write Summary and Critique

Summary of an article: Data Mining in the Detection of Higher Level Clones in Software
Name
Institution
Data Mining in the Detection of Higher Level Clones in Software
This article dwells on code clone which involves similar program structures that have a considerable size as well as significant similarity. According to Basit and Jarzabek (2009), knowing the clone’s location can assist one in comprehending the structure and for maintenance efforts. It is important to remove clones to evade their harmful impacts. Some of the methods that can be used include refactoring, replacing them with function macros and calls, or utilize unconventional metalevel approaches like Aspect-oriented programming and XVCL.
Basit and Jarzabek (2009) argue that the limitation of clone detection is a big problem that software developers experience since many researchers are engrossed by the duplicate code fragments or the simple clones instead of focusing on the bigger picture whereby the duplicated codes can be a part of a big replicated program. As such, they are known as larger granularity similarities clones. Most of the cloning is found within the system variants that originates from the common base code that is during the evolution process. This occurs due to the excessive copying as well as modification of the program files hence the clones either larger or small occur in the system variants. Basit and Jarzabek (2009) assert that the knowledge of structural cloning is prevalent during the creation, but the community lacks formal ways of making the presence of structural clone visible within software other than the utilization of naming conventions and the documentation process.

The knowledge regarding the difference structural clone is implicit as well and is easily lost when undertaking successive evolution and software development.
Basit and Jarzabek (2009) propose the use of follow-up analysis on the simple clones in the detection of clone structures. In this approach, one begins by searching for and identifying a simple clone that involves a similar code fragment. After that, the higher-level similarities are identified based on an incremental approach of data mining method for revealing the frequent closed item sets, as well as clustering.
Basit and Jarzabek (2009) implemented a structural clone detection method on the tool known as Clone Miner. Even if the Clone Miner can identify simple clones, the underlying structural detection method can work with output originating from simpler clone detectors. The article also presents the assessment of the Clone Miner tool based on its benefits by evaluating the structural clones that were detected within public and commercial domain software systems. Accordingly, this approach is scalable and beneficial since it results in better comprehension of the clone programs.
The article is systematic and extensive since the authors undertake a far-reaching literature review and research to examine the current and past clone detection method. Basit and Jarzabek (2009) found the problem with the past detection approach to structural clones and suggests a better approach which is scalable and more useful. The authors also present their future studies hence very informative to the readers. Moreover, the article is easy to understand and challenges the reader to engage in knowledge acquisition concerning clone types and detection approaches. Nonetheless, the presented approach of follow-up analysis can only apply to the physical location of the clones. The method is not very dependable in other aspects since the authors lack semantic association knowledge between different clones that can enable develop a better system. The work is very informative, but as the authors suggest, there is a need for further studies to detect different kinds of clone programs that are highly multifaceted.
References
Basit, H. A., & Jarzabek, S. (2009). A data mining approach for detecting higher-level clones in software. IEEE Transactions on Software engineering, 35(4), 497-514.