach, complex thought, is in harmony with a culture of uncertainty. Morin (1999) asserts
“But in life, unlike crossword puzzles, there are boxes without definitions, boxes with false definitions, and
no neat framework to define the limits” (p. 45). How can we manipulate these boxes into the computers?
An example is the evaluation of young ladies for top models. There are many requirements to be
fulfilled and nobody satisfies all of them. However,
Lolita
was disqualified because she is 1.79 meters
tall and there exists a clear (
crisp
) condition stipulating a minimum height of 1.80 meters. But what
about the other prerequisites met? What happen if
Lolita
was the best candidate based on the rest of
them? This system may be very unfair, due to the rigidity of precise conditions.
We have proposed the enhancement of an RDBMS to support flexible queries and help users to define
preferences depending on the context by using fuzzy conditions in their SQL queries.
The contribution of this chapter is to describe the design and implementation of flexible queries (using
fuzzy logic) in a Relational Database Management System (RDBMS) without affecting its performance
to manipulate large quantities of data in a considerable form. The approach is Tightly Coupling, with
modification in the source code on the query processor engine in order to process fuzzy queries.
BACKGROUND
Traditional RDBMS suffer from rigidity: data are considered to be perfectly known and query languages
do not allow natural expression of user preferences. In order to provide flexible queries over databases,
