A new approach to relational database management
One of the earliest relational systems actually incorporated many ideas which are
needed today - but are expressly condemned by a number of relational puritans.
Introduction
G-EXEC was one of the first relational database management systems,
developed in the 1970s by a team in the UK Institute of Geological Sciences
(now the British Geological Survey), but its ideas and design concepts are still
just as valid today.
One of the big problems in relational database management has always been the
treatment of missing data. The SQL 'NULL' is generally accepted to be an abomination,
but Date, Pascal, and others have reacted by attempting to forbid the
use of any missing data representation in relational databases. They have done this
by re-defining the relational database in terms of the 'closed world assumption'.
This states that anything not included within the database is false.
However, it was never Codd's intention that the
relational model should be
restricted in this way, and he explicitly accepted that both closed-world and
open-world assumptions have their place. Using the open world assumption may
require the use of multi-valued logic, but this is not the insuperable obstacle
which Chris Date presents.
New ssoftware is to be developed using the relational
principles as defined by E.F. Codd, rather than the much narrower 'revisionist'
principles of Date, Pascal, and others.
Another web site
www.vmine.com
will chart the progress of the development
of this software, and related developments such as new database definition and query
languages which will be needed. What will be discussed on this web site, www.g-exec.com,
will be the more theoretical aspects of the true relational approach, as envisaged
by Codd, and which allows the construction and use of relations using both 'closed'
and 'open' world assumptions.
