GIS QUALITY MODEL

A Metric Approach to Quality Models

Willington Libardo Siabato Vaca

Universidad Pontificia de Salamanca, Facultad de Informática, Madrid, Spain

Adriana Patricia Rangel Sotter

Universidad Pontificia de Salamanca, Facultad de Informática, Madrid, Spain

Keywords: Software Quality, ISO/IEC 9126-1, Quality Models, Metrics, GIS, Software Evaluation.

Abstract: In the past few years organizations and companies have developed new standards which have been proved

highly efficient both in the public and private sector. The most important of them is the Quality

Management System regulated by the standard ISO 9001. This standard proves how its implementation

represents sensitive improvements of the production system, optimizing the product and increasing its

quality. This phenomenon is not isolated within software engineering's frame. Many works have been

published, like Boehm and McCall, of which have raised many standards among which the ISO/IEC 9126 is

highlighted. Regarding this fact, it has been possible to create different solutions for multiple related

problems with IT. Nowadays, the Geographic Information Systems' project managers do not have a tool for

either selecting the software to implement their projects or supporting this selection in technical criteria. The

questions are: which one of the commercial software packages is appropriate to my project? Which one of

the software packages follows the requests of the project out? Which one of this software supports the needs

of the users? This article presents a quality model to support these decisions. This way, project managers

can make their decisions based on a set of metrics which are product of the deep evaluation of

characteristics, subcharacteristics and attributes of the software. These metrics has been developed to apply

for all models based on ISO/IEC 9126-1 standard. The mentioned elements allow user to know which of the

software packages is the best through a GIS Quality Indicator, generated through the model. This indicator

allows GIS’ project managers to take decisions based on a technical criterion. A model in accordance with

international standards related to product quality in software engineering such as ISO/IEC 9126-1.

1 INTRODUCTION

New trends to utilize the cartographic information

for evaluating resources and territorial planning

emerged in the sixties and seventies. It was noticed

that different coverages on the surface of the earth

were not independent, rather they kept some kind of

interdependence. The need to evaluate them in a

more efficient way was self-evident. At first, the

used methods were relatively simple namely,

transparent copies of maps were superimposed on lit

tables and points of coincidence in the distinct maps

were found. At a later time, this technique was

applied to emerging Information Technologies -IT-

and simple maps were created by means of overprint

of characters to produce different tones of grey.

Nevertheless, specialists did not find these

methods extremely useful and they were not

accepted by the professionals that produced, updated

or used cartographic information. In the late

seventies, the IT to generate cartographic

information progressed rapidly and it was tuned

many of information-technology systems for distinct

cartographic applications. In the same way, progress

was being made in related sectors namely:

photogrammetry and remote perception. Initially, the

fast development meant the duplication of efforts in

different areas relating to cartography. With constant

improvements to the systems, developers acquired

experience and the possibility to use different kind

of tools for working with spatial information had

been raised. The creation of these systems

contributed to the creation of solid Geographic

Information Systems -GIS- for general purpose. In

176

Libardo Siabato Vaca W. and Patricia Rangel Sotter A. (2007).

GIS QUALITY MODEL - A Metric Approach to Quality Models.

In Proceedings of the Ninth International Conference on Enterprise Information Systems - ISAS, pages 176-183

DOI: 10.5220/0002388701760183

 SciTePress

the early eighties, GIS had become a completely

operating system whilst the technology of computers

had started to grow and to develop rapidly.

Currently, GIS are being implemented in

vertiginous form on public agencies, laboratories,

research institutes, pedagogical institutions, private

and military industry. This big introduction of the

GIS has given rise to a need for the users of the

geographic information to know this technology in a

better way. Therefore, it is necessary to help them to

understand the GIS software through solid and

reliable tools which must be based on methodologies

created and developed for this specific objective.

This article describes a quality model for

selecting and verifying GIS software that will be

utilized as a platform for a specific project. This

software has to be reliable, fulfil the project

requirements analysis and satisfy the needs of both

the project and the user. This way, the paper presents

an innovative solution to help GIS professionals to

support the chose of the platform software on a

quality model created for this specific objective. A

model that solves a real problem which has not been

worked on yet, a problem that is observed in the

day-to-day routine of institutions and companies.

2 DESCRIPTION

Models are abstractions of reality and are designed

to make process easier. They offer to users the

adequate tools to interpret complex realities by

isolating and focusing on the principle components

and removing elements that do not affect the final

results meaningfully. Specifically, quality models

offer a series of elements that user must take into

consideration to insure that accomplished actions are

adequate to obtain the desired results. This way,

different kinds of variable are required in order to fit

the models up to the intrinsic and extrinsic

characteristics of the elements that will get involved

within the developed model or models.

Nowadays experts and beginners engaged in

development and implementation of information

systems, have many problems to select a GIS

platform namely: the big supply of GIS products, the

multiple versions available of one product, new

products or additional extensions to have more tools

and specific functions. These are enough causes the

users become confused, this way, a simple decision

such as to select a GIS platform software turns into a

difficult job.

When GIS projects are planned, objectives are

defined and user's requirements are identified; one of

the most important stages for their design and

development is to select the platform which it will

be implemented over. This decision implicates

investment of resources, run times, accomplishment

of requirements and needs, adequate functionality

and the customer satisfaction. For GIS' domain, it

does not exist a tool, methodology or model that

permits professionals to solve this problem. They

neither have enough support to bear his decision in

theoretic well-grounded concepts. Therefore, it is

evident the need of creating some method in order

that this decision is not based exclusively on project

manager's experience or, as it occurs for the most of

the cases, taking into account the most popular GIS

software without evaluating more alternatives.

The mentioned deficiency constitutes a problem

that needs to be solved. This solution has to support

one of the most relevant processes in the project

cycle, due to the fact that it affects deeply the

success or failure of the project.

Taking into consideration what's been said up to

this point, it is possible to define the main objective

of this article. Of course, it is to define a quality

model for GIS software packages, that permits

identifying through robust metrics which is the best

GIS software for an specific project and its

implementation within a frame of specific conditions

and under a set of specific requirements. Three

fundamental points would be mitigated with this

model: (i) speeding up the process of selection of

platforms and tools of the system, (ii) offering a safe

methodology to guarantee that the selected tool

fulfils the minimal requirements of system and user,

(iii) bearing the development of information systems

on standards and quality models designed and

developed specifically for this objective.

2.1 Quality Models

The first topic that must be discussed in this article it

is what quality means and how this characteristic is

integrated in information systems and software

packages. Just as quality has evolved, its concept has

borne several transformations over the time. Thus,

different definitions have been emitted by working

groups dedicated to its study. Formal meanings that

take into account the human dimension are

suggested by Dr. Joseph M. Juran (Juran, 1995):

quality consists in freedom after deficiencies; quality

refers to the absence of deficiencies itself; quality

consists in the product’s characteristics that are

based on the customer's needs. Other important

definitions are (Crosby, 1991), (Feigenbaum, 1991),

(Taguchi, 2004) and (ISO, 2000), that defines

quality as “set of properties or characteristics of

something (product, service, process) that made it

GIS QUALITY MODEL - A Metric Approach to Quality Models

177

apt to satisfy needs.” This definition not only refers

to the characteristics of the product or services itself,

rather introduces other aspects that can be shown in

the final service.

Analyzing the above definitions the quality

concept that will be taken into account to develop

the model exposed in this paper is the set of product

or service’s characteristics that have the ability to

satisfy the user's needs and expectations, permitting

to judge its value based in a set of attributes and

intrinsic properties, within a frame of reference well

defined.

2.1.1 Software Quality

Nowadays it is clear how computers and software

are utilized for a wide range of fields and

applications. That is why, development and selection

of high-quality software products are relevant, even

though it is considered that its development and

implementation in a right way implies the success or

failure of the processes that are borne on these tools.

Thus, the specification and the extensive evaluation

of the software quality is a key factor to ensure an

adequate quality and the success of the tasks based

on software products.

ISO (ISO & IEC, 2001) suggests that it is

important that each relevant characteristic of

software quality be specified and evaluated, using

valid and widespread metrics as far as possible.

Software producers are responsible for that these

characteristics are identified in order to define the

metrics that will permit to know whether an element

or attribute of the product is acceptable. Thus,

several elements will be considered both in the

process of development and the use of software.

ISO/IEC 9126 and ISO/IEC 14598 have been

developed considering these characteristics; in

addition associated metrics can be used not only for

evaluating software products but also to define

quality requirements and other uses. ISO/IEC 9126

was created for the specification and extensive

evaluation of the software product quality taking

into account metrics, specifying relevant quality

characteristics and describing a model for

production of software products from the point of

view of internal and external use. ISO/IEC 14598 is

related with the software product evaluation.

Considering that the proposed model aims

supplying a tool that permits users to select among

Table 1: Standard ISO/IEC 9126.

Standard Objective

ISO/IEC 9126-1 Quality model

ISO/IEC 9126-2 External metrics

ISO/IEC 9126-3 Internal metrics

ISO/IEC 9126-4 Quality in use metrics

the different bidding of the market, the software

which adjusts to specific user needs and fulfils the

entire quality requirements for GIS projects in a

specific domain. The developed model relies on the

ISO/IEC 9126 family standard and specifically in

the ISO/IEC 9126-1 quality standard due to the fact

that it lets to create hierarchies of quality features,

which are essential for building structured quality

models besides that it is a widespread standard.

2.1.2 ISO/IEC 9126 Quality Standard

ISO/IEC9126 is a family of standards that regulates

the software product quality taking into account:

models which are conformed by, internal and

external characteristics, the method to measure these

characteristics and the functionality of the proposed

model. The standard is conformed by four parts that

share the same general title: Information

technologies – Software engineering – Product

quality.

Xavier Franch (Franch, 2003) says that ISO/IEC

9126-1 specifically addresses quality model

definition and its use as a framework for software

evaluation. A 9126-1 quality model is defined by

means of general software characteristics, which are

further refined into subcharacteristics, which in turn

are decomposed into attributes, yielding a multilevel

hierarchy. At the bottom of the hierarchy are

measurable software attributes, whose values are

computed using some metrics, which are defined

and regulated by ISO/IEC 9126-2 and ISO/IEC

9126-3 standards. Internal metrics quantify the

software’s characteristics, while external metrics

measure the general behaviour and performance

implicating the system in which the software is

implemented. Finally, quality in use quantifies the

effects of using a software package in a specific

context; this is regulated by the standard ISO/IEC

9126-4 taking into consideration characteristics and

subcharacteristics.

The proposed quality model has been developed

exclusively taking into account the external quality

elements, because it is there where the end user

interacts directly with the final product and it is from

this that the user defines his quality perception. In

other reference frames, this perception can feed back

to the internal quality and create an ever improving

ICEIS 2007 - International Conference on Enterprise Information Systems

178

cycle of product. Nevertheless, this element was not

considered within the development of this work due

to the fact that it is focused exclusively on the end

user. The objective is to reach the necessary quality

to satisfy the real users’ needs.

2.2 Referenced Models

The software product quality should be evaluated

using a well defined quality model. In practice it is

not possible to measure all the software's internal

and external subcharacteristics due to the fact that

software products are normally too big. Also, it is

not practical to measure the quality in use for all of

the possible user-task scenarios. The resources for

evaluation must be assigned to the different kinds of

measurements depending on the objectives of the

business and the nature of the product or the design

process. ISO proposes evaluating the software

products quality based on a set of characteristics and

subcharacteristics of general interest.

Table 2 shows the six quality characteristics

defined in the ISO/IEC 9126-1 quality standard and

their decomposition into subcharacteristics. In

addition, subcharacteristic compliance is included

for all of the characteristics. These elements support

the proposed model. Combining this model with the

one proposed by Franch, it is possible to define an

adequate evaluation frame for GIS.

Quality models considered for the development

of this article are the ones that for its aim permit to

evaluate the performance of any application or to

create some kind of metric.

In the study accomplished by Chirinos (Chirinos

et al, 2003), he intends a requirements classification

which takes into account the quality views from the

first stages of development to provide the quality

requirements identification. The authors develop a

requirements classification model based on views of

quality with the aim of providing the quality

requirements identification.

Another interesting work is developed by

(Losavio, 2002). She proposes a form to specify the

relevant quality attributes implicated in the design of

architectonic process. An additional model to take

into account was carried out by (Calero et al, 2004),

whom developed the Web Quality Model –WQM–.

The main referenced model is developed by

Franch and Carvallo (Franch, 2003), quality models

in software package selection. This work proposes a

specific methodology to make structured quality

models to select software package involving

software's description and functionality. The model

Table 2: ISO/IEC 9126-1 standard.

Characteristics Subcharacteristics

Functionality Suitability, Accuracy,

Interoperability, Security.

Reliability Maturity, Fault tolerance,

Recoverability.

Usability Understandability, Learnability,

Operability, Attractiveness

Efficiency Time behaviour, Resource

utilization.

Maintainability Analyzability, Changeability,

Stability, Testability.

Portability Adaptability, Installability,

Coexistence, Replaceability

comprises six steps: defining the domain,

determining quality subcharacteristics, defining a

hierarchy of subcharacteristics, decomposing

subcharacteristics into attributes, decomposing

derived attributes into basic ones, stating

relationships between quality entities and

determining metrics for attributes.

It is in the determination of metrics for attributes

where the proposed model focuses its develop and

underlies one of the main contributions (since

current methodologies does not define real metrics),

and further on, a global quality indicator which

defines an absolutely quantification for the package

software quality evaluation.

3 MODEL DEFINITION

Through the proposed model, it is possible to know

which of the available software packages fulfil

minimal and necessary requirements in desired

conditions, and which of they get the higher

evaluation for each characteristic (dimensions)

defined in the model. This evaluation will permit

discarding those tools that not fulfil minimal

conditions, as well as reducing the possibilities to

determine a final decision based on another kind of

criteria like the cost-benefit ratio, because it is

possible that tools with a huge difference in their

cost, supply the same functionalities and guarantee

the same reliability in some designing conditions.

3.1 Characteristics &

Subcharacteristics

The flexibility of the methodology proposed by

9126-1 standard is unequivocal, their components

are not a straightjacket for the definition of the

model in a specific dominion but they constitute a

good starting point. This way, when the components

GIS QUALITY MODEL - A Metric Approach to Quality Models

179

Table 3: Modified characteristics and subcharacteristics.

Characteristics Subcharacteristics

Functionality Compatibility.

Availability.

Reliability Quality data verification.

Scalability.

Usability Complexity of personalization.

Usability compliance.

Efficiency Data protection.

Efficiency compliance.

Maintainability Maintainability compliance.

Portability Portability compliance.

and the modifying dimensions are examined

according to the needs of GIS software, some

subcharacteristics have been eliminated, others have

changed its meaning and others have been created

defining new concepts according to GIS software

packages. The result is showed in Table 3.

In general all the subcharacteristics definitions

are according to what is indicated by ISO/IEC 9126-

1. The subcharacteristics compliance for usability,

efficiency, mantainability and portability are

eliminated for being considered irrelevant for GIS

software. The six new incorporated elements try to

refine the model for a complete evaluation.

For the dimension functionality, the

subcharacteristic compatibility is defined like the

capability of the software to interchange data and to

maintain projects with others software package of

the same type. The subcharacteristic availability

evaluates the licenses and license administrator to

verify if the software guarantees the license service.

Data verification permits to verify input and output

of data quality. Scalability checks the software

capability to be adapted without problems to a

harder work as a result of new users addition,

increment of the traffic or execution of new

transactions. An important GIS subcharacteristic is

the complexity of personalization, defined like the

set of software attributes that determine the

capability and facility of software's personalization

for specific tasks. Finally, data protection compiles

the attributes for users’ administration and how the

data access is administrated in the main system.

3.2 Definition of Types of Measures

Once the model’s dimensions have been defined and

before defining the attributes that describe them, it is

necessary to indicate the types of measures that the

user will use to quantify attributes. For its definition

some points must have considered:

 Types of measures have to be represented by

quantitative elements to be able for operations.

Table 4: Type of measures for attributes quantification.

Measure Type Dominion Unit Symbol

Eyewitness Boolean 0-1 P

Time Integer s/d/m T

Level Integer 0-4 N

Ratio Integer 0-100 % R

 If the type of measure is represented by

qualitative elements, it has to be changed to

quantitative elements.

 For some types of attributes it is not enough to

express the measure of his behaviour with

single elements like boolean or integer, in this

case the attribute require a function to express

in a best-suited form its behaviour.

Taking into account Vallecillo and Bertoa’s work

(Vallecillo & Bertoa, 2002), the proposed model will

use the types of measures defined in Table 4. Each

one of these types will be utilized to quantify the

model defined attributes in following cases:

 Eyewitness (P) indicates if an attribute exists.

 Time (T) measures spans.

 Ratio (R) expresses a specific percentage.

 Level (N) indicates a grade of effort, ability.

The model establishes a five-level classification

method for the Level type according to Table 5. The

Eyewitness type is defined through a boolean value.

Table 5: Measures description.

Measure Value Scale

0 Very low

1 Low

2 Medium

3 High

Level

4 Very high

0 – False Exist Eyewitness

1 – True Not exist

Time and Ratio types are expressed in seconds

and percentage respectively. The percentage shows

how much the software gets close to the fulfilment

of a requirement. In cases like installation times and

configuration as well as capacitating and learning

can change the time unit second to day or month.

For operations of attributes quantified through

Time type, it is necessary to transform from Time to

Level, this way the metrics operation becomes more

efficacious. This transformation is done according to

Table 6. Each attribute needs to be transformed in a

different way. The attributes A, B, C & D

correspond to: duration of the product in the market,

efficient use, adequate configuration and efficient

administration. The values correspond to the same

scale defined in the Table 5.

ICEIS 2007 - International Conference on Enterprise Information Systems

180

Table 6: Attribute reclassification for Time measures.

Value Atributte

A (Year)

Atributte

B (Mth)

Atributte

C (Day)

Atributte

D (Mth)

0 < 5 < 7 > 5 < 4.5

1 5 - 10 6 - 7 5 4 - 4.5

2 11 - 20 5 - 6 4 3.5 – 4

3 21 - 30 4 - 5 3 3 – 3.5

4 > 30 < 4 <= 2 < 3

Once the dimensions and types of measure have

been defined it is possible to start the definition of

attributes, the main component of the GIS quality

model. The attributes correspond to the way that the

model's characteristics are described and a quality

property which is possible to assign a measure.

Through an analysis of GIS tools, compiling GIS

experts’ concepts and experiences as well as the

observation of projects, 370 attributes have been

defined for the dimensions. The complete list can be

consulted in the physical model or (Siabato, 2005).

3.3 Metrics Definition

The measures, which rate the attributes that describe

the dominion defined for the characterization of the

model, are computed through metrics. Two kinds of

metrics have become established in the model:

specific and general metrics. Specific metrics are

defined to deliver a general evaluation of the

evaluated software, in this sense, it will be a metric

which will compute not measured but other metrics

derived of the model such as general metrics, which

are defined as the elements that will be utilized to

quantify each one of the dimensions of the model.

3.3.1 GIS Quality Indicator

As was said, the model intends to show an indicator

that represents in a simple way the GIS software

quality evaluated, a specific metric is established (1).

∑

WNI

(1)

Where Ni represents each one of the dimensions

which conforms the model and Wi is their weight. In

order to define the Wi weights a group of GIS

experts replied to an opinion poll which tried to

classify the level of the dimensions. From the

obtained values and pondering the results the Wi

weights are described in Table 7.

Considering that grades granted by experts are

rated in the range 0–10 and they must be normalized

to the model's dominion (0-1); it is necessary to use

the relation described in (2).

Table 7: Weights for dimensions.

Dimension Value Weight

Functionality 0.26 W

Reliability 0.21 W

Usability 0.17 W

Efficiency 0.10 W

Maintainability 0.16 W

Portability 0.10 W

∑

iii

χχ

(2)

3.3.2 Subcharacteristics Evaluation

To define the metrics of the dimensions it is

necessary to take into account each one of the

subcharacteristics which compose them. Unlike the

definition of the indicator I, each subcharacteristic

will be calculated with the same weight.

A method according to the type of measure P, T,

N or R is determined for the evaluation of each

subcharacteristic. The dominion of evaluation is

established in the range 0 -1, 1 represents a total

fulfilment of the subcharacteristic and 0 represents

its absence. Taking into account these premises,

metrics M

, M

and M

are defined

according to the type of measure.

For evaluation of the type P attributes the

expression (3) has been defined, where Xi is the

grade P established by the user and n corresponds to

the number of attributes for the evaluation of each

subcharacteristic. With this metric is possible to

evaluate attributes such as applicability, accuracy,

interoperability, security and availability.

iTP

∑

(3)

For quantification of type R measures, the

previous method is valid if it is considered that the

dominion of evaluation R: 0-100 is comparable with

P: 0-1. The expression (4) defines the metric for the

evaluation of type R attributes. Subcharacteristics

such as compatibility and functionality compliance

are quantifiable with this method.

100

∑

(4)

For type N measures, the method to calculate the

grades is more complex. Due to the evaluation scale

(Table 5) it is possible to define the work dominion

in R5. Therefore, the best-suited the way of

quantifying attributes related to this type of measure

is applying the Euclidean rule. The expression (5)

represents the metric for type N measures. Where, Xi

GIS QUALITY MODEL - A Metric Approach to Quality Models

181

represents the grade N established by the user

between 0-4 and the element n, just like (3), is the

number of attributes for each subcharacteristic.

iTN

∑

(5)

Finally, for the quantification of type T measures

and keeping in mind the reclassification showed in

Table 6, which the dominion of evaluation for N is

homologate to type T measure, it is also possible to

apply the expression (5) to type T measures, taking

into account the respective reclassification process.

The metrics defined until now allows finding

values in ideal conditions in which all attributes are

graded with the same type of measure. Nevertheless,

there are some subcharacteristics that combine three

or more types of measures, this force to define a new

method for computing a single value for each

subcharacteristic. Ad hoc, it must be taking into

account that the final result belongs to the dominion

of the model, defining a general case that allows

finding the final value independently if the

subcharacteristic is quantified with two or more type

of measures. The expression (6) is the best suitable

relation to solve this problem.

∑

(6)

Where M

, defined as the metrics of general

type, is the general evaluation of the

subcharacteristics. Xi represents the metrics N, R, T

ó P established for the attributes that define the

subcharacteristic and n is the number of different

measures N, R, T ó P implicated in the evaluation.

3.3.3 Quantification of Dimensions

When the defined metrics are applied to the

subcharacteristics that compose the dimensions of

the model, a value between zero and one is obtained.

This value represents the level of quality for each

quantified item. This information can be useful for

some user that requires comparing a

subcharacteristic in particular. However, this

information is only an intermediate product utilized

for the final purpose, the presentation of the GIS

quality indicator (1). Once the sub-characteristics

were defined, the last step is to define Ni. Keeping in

mind that each one of obtained elements has been

normalized, in order to obtain each Ni it is necessary

to average the subcharacteristics which compose it.

(7) represents the expression to obtain each Ni value.

∑

(7)

Where, u

represents each one of the dimension

subcharacteristics, and n is the number of related

subcharacteristics. This way, the metric for each

dimension is obtained and the user will be able to

evaluate the result according to his needs and the

ones belonging to the project. In addition, we have

all of the elements to calculate the Indicator (I)

useful to determine which of the evaluated software

adjust better to the project.

3.4 Metrics Representation

The elements that must be represented are the GIS

quality Indicator and each one of the metrics which

are utilized to evaluate the dimensions. Two

methods of presentation have been established in

order to show the obtained results to the user:

 Numerical method. A numerical value that will

be utilized to represent the GIS Indicator (1)

and to compare the general evaluation of each

software package evaluated.

 Graphical method. A six-branch graphic that

will be utilized to represent each dimension.

The user will be able to evaluate which one of

the dimensions has the biggest evaluation.

4 SOFTWARE EVALUATION

There are multiple companies which offer different

solutions for the implementation of GIS projects.

The most outstanding are Intergraph

, MapInfo

Autodesk

, MicroImages

, Smallworld

, Bentley

and ESRI

. Once the model has been defined, three

GIS software will be evaluated in order to validate

the proposed model. The three evaluated GIS

package are ArcGIS

ArcInfo 9.0 from ESRI

TNT 6.9 from MicroImages

and Geomedia

Professional from Intergraph

To validate the model the set of metrics have

been implemented in a Microsoft

Office Excel

book. In this book, the user rates each attribute in the

dominion that has been defined for each type of

measure. Each sheet has the changes of scale and the

necessary operations to generate the metrics.

The results of the finished evaluation for the

mentioned software are shown in Table 8 and Figure

1. The showed metrics will permit project managers

make the best decision based on the metrics for each

subcharacteristic and the GIS Quality Indicator.

ICEIS 2007 - International Conference on Enterprise Information Systems

182

Figure 1: Metrics representation.

The results indicate that the best performed

software is ArcGIS

, followed by TNT and

Geomedia

. It is necessary to underline that this is

a general evaluation and takes into account the

global performance of the evaluated software and it

is not an evaluation based on specific conditions.

However, offers a good idea about which of the

evaluated software has the best performance. But it

is not possible to say that ArcGIS

is better than

the other ones. Figure 1 shows such the software

with the best functionality is TNT. This implicates

that if the user is looking for a package with specific

functions and properties to satisfy his needs, the

adequate election is TNT. This type of analysis is

possible because of the independent evaluation of

each dimension, e.g. if the user needs a highly

adaptable software the right election is ArcGIS

5 CONCLUSIONS

It is evident how any GIS project must select each

one of its components in the right way. This project

has covered this problem for the software platform

on which the project will be implemented. Taking

into account that exists many kinds of products, the

aforementioned platform must be selected with

technical criteria and keeping in mind users’ needs.

The proposed model constitutes an advance for the

definition and selection of GIS software packages,

based on international standards and focused on the

fulfilment of users and project requirements. This

model helps to GIS managers to select the platform

based on technical criteria and a safe methodology.

This a pioneer model, even though it is based on

existing methodologies, it does not exist quality

models related in the GIS dominion. Besides that,

this work proposes a set of completely innovative

metrics which can be applicable to any quality

model derived from ISO/IEC 9126-1.

Table 8: Results of the finished evaluation.

Dimension ArcGIS

Geomedia

TNT

Functionality 0,900 0,892 0,968

Reliability 0,932 0,811 0,840

Usability 0,918 0,850 0,856

Efficiency 0,855 0,862 0,868

Maintainability 1,000 0,809 0,676

Portability 0,772 0,780 0,905

Indicator (I)

0,909 0,840 0,859

The developed model considers the typical

evolution between measure, metric and indicator.

Each one of the attributes derived from the GIS

software analysis is quantified through a measure.

These measures are processed through a set of

algorithms which let to know an overview of the

evaluated software. Finally, each metric generated

for each dimensions is processed to generate the GIS

Quality Indicator, main objective of this work.

This paper had defined a new methodology

which can be used to support the GIS projects design

process on international standards, incorporating this

type of projects to international quality standards.

REFERENCES

Calero, C., Ruiz, J. & Piattini, M. (2004) "A web metrics

survey using WQM", In: Fourth International

Conference on Web Engineering, Munich - Alemania

Chirinos, L. (2003) "Una clasificación de requisitos

basada en vistas de calidad", In: Jornadas Chilenas.

Crosby, P. B. (1991) La organización permanece exitosa,

Mc. Graw Hill, México D.F. – México.

Feigenbaum, A. V. (1991) Total Quality Control, Mc.

Graw Hill, New York – USA.

Franch, X. & Carvallo, J. (2003) "Using Quality Models in

Software Package Selection", IEEE Software, 20 (1)

ISO (2000) International Standard ISO 9001:2000 Quality

systems - Model for quality assurance

ISO, IEC, (2001) International Standard ISO/IEC 9126-1.

Software engineering -Product quality- Part 1:

Quality Model, ISO/IEC 9126.

Juran, J. M. (1995) Juran y la planificación para la

calidad, Diaz de Santos, Madrid – España.

Losavio, F. (2002) "Quality Models to Design Software

Architecture", Journal of Object Technology, 1 (4) 165

Siabato, W. (2005) Modelo de calidad para paquetes de

software SIG, MSc. Tesis. U. Pontificia de Salamanca.

Taguchi, G., Chowdhury, S. & Wu, Y. (2004) Taguchi's

Quality Engineering Handbook, Wiley-Interscience,

New York – USA.

Vallecillo, A. & Bertoa, M. (2002) "Atributos de calidad

para componentes COTS", In: IDEAS 2002, La

Habana - Cuba 352-363

GIS QUALITY MODEL - A Metric Approach to Quality Models

183