Abstract:
The increasing usage of software in all fields of life, including safety-critical departments
of organisations has necessitated the need for the development of quality software. Although
software quality is of paramount concern to software development, it may be a challenging
task to software developers as it depends on ensuring that developed software meet the
standards of software quality design. Different quality models have been proposed by
researchers to serve as a benchmark for software quality design but most of these models
are tailored towards specific projects’ needs, hence, the need for the generic quality model
suitable for evaluation of all software projects. In this research, eleven (11) main software
quality attributes and thirteen (13) sub-attributes were identified and used for the quality
assurance model. These were attained from ten well-known standard software quality
models to rank the quality attributes. The Analytic Hierarchy Process (AHP) was used to
rank the software quality attributes and maintainability was observed to have the highest
score while cost had the lowest score. Mathematical models of the quality attributes were
used to evaluate software and the scores attained by each attribute was modeled with a
voting model and multiplied with the criteria weight from the AHP to get the overall quality
score for each evaluated software. Access to the quality assurance model was restricted by
the development of an access control model with the use of the Bell-LaPadula and Biba
model to regulate the people who use the model. To perform software evaluation, the
application must be hosted online, hence, must have a domain name. As a result, twenty eight web-based applications, grouped under six (6) categories, namely, Educational, E commerce, Company, Document Creation Software, Video editing, and Form creation web
applications were evaluated. Results from the evaluation showed that Document Creation
Software had the highest average quality score of 96.21% while Educational web
applications had the lowest average score of 84.16%. To evaluate the performance of the
software quality assurance model, recent works were used. The performance evaluation
showed that our model outperformed their models when evaluated against the attributes they
used and when extended to the use of eleven (11) quality attributes. The access control
model was also evaluated for accuracy, precision, and recall and was seen to have values of
0.93, 0.96, and 0.91, respectively. The study has established a model for assessing the
quality of software factoring in the major attributes of software quality assurance.