Feature Level Complexity and Coupling Analysis in 4GL Systems

Abstract

Product metrics are widely used in the maintenance and evolution phase of software development to advise the development team about software quality. Although most of these metrics are defined for mainstream languages, several of them were adapted to fourth generation languages (4GL) as well. Usual concepts like size, complexity and coupling need to be re-interpreted and adapted to program elements defined by these languages. In this paper we take a further step in this process to address product line development in 4GL. Adopting product line architecture is a necessary step to handle challenges of a growing number of similar product variants. The product line adoption process itself is a tedious task where features of the product variants play crucial role. Features represent a higher level of abstraction that are cross-cutting to program elements of 4GL applications. We propose a set of metrics related to features by linking existing program elements to metrics and by relating features with each other. The focus of this study is on complexity and coupling metrics. We provide a metrics based analysis of several variants of a large scale industrial product line written in the Magic XPA 4GL language.

Utilization of Underlying Semantic Information in Textual Data

“…LSI is already known to be capable of producing good quality results combined with structural information [5]. Our works also introduced more than forty new metrics [64] in total for 4GL feature extraction, which included properties based on complexity, coupling, size, and similarity. The advances listed above are elaborated in the following sections.…”