Speaker
Description
Scientific units nowadays commonly develop their own software in order to solve problems within their research. The planned user base is often limited to the originating unit which sometimes results in limited usability for external people. Oftentimes a software is only being developed to a prototype standard, so that it is usable for the group’s research, but not any further. When these stages need to be overcome, ad-hoc self-organizing management approaches can reach their limitations.
With this poster we present a user centric approach to the development of a software product not only to be used by our own group but industry, state organizations and other research institutes. We show how the core team, consisting of three software developers, one scientific advisor and tester, one IT project manager and the scientific project leader, organizes and communicates. A core tool for this is a variant of the agile project management framework Kanban implemented as GitLab issue boards, which serves as the main organizational tool for this project. We also explain how involving potential users through user meetings and early access versions of the software can be key to project success und feedback loops that can be used to ensure software quality.
ValidITy (Validation of Intelligent Terrain Feature Recognition Methods for Hydrographic Data [https://validity-project.eu/]) is a GEOMAR project to develop an object annotation and terrain classification software for gridded bathymetric data using machine learning and classification dictionaries.
