본문 바로가기
  • Home

A Study on an Object-orientation and Extensibility in Context-aware Systems

  • Journal of Knowledge Information Technology and Systems
  • Abbr : JKITS
  • 2017, 12(2), pp.353-373
  • DOI : 10.34163/jkits.2017.12.2.015
  • Publisher : Korea Knowledge Information Technology Society
  • Research Area : Interdisciplinary Studies > Interdisciplinary Research
  • Published : April 30, 2017

Choi, Jong Myung 1 Iksu Kim 2

1목포대학교
2숭실대학교

Accredited

ABSTRACT

There has been a lot of research on context-aware computing, but software extensibility in the development of those systems has not gotten much attention in spite of its importance in software engineering. In this paper, we introduce some extension requirements for context-aware systems, and identify four extension types for them: sensors, context inference algorithms, contexts, and context-aware services. For those extension requirements, we propose four extension mechanisms based on object-oriented technology: separation between abstraction and implementation of context, separation of context from sensors, modular and separate model for context-aware functions, and overloading model for context-aware functions. By adopting those mechanisms, developers or maintainers can add new sensors, context inference algorithms, contexts, or context-aware services without modifying the source code after deployment. Those mechanisms are all based on object-orientation. Our approach represents a context as a class, and context services as methods with context parameter. Context inference is a class with a method which understands sensor values and infers the current context from the values. This approach will reduce costs, time, and efforts in context-aware system maintenance which requires new context-aware features after deployment because it will increase the software reusability and extensibility. In this paper, we also specify a case study which shows how to extend a context-aware system with the extension requirements.

Citation status

* References for papers published after 2023 are currently being built.