In this paper we present some useful guidelines for the development of web-based applications. We comment on the current trend of development, its drawbacks, and then propose some guidelines to improve the quality of such practice. These guidelines are being investigated and it turns out to be very convenient and helpful.

Server side computation is very important to the end user. It is a performance issue, which is, usually, not well realized. In this paper, we provide a calculus that helps in expressing these performance issues at the design stage. The calculus is a subset of the Pi calculus with enriched timing constructs. We provide a formal syntax, an operational semantics and a bisimulation method. To illustrate the calculus an example from the web domain is given.

No abstract

The current practice in capturing requirements of web-based applications is accomplished by an ad-hoc and informal manner. In this paper we investigate the use of formal methods to formalize the requirements of a web-based application. We have used a simple yet powerful Definitional Specification Language (DSL) for this task. DSL is based on first order predicate calculus, and has three main concepts: a resource, a transitional, and a state type. These concepts help in specifying dynamic behaviors and the resultant specification is clear with no cumbersome manipulation of time variables. A brief description of DSL is given before we provide the informal requirements of our web-based application. We formalize the requirements using DSL and describe how this is done. We conclude that the benefits gained with such formalization have made the development clear and minimized the cycle of fix-and-test.

Selecting the right components to design a software system involve some deep thoughts and difficult decisions to make. In this paper, we present a model to ease the decision making process. The model is based on Commercial Off-The-Shelf (COTS) paradigm. COTS techniques aim to reduce development time and hence decrease cost compared to a traditional system development. First, an identification of components from the application requirements is made. Then, for each component, we specify a formal model, which is called the ideal-component. A structured first order predicate calculus is used as a tool to formalize application requirements and obtain these formal models. The evaluation of a possiblecomponent, from a vendor, begins with understanding the features and then an acceptance indicator is calculated. The acceptance equation combines three key factors: requirements and features match, vendor-viability and maintainability. Maintainability is a costly phase in any software system and this model caters for this issue during the evaluation process. The model is being investigated in small-scale systems with successful results.