As the authors state,

existing approaches for building mobile device applications are often targeted to a particular platform (e.g. mobile phones, PDAs, sensors) and software technology (Web Services, Microsoft COM, Java components).

This paper investigates middleware solutions that require different programming models.

Ueyama et al. describe “a generic component approach for the construction of adaptive applications that can integrate and re-use technologies (e.g. middleware and legacy components) and deploy them across heterogeneous devices.” The OpenMob architecture, a component-based development framework, consists of a runtime kernel, extension components, and target applications. The kernel is designed to be lightweight, to allow deployment on devices where resources are particularly scarce.

The authors present OpenMob, “a general-purpose middleware[,] through the abstraction of loaders and binders.” Due to this abstraction notion, OpenMob developers are able to deploy components that are coded in a variety of programming languages, for a range of middleware. The approach aims to provide a generic tool for the construction of a wide range of mobile environments, for a variety of applications and hardware platforms. Loader and binder extensions encapsulate the complexity of the underlying target deployment environment at which they are aimed, which simplifies the development of mobile applications.

The paper consists of seven sections. Section 1 is an introduction. Section 2 discusses basic concepts and research challenges. Section 3 gives a technical description of the OpenMob approach and states its benefits. Section 4 describes OpenMob implementation prototypes using Java 2 Micro Edition (J2ME) virtual machines, and discusses their deployment to heterogeneous hardware platforms and their dynamic reconfiguration capabilities. The performance results show that the prototypes “have an acceptable performance and resource consumption overhead.” Sections 5, “Related Work,” 6, “Conclusions and Further Work,” and 7, “References,” finish the paper.

The paper’s academic approach does not sufficiently relate to existing technologies. Mainly, it lacks empirical results on the tradeoff between the generality of the loader and binder extensions and the specificity of commercial environments optimized for specific mobile device platforms and software technologies.