Leveraging Context in Open Software Development Ecosystems

点击翻译按钮获取中文摘要

This proposal addresses the new ways of developing software when time to market and fast response to customers is key to survival. Software development has gone beyond a single office or company and into open, collaborative, and global development transitioning to software ecosystems', as: “networked communities of organizations or actors, which base their relations to each other on a common interest in the development and use of a central software technology”. Ecosystems support new ways of working based on an unprecedented participatory, collaborative approach to defining product features and software development. The province of British Columbia for example, has recently started a large initiative (www.bcdevexchange.org) in which public and private sector organizations collaborate in an ecosystem where public data is shared in the development of products that serve the province more effectively. Studies consistently show that only 37% of software projects are successful [11]. Poor understanding of market demands, project planning and performance are top reasons for failure. Today's organizations have the unique opportunity to stay competitive and avoid project failure by leveraging a wealth of information (context) from within their ecosystem. For example, projects plan better by analyzing feedback on the criticality of product features from stakeholders external to a product (e.g. customers or partnering organizations); project performance significantly increases when developers can track changes to code, including who made the changes and why, of their own or dependent projects in the ecosystem. Despite this significant impact of contextual information, software organizations do not currently have the techniques and tools to leverage it effectively. This research will develop (1) theories of the role and utility of context in software ecosystems that impact project performance and planning, and (2) techniques to automatically summarize this context for product managers and developers. In collaboration with industrial partners, we iteratively study developers at work, develop our theories and techniques and evaluate them in practice. Our research will significantly impact organizations and their new ways of working in open software ecosystems. Managers will be able to automatically analyze information from the ecosystem and use it effectively in software development planning and execution to ensure appropriate product direction and quality. Further, our tools will improve developer productivity by summarizing context from various sources, reducing the amount of information sources developers must scour. Particularly relevant to the Canadian industry, our tools will be available to our partners (e.g. IBM), but also Victoria-based startups and the BCDevExchange partners. For all these organizations, success depends on the ecosystem engagement and feedback. 0 0 1 422 2486 University of Victoria 46 9 2899 14.0 Normal 0 false false false EN-US JA X-NONE