A Structural Complexity Metric for Web Application based on Similarity
Woo-Sung Jung, Eun-Joo Lee · Journal of the Korea Society of Computer and Information · 2010
Abstract Software complexity is used to evaluate a target system's maint ainability. The existing complexity metrics on web applications are count-based, so it is hard to incorporate the understandability of developers or maintainers. To make up for this shortcomings, entropy-theory can be applied to define complexity, however, it is assumed that information quantity of each paper is identical. In this paper, structural complexity of a w eb application is defined based on information theory and similarity. In detail, the proposed comp lexity is defined using entropy as the previous approach, but the information quantity of individu al pages is defined using similarity. That is, a page which are similar with many pages has smaller i nformation quantity than a page which are dissimilar to others. Furthermore, various similarity measures can be used for various views, which results in many-sided complexity measures. Finally , several complexity properties are applied to verify the proposed metric and case studies shows th e applicability of the metric.▸Keyword:웹 어플리케이션(web application), 복잡도(complexity), 유사도(similar ity)∙제1저자 : 정우성 교신저자 : 이은주∙투고일 : 2010. 05. 02, 심사일 : 2010. 05. 18, 게재확정일 : 2010. 05. 30.* 서울대학교 공과대핟 컴퓨터공학부 박사과정 ** 경북대학교 IT대학 컴퓨터학부 조교수※ 이 논문은 2007년도 정부(교육과학기술부)의 재원으로 한국연구재단의 지원을 받아 수행된 기초연구사업 연구임(NRF-2007-331-D00407)