Empirical research on the impact and evolution of Bad Smells

不良气味是指对软件的可理解性和可变性产生负面影响的代码和设计异常，从而阻碍了软件的发展。这些糟糕的编程和设计实践可能会在某种程度上导致软件处于一种不再经济地进行更改的状态，从而缩短软件的生命周期。如果系统要长期存在，就必须检测、跟踪和处理坏气味——换句话说，积极管理。很多关于软件质量的研究都是针对不良气味的。甚至还有软件质量认证，如德国TÜVit提供的软件质量指数或可信赖的产品可维护性，都是基于不良气味的发生。这些问题的范围从重复的代码、高耦合、低内聚到过大的大小和许多其他代码方面。然而，臭味的真正影响还没有得到充分的调查，也没有得到实证证明。因此，不完全清楚，是否，如何，以及什么样的坏气味会对维护产生负面影响。除此之外，人们对难闻气味是如何演变的知之甚少。大多数研究通常只关注被研究软件系统的单个快照，而忽略了进化数据。我们提议的项目旨在全面捕获软件中的不良气味，并着重关注它们的演变。我们在收集到的关于坏气味的数据中寻找重复出现的模式和关系。我们调查进化数据是否有助于更好地定义、检测、理解和评估坏气味，并评估它们是否以及如何适合于评估内部软件质量。我们在研究界传播我们的数据，作为进一步补充和比较研究的基础。我们的长期研究目标是基于经验的可维护性模型，以支持长寿系统的进化。

Bad Smells are code and design anomalies that have a negative effect on the comprehensibility and changeability of software and thus hinder its evolution. These bad programming and design practices may at some point lead to a state of the software where it is no longer economical to make a change and thus shorten the life time of software. If systems are to be long living, Bad Smells must be detected, traced, and handled - in other words, actively managed. A lot of research on software quality is devoted to Bad Smells. There are even software quality certifications such as the Software Quality Index or Trusted Product Maintainability offered by the German TÜVit that are based on occurrences of Bad Smells. These range from duplicated code, high coupling, low cohesion to excessive size and many other code aspects. Nevertheless, the true impact of Bad Smells is not yet sufficiently investigated nor is it empirically proven. Consequently, it is not fully clear, whether, how, and what kind of Bad Smells impact maintenance negatively. In addition to that, little is known how Bad Smells evolve. Most studies typically look only at a single snapshot of the software system under study and ignore evolutionary data.Our proposed project aims at comprehensively capturing Bad Smells in software, with a strong focus on their evolution. We search for re-ocurring patterns and relations in the collected data on Bad Smells. We investigate whether evolutionary data help to better define, detect, understand, and assess Bad Smells and evaluate whether and how they are suitable to assess the internal software quality. We disseminate our data among the research community as a basis for further complementary and comparative studies. Our long-term research aims at empirically-founded maintainability models to support the evolution of long-living systems.