Code-based comparison of software applications' mobile and desktop versions: An exploratory study

Characteristics of a given software platform usually have profound effects on the design and source code of applications developed in that platform. There are fewer studies evaluating or comparing the features of desktop versions against mobile versions of an application while there are many studies in the literature comparing mobile platforms such as Android, iOS and BlackBerry, and applications developed for these platforms from different angles. With this thesis, applications developed for mobile and desktop platforms have been examined in terms of design criteria, source code size and platform dependency and it has been researched whether a similarity can be established between these platforms. In this direction, four research questions were constructed by adopting the method of exploratory research: (1) How is the design of mobile and desktop applications, in terms of object oriented metrics? (2) How is the size of mobile and desktop applications in terms of source code size? (3) How are applications developed for mobile and desktop platforms different in terms of dependencies? (4) Can similarities be identified between the applications developed for these two platforms? To answer these questions, the source code for the mobile and desktop versions of two open source applications has been examined and analyzed in a comparative way. It was found that the results obtained for the DIT and WMC metrics in the C & K metric set, which is used in the design evaluation, are higher in the desktop applications, and the values for the RFC and CBO metrics are higher in the mobile applications. Within the scope of the second research question, it was observed that the code size of mobile versions was larger than desktop versions for both application pairs. With the third research question, it is concluded that the mobile applications are more dependent on the platform. When the results of the last research question and the other research questions are evaluated, it is concluded that the obtained data and the conditions of the research are not suitable for making inferences by discussing the similarity of the platforms with feature-equivalent mobile and desktop application couples which have been developed with different languages.