Article Preview
TopIntroduction
Evaluators of software applications and websites strive to make sure that their software is up to the quality standards relative to others. They used metric tools and methods to be able to get the software characteristics and compare them with known standards. In order to make the approach plausible, those attributes should be gathered automatically through tools. Web applications have some characteristics that make their maintenance expensive. This includes: heterogeneity, speed of evolution, and dynamic code generation (Ghosheh, Black, & Qaddour, 2008).
The characteristics of every software or website can be classified into a wide range of types or categories. For example, there are several characteristics that are related to performance such as websites processing speed, and the speed of executing transactions. There are also several attributes related to reliability such as number of errors in pages, in scripts, the percentage of time that the website is running or available, etc. In some cases, some of those characteristics may not be fully measured unless the website is operational. This may include operational quality results from four characteristics: effectiveness, productivity, safety, and satisfaction. Such attributes can only be measured during the operating environments of the software.
Software or website attributes can be also classified into two major types: internal and external attributes. Internal attributes are those characteristics that can directly be measured while external attributes are goals that cannot be measured directly. Internal attributes are indicators to the external ones. For example, the number of links in a web page is the internal size metric is one that can be directly measured. Such internal metric has relations with several external metrics such as website size, complexity, quality, etc.
An ISO standard (ISO/IEC-9126-1, 2001) defines 6 major quality characteristics. Those are: functionality, reliability, usability, efficiency, maintainability, and portability. This quality model describes each quality characteristic of a software product by further sub-characteristics that elaborate each characteristic. The Goal Question Metric (GQM) approach provides a three-step framework for a quality model: (1) list the major goals of the empirical study; (2) derive from each goal the questions that must be answered to determine if the goals have been met; (3) decide what must be measured in order to be able to answer the questions adequately (i.e. the definition of the metric) (Basili & Weiss, 1984).
Websites’ evaluation can be distinguished from typical software evaluation. In typical software applications, for example, traffic and usage metrics may not be as much as important as they are in a website evaluation. Website users’ behavior can be studied through eye tracking, log files, studying users’ sessions, etc. However, one of the challenges in this evaluation is that there are some user related attributes that are hard to automatically collect through metrics such as loyalty, satisfaction, understandability, etc.
Usability, ease of use, user-friendliness, and ease of learning is a set of related metrics to indicate the ability of the user to communicate easily with the website and understand its functionalities with minimum effort. It can be measured through different ways such as the time it takes a user to perform a task or different tasks using the website (usually in comparison with other websites).
Productivity metrics try to evaluate the amount of resources consumed through using the website. Examples of those metrics include: time, memory, and some other attributes that relate it to the level of effectiveness (i.e. number of tasks achieved in a certain time) gained by the user. Indicators of productivity can include: time needed to complete a task, the parts of the goals reached within a unit of time, cost-effectiveness of tasks, and the amount of time it takes the user to perform certain actions.
The focus of this paper is on studying the web structural that can be gathered automatically and their relation with traffic metrics. The research hypothesis is to study the amount of correlation (if there is) between structural metrics such as links, documents, forms, etc with traffic metrics (mainly back or inlinks). The original motivation to the research is that initial investigation of top websites in terms of popularity shows that they vary widely in terms of their structural metrics and nature. One possible explanation is that the domain of the website may impact its size nature. Accordingly, in this study, several websites from several business domains are selected to study if there are any consistencies in the structure and popularity metrics based on the business domain.