CITI has stopped operations in 2014, to co-launch NOVA LINCS THIS SITE IS NOT BEING UPDATED SINCE 2013
citi banner
  Home  \  Publications  \  InProceedings Page Login  
banner bottom
File Top
Streamlining Code Smells:Using Collective Intelligence and Visualization

Code smells have long been catalogued with corresponding mitigating solutions called refactoring operations. However, despite the successful initiatives of integrating many of the latter in current IDEs (e.g., Eclipse), code smells detection has not gained the same status. Researchers have pointed out that the code smells detection process is inherently subjective and this fact is probably the main hindrance on providing automatic support. This research work focuses on the code smells detection process. To this end, it aims at proposing two contributions that, when used in a combined fashion, are expected to mitigate the aforementioned problem: crowdsmellling and smelly maps. We envisage that such features will be available in a future generation of interactive development environments (aka IDE 2.0) to help software developers to identify a set of code smells in Java source code. Crowdsmelling uses the concept of collective intelligence through which users around the world are encouraged to collaborate to a knowledge base that runs on a cloud server. This process will improve the assessment of the detection accuracy using a calibration process that occurs on the server after receiving data and suggestions from each user. Through the process of code smells detection based and referenced to a knowledge base (one for each code smell), we expect to some extent mitigate the subjectivity problem. Smelly maps builds upon the previous experience of setting up a software visualization infrastructure. It involves a strategy to represent detected code smells at different abstraction levels with the goal to increase software quality awareness. We expect to integrate both proposals with the aforementioned visualization infrastructure developed as an Eclipse plugin to facilitate refactoring decisions upon large software systems. We have laid out several evidence-based validation experiments, which will hopefully demonstrate that a step forward in the code smelling process will be achieved.


@ Proceedings of the 9th International Conference on the Quality of Information and Communications Technology (QUATIC’2014)

Publisher: IEEE Computer Society ( United States )

Pages: 306 to 311

Date: 2014


    Carlos Fábio Ramos Conceição (Instituto Universitário de Lisboa), Glauco de Figueiredo Carneiro (Universidade Salvador), Fernando Brito e Abreu
File Bottom