NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/comp/database/oracle/2627

Internet Message Format | 1992-12-23 | 3.5 KB

Path: sparky!uunet!cs.utexas.edu!natinst.com!hrd769.brooks.af.mil!hrd769.brooks.af.mil!not-for-mail
From: [e-mail] (InterNet News)
Newsgroups: comp.databases.oracle
Subject: Re: What is Oracle?
Date: 23 Dec 1992 14:37:11 -0600
Organization: Armstrong Lab MIS, Brooks AFB TX
Lines: 58
Message-ID: <[e-mail]>
References: <[e-mail]>
NNTP-Posting-Host: hrd769.brooks.af.mil

Stan,
     Oracle is a brand name of one company's (Oracle corp.) implementation
of a Relational Data Base Management System (RDBMS).  These Really Darn Big
Monster Software systems are supposed to reduce the amount of REAL programming
necessary to access data stored in the computer.  It does this in two major
ways.  The first is at the data design level, the second is at the access
level.

     At the design level, very basically, trained designers break data into
small groups of related information through a process of 'normalization'.
This normalization eliminates redundant data, hidden relationships, and
a few other effects that could become a programmers nightmare.  These groups
of data are not linked together by the data base system but by the mind of the
programmer in the code he/she writes.  This allows for relationships to be
constructed that the designer had not thought of (good or bad).

     At the access level, most implementations of data base systems are moving
toward a Structured Query Langauge (SQL) to access their data.  The SQL allows
the user to INSERT, SELECT, UPDATE, and DELETE data without resorting to
writing a program in FORTRAN, COBOL, etc.  In essence, looking at all those
3GL programs written in the past to access data files, combine all of the
common routines; data definition, opening of files, looping though the files,
reading records, qualifying records for retrieval, sorting, etc.  Those
routines are the meat of the data base systems.  That leaves you with:

Insert into table values ('a',1);
Select field1, field2 from table where field1 = 'a';
Update table set field2 = 2 where field1 = 'a';
Delete from table where field1 = 'a';

These are simple examples to give you the idea of how time can be saved
for simple tasks.  If it were all that simple this wouldn't be the
competitive industry it is.

     One of the special requirements of a relational data base is that
it use it self to manage all of the data definitions, locations, and the
myriad of other data dictionary and house keeping chores.  The data dictionary
could be considered the heart of the RDBMS.  Even with all of the work saving
tasks performed by the system, I have seen people use it as a flat file
system.  Using an RDBMS just because the specs called for one.

     As to Oracle itself, it is not just the RDBMS that is the product.
There are several tools surrounding the kernel; an interactive command driven
tool (SQL*PLUS), a report writer, a data entry screen generator, a data loader,
and a few others.  The most interesting and most marketable point is that
Oracle software is designed to run on over 80 different combinations of
hardware and software.

     When considering whether or not to use one of these Real Dollar Busting
Money Sucking systems, do not look at the short term.  You must look at a
future when none of the original designers or programmers still work the
projects.  Look at the maintainability, ease of use, and the consistency
from one product to the other.

These are the basics as I see them, others may/will differ.  That's great.

Mike Stout... :-)