Presence and Performance of Mobile Development Approaches

Abstract

At the centre of the multi-trillion dollar mobile app economy, we find the mobile apps - smartphone-optimised software downloadable from platform marketplaces, including Google's Play Store and Apple's App Store. While the development of these apps has traditionally been conducted on a perplatform basis, using the native platform-specific programming languages and architectures, in this thesis, I explore alternative development approaches for cross-platform mobile apps. Original contributions to knowledge include new empirical insight into the presence and performance of these alternative approaches. The background is an identified lack of empirical studies on these topics, which has spawned claims and allegations across practitioners' outlets and scholarly research. This thesis follows the design science research methodology, focusing on the development and evaluation of mobile apps for performance testing, and scripts to facilitate studying the presence of technical frameworks in published apps. In terms of performance, the findings indicate that while the native development approach provides an overall better performance output, cross-platform frameworks can deliver better performance output in certain situations and for specific hardware metrics. Adding to the complexity is the challenge of significant performance variations between the assessed frameworks' generated Android and iOS apps. As for framework presence in published Android apps, the findings indicate that cross-platform frameworks are present in approximately 15% of the sampled dataset (n = 661 705 apps) with adoption fluctuating between app categories and that the choice of technology and framework has a significant impact on compiled app file size.