文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > Validating Object-Oriented Design Metrics on a Commercial Java Application
侵权投诉

Validating Object-Oriented Design Metrics on a Commercial Java Application

de l'information

2 Background

Below we present a detailed cognitive theory of object-oriented metrics. The theory explains themechanisms as to why object-oriented metrics are associated with fault-proneness. By giving themechanisms, it predicts which metrics are likely to be associated with class fault-proneness, andthe direction and functional form of the association. In our study we then empirically test thesepredictions.

It is important to articulate this theory in detail so that it can be tested. If the theory can besupported empirically, then it is possible to devise better metrics (i.e., with greater predictivepower or that can be collected early in the life cycle). Furthermore, if we understood themechanism then it is plausible that we can improve the design and programming process to avoidthe introduction of faults altogether. It is also possible that hypotheses derived from a cognitivetheory cannot be empirically supported. In such a case cognitive complexity may not be thereason why certain structural properties of object-oriented applications are problematic, and wehave to focus our attention on finding the mechanism of fault introduction.

2.1 Overview of the Cognitive Theory

In order for object-oriented metrics to be associated with fault-proneness, there has to be amechanism causing this effect. In the software engineering literature, this mechanism is believedto be cognitive complexity. For instance, Munson and Khoshgoftaar (Munson and Khoshgoftaar,1992) state that “There is a clear intuitive basis for believing that complex programs have morefaults in them than simple programs”. It has been argued that a detailed cognitive model is anecessary basis for developing software product metrics (Bergantz and Hassell, 1991; Ehrlichand Soloway, 1984). A further argument has been made that unless there is an understanding ofthe cognitive demands that software places on developers, then only surface features of thesoftware will be measured, and such surface “complexity” features will not be effective (Sebrechtsand Black, 1982). This implies that a cognitive theory of object-oriented metrics that explains howand why certain metrics are associated with fault-proneness is necessary.

Figure 1 depicts the cognitive theory that we describe. The rationale for the theory is the commonbelief that the structural properties of a software component (such as its coupling and inheritancerelationships) have an impact on its cognitive complexity (Briand et al., 2000; Cant et al., 1994;Cant et al., 1995; Henderson-Sellers, 1996).

上一页第5页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email