The Research Of GUI Design And Code Structure Improvement Techniques On Mobile Platform

With the rapid development of modern mobile device, such as mobile phone and PDA, and the increasing popularity of Graphic User Interface (GUI) on mobile platform, more and more researchers pay attention to how to implement efficient software applications on mobile platform, after fulfilling the basic functionality. GUI design is an important part in modern software development. Particularly, GUI design on mobile platform has the main objectives of providing a human-centric vision optimization, more recognizable icons and other visible elements, and ease-to-use and highly extensible user interaction interface for end-users.In order to develop efficient software applications, current researchers proposed various approaches to optimize hardware architecture, software design, and mobile operating system, and made significant progress. However, as pointed out by previous research, GUI part in mobile computing environment is an essential role to deal with most human-computer interactions and operation events. Unfortunately, there seems to be few research focused on the GUI design in a human-computer interaction perspective, to improve the operation efficiency. Moreover, with the expansion of GUI code size in mobile software development, traditional object-oriented programming methodology, though succeeds in greatly improving the software modularity and reducing the source code complexity, fails to solve the crosscutting concern problem, which might easily lead to code scattering and tangling in a large software system. Therefore, researches on how to optimize GUI operation efficiency and source code maintainability, in both end-users and mobile software developers'perspectives, are needed. In this paper, we present an approach to optimize both operation efficiency and code maintainability of GUI design on mobile platform. We first investigate and summarize the state-of-the-art GUI design techniques as well as their flaws. Then we introduced a highly adaptive dynamic menu component implementation. Compared to the traditional GUI components like menu and button, the adaptive dynamic menu component facilitates the basic UI operations and has the unnecessary menu operations significantly. To improve the source code maintainability, we present an approach to optimize the GUI design based on aspect-oriented refactoring. In our approach, we use a novel scenario pattern-based approach to abstract aspects from the original software system, to handle the crosscutting concern code, making the GUI design easier to maintain and reuse.Another contribution of this paper is the introduction of a system evaluation model, based on the identification of source-code-level semantic changes and dynamic coupling metrics. Our proposed evaluation model is based on multi-factor statistical analysis, which makes the evaluation process fully automated. It first utilizes a set of 23 standard benchmarks to train model automatically, then it uses a set dynamic coupling metrics proposed in this paper to quantitatively reflect the system maintainability changes. In our experiment with one standard aspect-oriented refactoring benchmark and the GIS system, we found this model practically accurate. We also believe this model can also be used for software understanding, bug predication, and software testing tasks, apart from system maintainability evaluation.