Vol.3 No.1 2010

76/110

Research paper : A field-scientific approach to Clinico-Informatics (Y. Kinoshita et al.)−73−Synthesiology - English edition Vol.3 No.1 (2010) took place in an extremely expert-oriented style where the technology was thoroughly taught to one engineer. This engineer participated in several research projects at CVS and went back to Q, after two years, with ample experience in conducting “restoration experiments” and “pre-development experiments.”As a result, the trained engineer has been active and has established a study group of model checking after the completion of the fieldwork project. The group extended even to other companies and he created his own model checking training course for engineers, which is readily used in Q.The two cases of fieldwork above differ in their strategy for propagation of the technology inside the industrial partners, even though both were trials to apply the same technology. The reason seems to lie in the diversity of software development. The field to which the information technology is applied tends to be diverse and that implies that qualitative research is effective here.7 Discussion and conclusion7.1 IssuesWe are still in the process of collecting experiences in the fieldwork for technology transfer. Therefore, the scenario presented in this paper should be regarded as tentative. There are, however, at least two issues for the future.a) There is room for further systematic study on observations obtained through our fieldwork. It seems qualitative approaches, such as the KJ method, are effective for such a study.b) While there are active discussions on measurements and quantity for the software development processes[15][16], in clinico-informatics, how to use such measurements in concrete technology transferNote 10) still remains as an issue to be discussed.Just as there are issues of finding the cause of a disease and clarifying the mechanism of a disease in clinical medicine, there are issues of finding the cause of risks and clarifying the mechanism of faults in clinico-informatics, and these issues must be solved by collaborating with the basic research of informatics. The risks (or dependability, if one looks at the other side of the same phenomenon) of information systems are widely studied nowadays, and an approach that takes into account the diversity and complexity of information processing, as presented in this paper, seems to be worth emphasizing more.7.2 ConclusionIn this paper, we attempted the systematization of clinico-informatics based on the field scientific methodology by Kawakita. The KJ method seems particularly effective as a method of information gathering in the requirement analysis, as it is one of a few methodologies about abduction. To apply the KJ method to the actuality of clinico-informatics is our future theme.AcknowledgementsThis study was done using the experiences of the technology transfer activities conducted by a succession of research groups: Collaborative Research Team of Informatics, Laboratory for Verification and Semantics and Research Centre for Verification and Semantics, which continues since the establishment of AIST. We are grateful to all enterprises involved, the engineers thereof, and the researchers of AIST who participated and contributed to these technology transfer activities. Professor Hideyuki Nakashima and Professor Naoto Kobayashi kindly reviewed the paper, and provided several essential, synthesiological comments. In many cases in the review process, the authors’ understanding was improved in an essential way. We are deeply grateful to be given this opportunity of writing this paper and of discussion.NoteNote 1) We write Kawakita’s family name before his given name, as is usual in Japanese. We guess he had a definite opinion on how his name should be put, from the fact that the KJ method was named as it is after his own name.Note 2) Model checking is a technique (or technology) in system and software engineering. It is an application of mathematical logic to software engineering, which is generally called Formal Methods. For an overview to Formal Methods, refer to Reference [17].Note 3) Reference [18] is a proposal for the terminology of the dependability and risk of information systems such as fault, error and failure. We try following them. The notion of fault removal is explicitly defined in [18], and it includes verification (finding the fault).Note 4) A transition system is a mathematical structure given by a pair (S, R) of a set S, whose member is called state, and a binary relation R, which is called the transition relation, on S. If s R t holds between states s and t, we say there is a transition from s to t. Automaton is obtained by adding some additional data such as input and output symbols and relations around them to a transition system. A state transition diagram often used by programmers may be seen to denote a transition system. In that sense, transition systems are widely used in practice as a mathematical model of information system.Note 5) In some cases, all stakeholders come to an

※このページを正しく表示するにはFlashPlayer9以上が必要です