2.1     Introduction

The age of computers can be said to be synonymous with the age of information. Modern computers of the 90’s era have become more powerful than ever but more importantly, available to a large mass of people as prices of personal computers come within reach of the average consumer. In other words, a wealth of information can be made more readily available to the general public. Computers now provide a great capability for integration and analysis of different types of geographically referenced data and for displaying information in an effective, impressive, convincing and attractive manner. As town and country planning by its very nature is an information-rich discipline, one would not be surprise to find the application of geographic information system for such a purpose a very attractive proposition. Although town and country planning has evolved over time, the modus operandi of problem solving has still remain consistent. Thus, if more information could be made available to address a particular problem, this would increase the probability of formulating more appropriate and effective solutions. Town and country planning operates at the national, regional, state and local level consequently, the type of information required by the planning authorities are varied and extensive. In Malaysia, the legal requirements of town and country planning make it imperative that decision-making be made objectively and if necessary, be recalled if an appeal is made by an aggrieved party. Therefore, the collection and storage of relevant data and information is most relevant and the use of GIS would be a most appropriate tool for the storage, retrieval, and manipulation of data to assist decision-making in town and country planning.

As GIS is relatively new and is developing at a fast pace, this section begins with a definition of GIS to enlighten the reader with the complexities faced by GIS resulting from the heterogeneity of use. This is followed by a review of the current trend of the GIS industry and looks at how IT has influenced GIS in particular, the impact of the Internet. Object-oriented technology too has been creating a lot of interest in the GIS industry and deserves some mention as being very revolutionary in approach. Efforts to utilise GIS as a decision-making tool are exemplified in spatial analysis along with other GIS models namely, a land management system, long-range decision support system and local authority information system. The application of GIS as a decision support system in the field of town and country planning follows suit. Fundamental problems in the use of a GIS, in particular planning are highlighted followed by factors leading to the successful implementation of a GIS. The section concludes with a summary of the future of GIS and its suitability as a decision-support system in the field of town planning.

2.2     Definition of Geographical Information System [GIS]

GIS is a relatively recent phenomenon. The outcome of the computer revolution over the past 30 years helped propel its rapid development and could be attributed to the following factors:

1. GIS has been involved in a very diverse field of application;
2. The many attractive market opportunities in GIS;
3. Different ways of defining objects and subjects; and
4. The academic debate on the basic form of GIS.

Flexibility seems to be one of the greatest strength of GIS. This is supported as it proves to be a very adaptive tool for a wide range of applications. Such diversity has seen GIS expand from the traditional world of cartography, planning and environment management to new challenges such as business geography. The irony is that this has resulted in a variety of interpretations on what constitutes and defines GIS. Terminology are not standardised and key issues on the nature and scope of GIS become blur and unclear. The many definitions of GIS merely show there is no common unison among the various sectors of the GIS community.

To illustrate this point, Tomlinson [1972] whom is recognised by most in the GIS industry as the father of GIS specified GIS as a tool for computerised mapping to help produce maps more cheaply and faster. Burrough [1986] associated GIS to functions and defined GIS ‘as a powerful set of tools for collecting, storing, retrieving, transforming, and displaying spatial data from the real world.’ Cowen [1999] defined GIS as ‘a system of hardware, software and procedures designed to support the capture, management, manipulation, analysis, modelling and display of spatially-referenced data for solving complex planning and management problems.’ Others acknowledged GIS as an information technology. [See Fig. 2.1] M. Juppenlatz and X. Tian [1996] saw GIS as ‘a system of ordering, managing and accessing large quantities of information.’ Duane F. Marble [1990] defined GIS as a system, which should contain the following major components:

1. A data input system;
2. A data storage and retrieval sub-system;
3. A data manipulation and analysis sub-system; and
4. A data reporting sub-system.

Fig. 2.I Spatial Information System

Source: Geographical Information Systems for Urban and Regional Planning, 1990, Kluwer Academic Publishers

From an academic view, GIS is possible without the aid of a computer but this is not practical anymore and it may be true to say no GIS currently is not computer aided. Indeed, the benefits of a computer system to handle a huge volume of data and rapidly process and analyse data to assist decision-making is universally accepted and strengthened as computers become increasingly powerful, cheaper and accessible by the day. This could be probably why the U.S. Geological Survey [1999] defined GIS ‘in the strictest sense, a GIS is a computer system capable of assembling, storing, manipulating, and displaying geographically referenced information’. Similarly, Environmental System Research Institute, Inc. [1999] defined GIS as ‘a computer-based tool for mapping and analysing things that exist and events that happen on earth.’

D. R. Taylor [1991] believed there is no universally accepted definition of GIS as result of the many related disciplines but it appears there is a general agreement that a GIS has the ability to analyse data. So the keyword ‘data analysis’ is seen by many as a component without which, strongly lacks a true definition of GIS. Admitting there is no easy answer to define GIS, Ron Briggs [1999] stated it should at a minimum comprise the capability for input, storage, manipulation and output of geographic information. He defined GIS as ‘a system of integrated, computer-based tools for end-to-end processing of data using location on the earth’s surface for integration in support of integrated decision making.’

The dynamism and evolution of GIS is illustrated by Crain and MacDonald in 1984 [Maguire1991] who suggested that the development of GIS is a three fold overlapping scheme:

1. Generic, location query- emphasis on data inventory
2. Analysis application - emphasis on analysis
3. Decision support system- emphasis on management

But as GIS evolved over time, Maguire [1991] felt it would be wrong to define GIS in a minimum essence for fear that would not be able to present a minimal holistic view of GIS. He sees GIS in the form of a series of views that is mapping, database and spatial analysis. This is because maps have always been central to GIS for purposes of map processing and map display. Secondly, the key feature which differentiates GIS from other information system has been the general focus on spatial entities and relationships of its database. Lastly, spatial analysis through search and overlay operations has always been a key functional feature of GIS.

As new modus operandi are developed as exemplified by the current research towards a fully object-orientated GIS, this may further lead to the discovery of other hidden potentials never perceived before. If GIS were such a dynamic and evolutionary subject, would it be safe to say that the definition of GIS is as dynamic as its evolution? In a nutshell, the development of GIS is only hampered by man’s own creativity in optimising ways it can be of service to mankind where the definition of GIS will largely be a result of that creativity.

2.3     The Current Trend of the GIS industry

The GIS industry is creating waves. In the Asia Pacific region, Hastings [1999] sees the real challenge is overcoming the current economic crisis placed on the IT industry. Through her discussion with a number of key software/hardware vendors, associations and government, it was felt that the GIS market in the Asia Pacific over the next three years would concentrate on:

1. Geographic Positioning System [GPS] telecommunication and utilities sector development;
2. Business GIS application and system integration;
3. Internet and world wide web based technology;
4. After sales service, improved customer relation;
5. High-resolution images;
6. Simulation; and
7. GIS data capture.

Needless to say the economic crisis has had a profound impact on the GIS related industries: some had to reduce cost and revise their business strategies while others with new approaches logically, preferred to concentrate on product introduction and training. GPS specialists strengthen their service, support capabilities and distribution while those that survived believe it was because of the robustness of their reseller network.

Odenwalder III [1998] felt that whilst there has been a growing collection of GIS application and third party plug-ins, there has been a fundamental shift in the approach to technology where the user, not the technology has become central in the development of the GIS software. In the user-centric model, developers start with the users, learning their objectives, job and current reference then tailor their application around the user. Three trends are expected to emerge:

1. Geospatial data management will become a core information technology that will embed deeper within the general IT infrastructure;
2. The Internet model will emerge as the de facto standard for computer/remote networking and communications and make it easier to link systems throughout an organisation and the world; and
3. Periphery technologies will extend the scope of geospatial data beyond the office-bound workstation or PC.

Joe Astroth of Autodesk [Hastings 1999] indicated four major developments would continue to dominate the GIS industry in the near future:

1. The influence of Internet on GIS;
2. The development of content–rich GIS information sites;
3. The movement of GIS technology to Windows 95/98 and Windows NT; and
4. The rapid convergence of CAD and GIS technology.

Of the four, Walter Mayr of ERSIS, Henry Tom of Oracle Corporation and Tony Hart of MapInfo Australia share concern for the impact of the Internet on GIS. Glifford [1999] feels that this is because the idea of integration between GIS and Internet technology has been one of the most important problems inhibiting information utility. Thus, it is expected that the development of the GIS for Internet use in the next few years will be substantial. However, he pointed out the Internet is a slow and unstable network that is not very compatible with large GIS database. Furthermore, client applications on the Web were not optimised for complex interaction with GIS application. Modern browsers were also initially not designed with GIS in mind and do not allow for multiple mouse clicks to define a viewing window and it was only recently in February 1999 that an approved standard for vector graphics on the Web was established.

So why is the Internet is the current buzzword in the GIS/IT industry? Glifford believes the widespread availability allowed by a standardised set of information deployment technologies far outweighs the limitation that such an environment imposes. It is also possible that whatsoever information technology that does not adapt to the Internet is bound to be obsolete in a short time.

There are basically two approaches to deploy GIS on the Internet: server or client application. In a server application, a web browser is used to generate server request and displays the results where the hardware usually consist of a GIS server and the GIS database while functionality resides in the server. In the client application version, the client is enhanced to support GIS operations. Advantages and disadvantages between the server-side and client-side GIS are shown below:

Glifford anticipates that client applications will become more prevalent but if real Internet GIS is what the market prefer, then the server-side GIS will the choice. This is because the issue of standards has not yet been deliberated. On that point, it is felt that in a matter of time, the server-side GIS will eventually win the battle between the two as its concept allows for a true Internet audience with different platform and network capacities.

On a local scene, Munit, Latip & Khoo [1998] indicate this approach has already being employed in Selangor’s state enterprise system known as Darul Ehsan Geographical System [DEGIS]. DEGIS was developed to implement an internal land management system and provide the general public information on land matters. All users have access to DEGIS via its Internet homepage though subscribed users enjoy access to check the status of their applications, cadastral maps and correspondence with the land office. Common Object Request Broker Architecture [COBRA] was used to enable ease of maintenance and integration. For use of GIS on the Internet, the system uses a SICAD-Internet Map Server [IMS] web module together with a SICAN map server where these modules communicate via a Common Gateway Interface [CGI] with the Web server.

2.4     GIS and Object Oriented Technology

Object-oriented technology is rapidly craving a name in GIS technology and deserves special mention. It views the real world as a collection of natural things or objects and arguably makes this more realistic for GIS than the georelational method. Among the advantages of object-oriented technology [S. J. Fletcher 1999] are its excellent storage and spatial modelling capabilities. This highlights a key point of object-oriented philosophy that powerful functionality lies in the data rather than the application. Furthermore, Addison [1999] pointed out this provides the means for logical system design and maintenance and affords considerable flexibility. The data is always accurately linked through heredity information and makes it impossible to bypass data other than through the object. Object-oriented technology also excel in versioning, the process which updates are copied and stored to enable close monitoring of the evolution of data between different dates. It also resolves the problem of long transaction problem when data has to be updated by multiple users over an extensive period, an area of difficulty in non object-oriented spatial software. Thus, it is not surprising that as object-oriented technology is revolutionising GIS. [Berry 1996].

However, Kufoniyi [1995] argued despite the initial enthusiasm that heralded the introduction of object-oriented modelling in GIS applications, its implementation has not been as rapid as expected. He says a major reason is the lack of a standard object-orientated query language comparable to SQL, the Standard Query Language. Another probable reason he offered is that many GIS users are not familiar with the concept involved in object-oriented modelling. This implies relational structure will continue to play a significant role in database management either in conventional form or as a base for object-orientated query language.

If proof is in the pudding, the commercial success of object-oriented GIS is proof enough that object-oriented GIS is not an option but a near future requirement. Sargent [1999] indicated that even heavyweights users like the US Geological Survey, UK Ordinance Survey, Land Information New Zealand and the Czech Land Survey Office have entrusted object-oriented technology into their future geospatial installations. In addition to that, organisations in more than 30 countries have benefited significantly from using object-oriented GIS taking into account the first object-oriented GIS products were only sold eight years ago. There obviously must be something in object-oriented technology that [Sargent 1999] makes the file-based and relational database GIS vendors scramble to catch up with object-oriented GIS capabilities.

Maguire [1999] finds the key reasons for moving to an object-component model include having a GIS that more closely match the real world; more scope for extension and customisation, in particular, the support for custom data models with specific features and modern approaches to software engineering leading to higher quality software that is easier to maintain. The georelational data model has proved its flexibility, extensibility and good performance but begins to show its limitation for modelling rich geographic objects and cannot be easily extended to support user or domain specific features. Although the object-oriented GIS software had suffered problems of sharing parts of the system, recompiling, lack of good modelling language and proprietary interfaces, it has managed to resolve these problems. The key to this success lie in the ability of components to implement in a practical way many of the object-oriented characteristics of encapsulation, inheritance and polymorphism making it superior to an object approach. So while the vendor can only carry out complete customisation capabilities under the basic object model, the user can undertake this facility himself under object-component model.

It is most likely that the object-component approach to GIS design will quickly become the implementation norm for the following reasons:

1. Object-components are based on standards making them clearly defined and well understood;
2. New components can easily be created because components have very good extensibility qualities;
3. They support run-time which means a system can be extended while it is still running;
4. The programming language is neutral in particular, Component Object Model [COM] object components;
5. Object behaviour is in the data model which makes it fully recognised by the system and is available to all users; and
6. Object-oriented software is a powerful design and analysis tool and has already been successfully used.

Myers [1996] argued although the integration of a GIS and object oriented application offered the use of two systems by a wide audience, unfortunately the pace of development for both systems and computer languages are so fast that for one to fully understand their capabilities and effectively utilise them presents a real challenge. Even so, with the majority of end-users committed to the popular relational database, does one jump on the bandwagon and shift to object-oriented GIS? To this issue, Sargent offers the following guidelines:

1. An extended relational database is an adequate, rudimentary solution when simple location requirement or post coding is the only aim;
2. The object-oriented GIS is preferred when rich spatial structure need managing and manipulating and a rich active data repository is required; and
3. The object-oriented GIS is preferred when high integrity, long term update and maintenance are required and feature meaningful segments together with long transactions with version control are essential.

To summarise, object-oriented GIS technology is crucial when placing much investment in the structure of their spatial data including:

1. Complex features and relationship among data;
2. Layers in interpretation and derived information from raster images; and
3. Active data in features or particular types share behaviour.

2.5     GIS for Spatial Analysis

To many it is query and more especially, the analysis function is the heart of GIS. The ability to analyse geographical pattern and relationships differentiate GIS from other computer systems [D. J. Maguire & J. Dangermond, 1991]. From that, S. Openshaw [1991] noted both spatial pattern and description could be applied in three markedly different contexts:

1. Testing a hypotheses about patterns and relationships present in spatial data;
2. Efficient spatial pattern and relationship description; and
3. Analysis for purposes of decision support and spatial planning.

However, he argued key functions are missing from current GIS. These are basic exploratory geographical analysis tools and relationships that may exist in spatial database Hence, the development of a truly integrated GIS which would allow simultaneous query and analysis of both raster and vector data than restructure has yet to come. In the meantime, Chen [1996] still believes GIS has shown promising prospects for performing often-complex spatial analysis and suggested some popular analysis methods and tools frequently used in GIS decision support applications:

Multiple Data Layer Analysis

Layering geographic data has been a powerful method of analysis used in many fields. Methods of analysis include Boolean operations, weighting and fuzzy logic at one or multiple levels and by buffering. This technique also enables the creation of new thematic layers to further assist decision-making more objectively and comprehensively.

Simulation

Simulation represents a high-powered analysis method based on a comprehensive understanding of the natural and social phenomena. In order to ensure that the simulation is close to the actual phenomena, certain rules must be accepted and is highly dependent on recent and accurate data in order to produce realism.

Prediction

Prediction is the sophisticated end of simulation but a predictive model requires much time-tested trials and experiences before it can be termed reliable. This implies a philosophy that increased frequency of success of a particular phenomena correlates directly to the probability of a prediction coming true.

Decision-Making

Decision-making is a complex process influenced by both human and non-human factors. Though GIS cannot replace the act of decision-making by human, the development of many simulated alternative results should be able to guide decision-making. When clear rules are strongly adhered and openly used, it provides more transparency in decision-making.

2.6     Types of GIS Models

The idea of ‘modelling’ is fundamental in most information systems. This is because a model represent, formalise concepts and ideas, description and simulation of the real world Thus, before one can extract valuable information, relevant geographic data need to be ‘modelled’ before it can be entered into an information system. Modelling involves several stages and includes data capture, interpretation, encoding and structuring. Sepakat [1997] indicated that the basic information of a data model should include:

1. Spatial data types;
2. Feature classification codes and coding standards;
3. Unique identification number;
4. Spatial location number;
5. User defined attributes; and
6. Spatial relationship [topology].

Nordin [1997] pointed out that due to the multi-dimension and heterogeneity of geographic phenomena and events, the term ‘model ‘ used within GIS has a diversity of meaning and nuances. Because of that, it would be impossible to develop a scheme of GIS functionality that is completely comprehensive [David J. Maguire & J. Dangermond 1991]. McMullin [1999] argued if a GIS is to be a spatial decision support system, it must provide modelling capabilities on par with in disciplines such as statistics and management sciences. However, he pointed out presently this is not the case as basic GIS software have virtually no models or support for directly integrating them into proprietary packages. The standard GIS software also is too generalised and falls short of the real GIS requirements of many disciplines. Consequently, one has to customise specific models for particular needs. Some GIS models used to support decision-making include:

1. Land Management System [LAMS]
2. Long Range Decision Support System [LRDSS]
3. Local Authority Information System [SmartMAP]

Land Management System [LAMS]

New Zealand has an agriculture-based economy that is crucial to changes in landuse. Pastoral management systems too are changing and have a direct effect on the environment. Coupled to that, constant changes of the landuses in response to market demands along the flood plains place a premium on water quality and water resources. Since many of the risks are water-related, GIS applications were used to identify and quantify the impact of risks in relation to these changes thereby assisting decision-making pertaining to landuses. Luckman [1996] felt the LAMS modelling tool designed by Landcare Research, a government owned New Zealand Research Company provided a robust approach for making predictions as well as managing and maintaining knowledge about the effect and impact of changes in landuse. The process involves the identification and importance of different sources of sediment where each source is investigated as to their effect on landuses. Central to this is a comprehensive knowledge of factors to assess sedimentation risk; hence, a knowledge-based spatial model was used to identify risks by illustrating possible scenarios and changes to various levels of risk.

However, such a system brings about issues that need special attention. For instance, risk is a subjective matter and ought to be define by the community not the GIS developer. If probabilities are used in the models, they must be clearly defined to ensure effective analysis. Modelling tools need to be user-friendly and accessible to end-users whom are not always programmers. Finally, this approach requires a critical evaluation of changes in the state of key sedimentation sources. Ironically, agriculture rich New Zealand lacks detail data to assist LAMS analysis so current update of its database is a logical step in the right move.

Long Range Decision Support System [LRDSS]

India is currently faced with rapid economic and export growth that has created pressure for rail infrastructure. Consequently, the identification of the most cost-efficient means of achieving capacity expansion was a top priority for the Indian Railways. P. Cook & A. Mukerjee [1996] pointed out that the development of a Long-Range Decision Support System [LRDSS] was awarded to GIS Trans Ltd. It helped facilitate the planner regarding queries on traffic congestion by aiding forecast total traffic flows and provided alternative traffic patterns with an aim of eliminating bottlenecks. To meet their objectives, the LRDSS was structured around six key modules:

1. Traffic forecasting ;
2. Facility performance ;
3. Traffic assignment ;
4. Cost-benefit analysis ;
5. Financial forecasting; and
6. Market analysis.

However problems were many: Database proved to be a key issue because traffic forecast by origin-destination had not existed before. Other than that, data wherever available were stored on unreadable format [tape] and were not consistent with the zoning of railway lines. Available statistics were also aggregated, not suitable for analysis and difficult to acquire. Faced with limited data processing facilities, data collection was also time-consuming. Hence, the problem with the LRDSS was basically that it a rather ambitious project. To make things more effective, the authorities should have initially resolved these issues before embarking on the LRDSS project.

Local Authority Information System [SmartMAP]

The application of GIS in developed countries is popular for organisation purposes and M. Juppentatz and X. Tian [1996] noted that overall, local government is one of the biggest GIS user groups. On the other hand, Nappi [1990] stated that the growth of GIS in developing countries is linked to the relationship between the driving forces which create a set of circumstances that require increased management and attention, and the enabling forces which supply the technology to manage and observe the driving forces.

In Malaysia, Idrus [1999] stated that advancement in GIS technology and demonstrations in a variety of successful government applications led many decision-makers to accept GIS as potentially useful for many municipality functions. In the case of Shah Alam Municipal Council, GIS assist the local government at two levels of management or decision making process. At the macro level, the development of a reasonable comprehensive land-related geographic database would eventually allow for easy, quick retrieval and reference of information to assist policy and decision making by the State Authority. At the micro level, GIS would be optimised for planning, management and administrative purposes. For example, integrating various types geographic and geo-referenced data for use by the relevant departments would help expedite the processing and analysis of data by the planning department that could later be used for monitoring and operational purpose by the engineering department. Thus, GIS was made to function as a multi-user system because:

1. It minimises the duplication of effort in data collection and maintenance ensuring access and convenience of sharing a set of standardised set of database;
2. It is more cost-effective by reducing duplication investment in GIS for the same agency;
3. It allows for the sharing and optimisation of human resources in particular, of GIS and computer specialists, to staff and administer the system;
4. It reflects and facilitates the actual interdependence of data needs and functions among the various departments within the same agency;
5. A digital database allows for more manipulation function through specific applications allowing for rapid decision-making; and
6. It provides an avenue to recover substantial cost and make some form of revenue by selling its geographical information to the public.

The SmartMAP model was used to optimise this technology. One of the main strategy of this model was to help manage, supervise, maintain and monitor the many public utility services under the responsibility of the local authority as legally required under the Local Government Act, 1976 [Act 171]. Town planning is also a key responsibility of the local authority and therefore incorporated into the system. Although the SmartMAP has six modules [See Fig. 2.2], three main modules concentrate on management:

1. Infrastructure planning and management system;
2. Landuse, planning and management system;
3. Property information management system;
4. Areas of Interest;
5. Public Complaint system; and
6. Plan printing.

SQL has been employed in SmartMAP to facilitate decision-making. As the local councillors often are not technically trained, the state of affairs in the local authority could be clearly brought to their attention in various forms of charts and diagrams through the use of GIS software. Queries on social and communal facilities have also been made according to specific areas, type and status of the local authority. Queries have also been designed on land matters, property assessment, rental, landuse classification and property ownership. SQL were also devised on matters relating to the general population, bumiputera participation in business, income level by race, section and employer as well on commerce and housing to gauge the socio-economic development of their community.

Fig. 2.2 SmartMAP Information Model

Source: Laporan Rekabentuk Pangkalan Data GIS untuk Majlis Perbandaran Shah Alam, Integrated Geographic Design Sdn. Bhd.

The SmartMAP pilot project was reportedly a success and consequently the Municipality is currently embarking on the second phase of the GIS project that involves the collection of relevant data for the whole area of the Municipality. However, Idrus [1994] stressed the materialisation of a successful GIS in a local authority to a large extent, is dependent on factors beyond the technique of GIS itself, for example, on matters like cost, manpower, data availability and technological development of IT.

2.7     GIS as a Decision Support System

Firstly, there is a need to distinguish between spatial decision support system [SDSS] and decision support systems [DSS]. SDSS have evolved in parallel with DSS but SDSS defer slightly in that additional characters of a DSS are added and include the capability to:

1. provide for spatial data input;
2. allow storage for complex structures common in spatial data;
3. include analytical techniques unique to spatial analysis; and
4. provide output in the form of maps and other spatial forms

Although there has been an increasing interest to use GIS to provide decision support, Densham [1999] mentioned many spatial problems are complex and require the use of analysis and models. Furthermore, many spatial problems are semi-structured and because of these aspects cannot be measured or modelled. He argued that GIS fall short of the goals of SDSS because:

1. Analytical modelling capabilities often are not part of GIS;
2. Many GIS database have been designed for cartographic purposes display of results;
3. The set of variables or layers in a database may be insufficient for complex modelling;
4. Data may be insufficient scale or resolution; and
5. GIS designs are not flexible enough to accommodate variations in either the context or the process of spatial decision-making.

Therefore, in order to make GIS more effective for spatial decision-making, he suggested a need to integrate GIS capabilities found in database management systems, graphical display and tabular reporting with analytical modelling capabilities and the decision-makers expert knowledge.

K. Fedra and R. F. Reitsma [1990] are of the opinion almost any computer-based system from database management or information system via simulation models to mathematical programming could conceivably support decisions. This is because of the capability of GIS to analysis, display and to a certain extent model, that make people regard GIS as a special class of decision support system. But this is done primarily in combination with simulation and optimisation models, database of non-digital data, artificial Intelligence, expert system technology and decision support tools. Kennan [1997] felt that many GIS based system are being described as DSS not based on a academic definition but on the basis that GIS assisted in the collection of data used by the decision maker. He suggested three categories where SDSS may make a contribution:

1. This group covers the traditional uses of GIS such as geology; forestry and landuse planning where speeding up the processing of spatial data and the completion of activities contribute directly to productivity;
2. This group covers routing and location analysis where in the past DSS design was predominately promoted by management science models; and
3. Marketing is an area where spatial data and modelling have presently somewhat neglected.

What one finds is that while many widely accepted definitions of DSS identify the need for a combination of database, interface and model components directed to a specific problem, a GIS would not constitute a DSS as it lacks support for the use of problem specific models. Common to all definitions of DSS is a sense that these systems must support a particular type of decision. Whilst there is evidence GIS software are becoming increasingly suitable as a generator for a SDSS, a GIS is not a complete DSS because of the almost complete absence of problem specific models or support for the organisation of such models. Therefore, SDSS can be seen as an important subset of DSS which potentials have yet been exploited.

But for it to be of interest, real world problems need a spatial component. For example if a GIS could model flooding along the banks of a river, it would assist planners to restrict development in those flood prone areas. In DSS applications, the focus of the decision-maker is on the decision being made where the output from the DSS is of interest only to the extent that it facilitates decision. Thus, GIS software must allow easy automatic interchange of data between GIS modules and modelling techniques operating on non-spatial elements of the data. In other words, the GIS software must make available data in a format appropriate for modelling techniques to be drawn from other disciplines. Unfortunately, this lack of integration hinders the comprehensive use of GIS as a SDSS.

M. Enache [1994] suggested that the missing link between sophisticated GIS and refined choice models, pattern-seeking systems and integrated decision support tools seem to be a planning/policy analysis methodology based on a planning/policy analysis theory. While SDSS generally include pattern-seeking models of geographical analysis, they would benefit from the integration of choice models in a flexible, decision research approach.

A good example how a GIS can be exploited as a planing tool to assist decision making is an Integrated Planning Decision Support System [IPDSS] developed by Mario Meja-Navarro in 1994. M. Meja-Navarro and L. A. Garcia [1995] indicated an IPDSS is a decision support system to assist government and communities in the evaluation of geological hazard, vulnerability and risk. It was also designed to assist a town planner in organising, analysing, modifying and re-evaluating existing or needed spatial information within land-use planning activities. The IPDSS incorporates GIS, Geographical Resource Analysis System [GRASS] and engineering numerical model within a Graphic User Interface [GUI]. This provides the user with dynamic user-friendly environment for modelling capabilities in vulnerability and risk assessment. IPDSS is designed to assess hazard such as debris flows, subsidence and could be applied to some hazards with probable maximum precipitation and seismcity as triggering factors for susceptibility scenarios It incorporates information on topography, aspect, bedrock and surficial geology, structural geology, geomorphology, soils, land cover, land-use, hydrology, sociology, precipitation, Federal Emergency Management Agency floodway maps and historic data to assess hazards. Regular items considered in vulnerability analysis include ecological sensitivity, economic vulnerability and social infrastructure vulnerability. Risk is assessed as a function of hazard and vulnerability.

Geological hazard were initially modelled based on subsidence and debris flows taking into account other factors like floods, rockfall and landslides. Landuse vulnerability assessment is carried out with consideration of community infrastructure taking into account the density of the local population affected while lifelines factors considered include the buffer areas built around road, water, telecommunication and electric lines. Risk assessment is considered the most important objective in decision-making for town planning because it involves the issue of human lives and the impact on the urban infrastructure for the community at large. In order to ensure user-friendliness, the IPDSS interface design is built in a way that the user applies the triggering factors directly through the hazard pull-down menu, clicking on the hazard of interest. The result is obtained interactively by pressing the button for the user’s interest, which activates a pop-up editor on the screen. The risk pull-down menu allows the user to selectively evaluate the geographical distribution of potential damages affecting social features selected.

2.8     Planning Rationale

Town planning traditionally has been concern for the improvement of the social conditions in towns and cities, in the eradication of social ills, overcrowding, poverty, ill health, unemployment, unsanitary and inadequate as well as inappropriate living conditions. However, the purpose for town planning within the loose context of physical planning has been somewhat more difficult to determine. Bruton [1974] referred physical planning as a physical design of something which already existed or might exist in the future and is a in a geographical or spatial representation of actual physical structures or elements. But if one looks at the essence of planning, one will agree with Ratcliffe [1975] that planning serves as a reconciliation of conflicting objectives: between social and economic aims, of public and private objectives.

As towns expand and develop, conflicts often arise as urban areas encroach into the rural areas. This is because land, in most cases, is limited in supply and a location that is lucrative for development on the opposite side of the coin, may be suitable for conservation. So the question arises, what is the optimal use of a particular piece of land? Town planning has always the difficult job of balancing between a sensible and acceptable blend of conservation and exploitation of land against the background of human activity. In order to achieve these utopian objectives, it is inevitable that there will have to be control of the layout and design of the urban settlement. This involves a systematic approach towards problem solving which invariably involves decision-making.

In Malaysia, town and country planning has been identified as a concurrent function of the federal and state government as stipulated in the Ninth Schedule of the Federal Constitution. At the federal level, the Federal Department of Town and Country Planning provides technical advise on all planning matters to the Minister of Housing and Local Government while at the state level, state planners corresponding provide technical advise on all planning matters to the Chief Minister and the State Authority. At the local government, the Town and Country Planning Act, 1976 [Act 172] empowers the State Authority to delegate the control and regulation of town and country planning to the local planning authority. Thus, it can be seen that town planning operates at different levels of organisation.

Act 172 requires the local planning authority to prepare a development plan for the purpose of guiding and controlling development and landuse. The local government is represented by a two-tier level of planning. The structure plan, a written statement of policies and general proposals outlines how land shall be to developed and utilised. It also states the relationship of proposals to the general proposals for development and use of land in the neighbouring areas that may be affected. The local plan in simple terms is a detail interpretation how the policies and general proposals of that structure plan are implemented at the local level. This shows that local planning is very site or lot specific. As new techniques are sort to cope with the fast changing trends, GIS is looked upon as a ideal tool to assist town planning because it holds many qualities suitable for planning: Common denominators include:

1. They both deal with the land matters;
2. They both handle a large volume of data; and
3. They both deal or are related with the process of problem solving.

Hidehiko Sazanami [UNDP 1990] felt GIS is an important tool in many kinds of development planning for 2 principal reasons:

1. It is unequivocally orientated to the spatial components of development and comprises the use of land for human settlement, land resources as well as the management and protection of natural environment; and
2. GIS has informational content and provides a mean to store, access and manipulate data and to utilise its informational content in a spatial context.

2.9     GIS and Planning

Planning and management are basically about problem solving and are based on the generic problem-solving process. This process begins with problem definition and incorporates various forms of analysis that might include simulation and modelling. Then this moves to prediction and subsequently to prescription or design that often involves the evaluation of alternative options. Clercq [1990] stated the synthesis of the planning process consists of five elements:

1. The problem to be met;
2. The opportunities and threats pertaining to the planning issues
3. The images of the situation, usually expressed by means of policy statements and sketch designs;
4. The options for decision-making, and
5. The decision-making process after considering the impacts on all the parties involved.

M. Juppenlatz & X. Tian [1996] also share the view that planning is a process encompassing the following steps:

1. Problem identification;
2. Goal Setting;
3. Data collection;
4. Refinement of goal;
5. Development;
6. Development of alternative plans and/or policies;
7. Evaluation of alternative plans and/or policies;
8. Adoption of the preferred plan and/or policy;
9. Implementation of plan and/or policy;
10. Monitoring and evaluation of results; and
11. Feedback.

Fig. 2.3 A GIS In Planning Practices

Source: Geographic Information System and Remote Sensing, 1996McGraw Hill Book Company

These steps reiterate planning is very much a tool for decision-making and decision-making occurs at every stage of the process. GIS too plays a crucial role in planning and decision-making as it can be found throughout the whole spectrum of planning process irrespective of their technicalities. However, M. Batty and P.J. Densham [1996] argued that in terms of planning and problem solving processes, to date there has been very little emphasis on formal analysis, simulation and modelling and hardly at all on design and decision-making aids. This view is also shared by C.A. de Brujin, S. Amer and D. Dougali [1996] who felt there is a lack of explorative and predictive analytical power of contemporary GIS software for planning applications that can interrelate and interact between people and their environment. Such situation is attributed as a result of:

1. the heavy dominance of information management in the GIS marketplace;
2. the absence of a coherence conceptual framework for spatial analysis; and
3. The complexity and obscurity of most analytic and the lack of expertise and experience with these technique among GIS users.

This implies that integration of sophisticated spatial analysis and GIS will remain an academic issue for some time. So far, little progress has been made probably because of the complexity of tight coupling of GIS software with analytic routine. Furthermore, if GIS is to be more popular among planners, simpler user interface have to be developed as planners generally have a comparatively lower level of computer skills than other technical departments. Still, Ahris [1997] felt that the use of GIS is valid and can generally aid strategic planning of an organisation in the implementation and monitoring of development projects and identified 5 specific areas:

1. Regional planning and resource management;
2. Urban development programmes;
3. Impact study of proposed projects;
4. Preparation of development plans at local government; and
5. Development control and urban development.

The Planning Data Model

The design of a planning information system is based on the clear understanding of the function, responsibilities of the various divisions of an organisation and the activities being practised in the department. The tasks and supporting data provide the fundamental framework which a conceptual model of geographic data entities and their relationship can be developed. Sepakat [1997] stated that the conceptual data model for planning is based on the land parcel as the central entity. This is because most planning processes are related to the lot parcel for the preparation of the formal development plan and development control of planning applications. Hence, the basic geographical layer for planning purpose is the cadastral map. Such a model represents the basic component for the Town and Country Planning Department.

But as town planning is a multi-disciplinary profession, the collection of relevant data from other sectoral studies is fundamental in town planning. This point is even stipulated as a legal requirement under Act 172 which require development plans to examine matters among others, economic, sociology, transportation, communication and environment. Consequently, the components of the planning data model [See Fig. 2.4] incorporates the following modules:

1. The Base Map;
2. Land Records Module;
3. Town Planning module;
4. Utilities module;
5. Environment module;
6. Transportation module;
7. Trade module; and
8. Administrative module.

Fig. 2.4 Planning Data Model

Source: Information System Planning Final Report for the Federal Department for Town and Country Planning.

An example of a planning model adopted locally is illustrated by the SUMBER-PUTRA system currently developed at Perbadanan Putrajaya. [See Fig. 2.5] This system is a large local computerised system designed to ensure effective administration of the many activities for the utilisation of land and construction of buildings at Putrajaya. The system platform for the SUMBER-PUTRA system is called the Enterprise Wide System links all the operational sub-systems together as a unified system and provides user interface, access control and security, workflow management and communication between users. The SUMBER-PUTRA system comprises over forty application sub-systems in which town planning activities are managed by the City Planning Department under three sub-systems: the Development Planning, Planning Permission and Land Management sub-systems.

Fig. 2.5 SUMBER-PUTRA Phase I Chart

Source: User Requirements & System Design Specification, SUMBER-PUTRA System, 1998

The Development Planning sub-system is a database that contains all relevant graphic and textual data relating to development planning and focuses on three different data sets:

1. The existing data which provide the present graphical and non-graphical information of Putrajaya;
2. The Development Plan which contains the published master plans and related guides; and
3. The Urban Design Guidelines which contain the overall design concept and detail development guidelines of Putrajaya.

Design of the database comply with the requirements of the Town and Country Planning Act, 1976 [Act 172], the Putrajaya Structure Plan and part of Sepang District, local plans and Urban Design Guidelines. The Planning Permission sub-system is designed to assist processing of planning permission applications and provides a broad information base and retrieval tool for landuse studies and other land specific evaluation. It performs three main functions:

1. To identify and ensure elements mentioned in checklists appear in the submitted plans;
2. To ensure the total development area and units submitted by the developer are correct; and
3. To ensure elements proposed by the developer conform to the structure plan and Urban Design Guidelines.

The process can cover a variety of applications:

1. Applications for layout plan, pre-computation plans, the erection of buildings and extensions to planning permission and temporary buildings;
2. Additions and alterations to residential buildings which also include existing residential buildings; and
3. Change of use, which covers applications to change the usage of a building, parts of a building or individual units.

Because these activities are closely inter-related, the Planning Permission sub-system interfaces with the following sub-systems:

1. Billing and collection ;
2. Development planning;
3. Land management;
4. Building control and legal database.

To simplify the planning permission process, a Planning Permission Expert System Ver. 1.0 was designed and represents a sub-module of the Planning Permission system. It automates the checking routine of development plans but is only effective if plans are submitted in the correct format. General rules ensure no contraction between the proposed development and the requirements of the master plan and the Urban Design Guidelines.

In short, the planning permission process is complex and covers a variety of applications. The success of sub-modules like the Planning Permission Expert System is largely dependent on an early adoption of a standard format for all layout plans to adhere to.

2.10     Factors leading To the Successful Implementation of A GIS

While developing a GIS can be very challenging, it has to be acknowledged that problems still persist because the system has not yet been perfected. However, many problems have been identified and stem mainly from the very nature of GIS, the way it handles huge volume of spatially related data, the composition and nature of the data itself and the wide spectrum of different organisations using it. R.G. Newell & D.G. Theriault [1998] stated these problems could be categorised as data capture performance, customisation and integration. They highlighted key fundamental problems which manifested from the above:

Topology Capture Problem

Although progress to increase data capture efficiency, save cost and reduce time have been made through scanning and vectorisation technology, the generation of meaningful topology, the choice of user data and the addition of feature information is very much a human function. It seems the only way out is to put more time and effort into this problem.

Large Data Volume

The relational database system proves popular and effective for storing and retrieving data but it is questionable whether it can adequately perform analysis as GIS often involve a large volume of data and its query capability is quite tedious when used simultaneously. Alternative solutions are possible but do not solve the problem spot on. A system that can handle spatial and aspatial data as one entity similar to object–oriented GIS software would sound attractive but still remain a problem for those committed to a non object–oriented GIS software.

Accommodating Existing Database

This brings to the next point when organisations of a particular database system wish to optimise their database for GIS purpose. For example, most local developers and consultants in the building and construction industry usually store their projects in CAD files because layout and survey plans are traditionally CAD based. As they work are related to planning submission they may want to capitalise on GIS to make a strong impression when making presentations for plan submission to the local authority. In this case, it would be wiser and worthwhile to ensure that one’s investment in database is adaptable for GIS.

Continuous Mapping

The real world is a single, continuous, seamless landscape. For GIS to provide an accurate representation of the real world, it should portray that same environment. This can achieved by a process known as continuous mapping. However, the transformation from original hard copies of maps and plans to digital versions has been popular by the map sheet approach because of similarities to the conventional filing system. Unfortunately, this approach creates problems during query and analysis because their geometry, topology and relationship between graphics and application data cannot be said truly integrated.

Version Management

While GIS and database management systems normally have a locking system to preserve the integrity of the database, it would be ridiculous to enable that to function at all times as this would not make it accessible for multiple users. In GIS design database, a transaction could take quite a while to realise and it would not be practical to lock the system. Versioning that among others covers the management of alternatives, chronology and policing of change has not been addressed in current DBMS.

Vector-Raster Database

Vector based GIS tend to separate from raster based image system. Image processing systems have a key problem extracting information from the sheer bulk of data from remote-sensing but vector based systems have problems trying to extract information from their data capture. Rather than dispute between is the better [sources], both sources should be integrated as one in a vector-raster database and capitalise on the best of both worlds.

Overlay Analysis

Overlay analysis would be extremely tedious without GIS when used in a multi-thematic base map. However, if data had not been thoroughly cleaned, the robustness of the database breaks down. When themes are combined, small polygons called silvers may occur and its removal is costly and time-consuming. One also has to be cautious about the level of accuracy of the results because the combined effect of individual errors may be difficult if not impossible to gauge.

Front End Language

Every user has different needs in developing his system. Although customisation tools are available, it may take many man-months to make modest changes which is so ineffective. What is needed is a system where all data and functions can access and manipulate seamlessly. While it can be argued that object-oriented programming is still new and may take a while before it becomes fully develop, its concept looks right as the language has a more friendly syntax and enables casual users to write simple programs for customisation and usable in interactive mode.

Query Language

Modern query languages such as SQL may be good for simple selective queries but can be difficult and clumsy in complex cases. However, problems in constructing queries for both databases are anticipated in situation where two databases are used in parallel where one driven by SQL and the other a geometry database. Much work needs to be carried out, as they have not fully matured.

H. Sazanami [UNDP 1990] said that GIS was recognised for having much potential at improving geographic information for planning but lack an effective means for interrelating planning with GIS to develop practical applications and models for planning. He said that while GIS has the capability to support a wide range of planning applications, information retrieval, site selection, environmental impact assessment; land suitability and transportation network modelling, GIS is weak in designing decision support systems and expert systems applications. From the point of view of using GIS for planning, he felt many issues remain unresolved:

Data