1905 An Approach to Human-Error-Prevention Learning from Failure Cases

Abstract

Most of incidents in IT systems result from mistakes in software. The humans develop the software and operate the systems, but they sometimes make mistakes. So we have to reduce the mistakes by those humans (=human errors) as soon as possible, when thinking about the seriousness of the social influence due to those incidents. So, we decided to use those incident reports in our company which describe the details to solve the problems, and complaints from the customers, to prevent recurrence and occurrence from those mistakes/human errors. We add the Action Mandala, which effectively helps to solve the problems to the Failure Knowledge Database, which is produced by JST, and name it "Human Error Management System (HEMS)," and we propose it as a new cause-action model. The new model can verify easily adaptability and effectiveness of the cause-action items. And we show effectiveness of HEMS by explaining examples of applying HEMS in our real software development and operation business.