Research On The User-Centered Process For System Analysis, Modeling And Design

A software process is a framework for the tasks that are required to build high-quality software. Traditionally, most software processes made their focuses on how to fit software systems into their functionality, that is, what the systems can do. However, this trend has suffered from a great challenge in recent years from the changed usage of computers. Computing is becoming ubiquitous with information technology, such as the Internet, information appliances, or many e-services, and reaching an increasing number of users. In many cases the users are neither computer professionals nor domain experts, and they use a software system mostly for their lives instead of for their work. This demonstrates that the computer is becoming an information tool and hence, it should be ease of use as the paper and pen. Therefore, usability is increasingly becoming a significant issue in software development.This thesis explores how a process improvement can be done with the consideration of usability. In order to promote User-Centered Design (UCD), we recognized first such an improvement should start from the consideration of what users could do since functionality should fit its usage;second the improvement should be still compatible with the existing procedures and techniques because developers were accustomed themselves to the traditional process that transferred from business domain to implementation domain in terms of two crucial models: conceptual and architectural models.To achieve these goals, we have identified the nature of the changed context of usage is the interaction between people and the social and technical environments bymeans of new findings from empiricists' researches rather than traditional rationalists' methods. Furthermore, the Context-Directed Interaction Modeling (CDIM) process presented in this thesis proposes such an improvement. The CDIM process is a complement of existing processes rather than a replacement for them. It is composed of three phases: a context analysis, a conceptual modeling and an architectural modeling phase, each of which is driven by a corresponding ontology. An innovative feature of the CDIM process is that the corresponding ontologies give existing models the meanings of user-orientation with the terms of interaction. This enables to transfer knowledge about the interaction into the existing models rather than to develop a particular set of models. In addition, as a true complement the CDIM process does not provide any procedures or techniques for the realization of functionality except some glue connecting usage to functionality.To demonstrate that the CDIM process can connect system implementation seamlessly, we also present here an Object-Oriented Interactive Graphics Toolkit (OOIGT) and an approach to the formal specification of dialogue based on attributed grammars. They illustrate the rationale behind the CDIM process indirectly, that is, although the toolkit and the approach are useful for the specification of system components, we still need to specify the content of those components. The information about the content comes from the context of usage rather than from the developers' understanding to the system.