Back
News

Property databases

This month’s column is devoted to a single topic: the development of property database systems. It is an attempt to review the latest applications of information technology to the property sector and the methodology behind database design.

Although the electronic digital computer has been used for more than 50 years, it is still an emerging technology. However, the principles behind its design have changed very little. In contrast, the way in which we apply computer technology is being altered constantly. For example, by the introduction of fourth generation languages (4GLs), the use of expert systems and the application of the database.

Traditional information systems

The traditional approach to computerising information systems is to create an electronic equivalent of the normal office filing system, with files, records and fields. In other words, we simply replicate the office system on the computer. The only significant difference is that when information is stored electronically we have at our disposal much more powerful facilities for indexing, sorting, searching etc.

However, a number of problems have been identified with traditional computer systems. First, organisations using conventional systems, both computer-based and paper-based, inevitably duplicate much of the data which they store. Second, where organisations hold duplicated data it follows that inconsistencies will arise in the data stored. Third, system users are often frustrated in their attempts to make non-standard queries of conventional systems. And, finally, there is normally a strong link between programs and data structures within conventional systems, such that if programs are altered the data structures have to be modified, and vice versa. The database represents a way of minimising these problems.

The database approach

There are three types of database model: hierarchical; networked and relational. However, it is the latter approach which offers the most potential. The relational model is based on the concept of storing information about entities and the relationships between those entities. Entities are simply things about which we wish to store information (eg properties, landlords, tenants).

A number of relational database systems are marketed currently, including DB2, Ingres, Oracle and PICK. My own experience is with Oracle, which I have used to develop a property database system.

The ORACLE RDBMS is a relational database management system that is compatible with SQL/DS and DB2 — IBM database systems that run on large IBM computers. ORACLE also runs on many different mainframe computers, minicomputers and microcomputers.

In addition to the RDBMS, Oracle supply application tools that enable the user to exploit the benefits of the database, including SQL*Plus, SQL*Forms, SQL*Reportwriter and SQL*Menu.

SQL*Plus is an interactive SQL interface that enables users to manipulate an ORACLE database using a powerful 4GL. SQL*Forms is an interactive application development tool that enables the user to design and use forms to enter and retrieve database information. SQL*Reportwriter is a very powerful report generation program that allows the user to create single and multi-query reports using a number of reporting formats. SQL*Menu is a program that enables users to create and generate menus with in-built security. Finally, a product called Pro*COBOL provides a programmatic interface for the COBOL language: a precompiler that allows the user to embed SQL statements directly in COBOL code.

SSADM

In order to create a database it is necessary first to analyse logically a user’s requirements, and this may be done using a series of related techniques known as Structured Systems Analysis and Design Methodology (SSADM). SSADM consists of six stages: analysis; specification of requirements; selection of system option; logical data design; logical process design; and physical design.

The principal technique used is Entity-Relationship (E-R) modelling. An E-R diagram is formulated which represents entities, the relationships between them and the attributes used to describe them. A process known as normalisation is applied to produce a data structure in what is known as “third normal form”.

First normal form (1NF) is the removal of repeated attributes or groups of attributes. Second normal form (2NF) is where attributes are removed which are dependent on only part of the unique identifier (eg property reference number). And third normal form (3NF) is the removal of attributes dependent on attributes that are not part of the unique identifier.

CASE*method

To date, most software development has been aimed at increasing programmer productivity. For example, the introduction of 4GLs. Recently, however, attention has focused on the role of the systems analyst/designer. This has led to the development of Computer-aided Systems Engineering (CASE). That is software which helps the analyst to perform the above tasks.

Oracle has a suite of CASE products that are primarily aimed at supporting Oracle’s own development method, which is a derivative of SSADM. These are: CASE*Designer; CASE*Dictionary; and CASE*Generator for SQL*Forms.

CASE*Designer uses workstation technology to provide multi-user interactive CASE graphics that automatically update the CASE*Dictionary database. The CASE*Dictionary provides a multi-user ORACLE database for the capture of analysis and design details. CASE*Generator for SQL*Forms is the first of several generators which use analysis information in the dictionary to automatically generate sophisticated applications.

Future prospects

Users are more committed to systems when they have been closely involved in the design/development process. SSADM employs techniques which are readily understandable by users, and allows them to define logically their information requirements. In addition, CASE tools provide a stimulating environment within which analyst and user may co-operate to produce the required systems.

However, there is no value in “re-inventing the wheel”. Some excellent software has already been written for property applications. For example, packages designed to undertake investment valuation and portfolio analysis and development appraisal. Instead, attention should be focused on how to link such systems to databases designed using the above methodology. In other words, in property terms, we should seek to refurbish rather than to redevelop.

Finally, I recently came across a quotation from Jean Baudrillard which I feel system designers would do well to consider when creating information systems using the above technology: “Information can tell us everything. It has all the answers. But they are answers to questions we have not asked, and which doubtless don’t even arise.”

Up next…