Design and Methodology for Cross-Platform Mobile Applications Project

This project focuses on the design and methodology of developing cross-platform mobile applications to reduce development time and costs. The design involves a source application (Android), a target application (iOS), and a server, where the source application's UI is converted and user operations are processed. The system uses Aspect Oriented Programming (AOP) to capture UI instances, XML structures for UI mapping, and an emulator to run the source application. The target application utilizes Swift, with WebSocket4Net for data transmission between the source and target. The methodology analyzes latency of IOS phone interaction, GUI updates, and network communication. This approach aims to minimize the need for separate cross-platform development platforms by running the source application in the cloud and displaying it to the user, with UI actions sent to the cloud for processing.

Design and Methodology

Summary
The basic design methodology used in the platform is based on the problem of making the
mobile applications for different platform individually. This leads to wastage of time as well
as money. The project focuses on developing mobile applications cross-platform or multi-
platform so that the features of the application can be enhanced. Along with this, it will
decrease the improvement and support costs to a sustainable amount.
Design
The design part is done in a way that it contains a source which can be an android app, a
target which can be an ios app and a main server. The system basically converts the opening
screens of the source and then it distinguishes the user operations and accordingly it sends the
data to the respective server [1]. The whole process can be concluded in following steps:
1. The system sends actions.
2. The user then runs the app in the cloud.
3. Accordingly it receives UI updates.
The designing of the system is done in such a way that first of all the source application will
seek the instance of the user interface. This instance is catch with the help of Aspect Oriented
Programming i.e. AOP. An assumption is made that all of the mobile platforms holds true.
After this runtime processing takes place, the system winds up the source application as XML
structures the view of the application [2]. The user interface is mapped with the XML for
android layouts like RelativeLayout and LinearLayout, and components like TextView,
ListView, ScrollView and MapView. The souce setup is basically an emulator on which the
source application runs. The events processed at the target are implemented by a listener. At
the same time, the emulator receives these listeners and capturs the events performed. A tool
known as MonkeyRunner is used which typically provides the API [3]. Apart from this,
python scripting is used which converts the positions of the ios mobile screen which is same
as that of the android emulator.
At the target end, the system easily converts the language according to the target application
with respect to the intermediate XML form which is generated at the source. Here, the target
application uses the swift programming language. WebSocket4Net is used to transmit the
data back and forth between the source and the target. Swift language uses objective-C which
contains fields like id, width, height and position. With the help of Swift programming, the
events of the object’s listeners which are captured are transmitted to the source emulator
using the web sockets [4]. To obtain data from the sensors like GPS are used to process the
sensor data of different files. At last, the system runs the application from the target by
communicating with the emulator. Web socket is used to communicate from the target to the

source because it provides efficient asynchronous communication as it lessens high network
latencies of single user int5eraction and is also based on the events.
Method
The working of the application system is done on parameters like Latency of IOS phone
interaction, Latency of GUI updates and the Latency of Network Communication.
With the help of this method, the need of using cross-platform development platforms is
minimized. The method depicts that the actual application which is the source is basically
running in the cloud using an emulator and then is displayed to the user who is the target. At
the background, the UI actions which the user mobile performs are sent to the cloud over the
internet. These actions are received by the emulator which is currently running in the cloud
and hence performs the functions of the app.

References
[1] E. Holder, E. Shah, M. Davoodi, and E, Tilevich, “Cloud Twin: Native Execution of
Android Applications on the Windows Phone,” in ICASE 2013, pp. 598-603.
[2] S. Baride, K. Dutta, “A cloud based software testing paradigm for mobile applications,”
in ACM SIGSOFT Software Engineering Notes 36.3 (2011), pp. 1
[3] Google.MonkeyRunner Android Development Tools.
http://developer.android.com/tools/help/monkeyrunner concepts.html.
[4] Apache. WebSocket4Net. http://websocket4net.codeplex.com/.