60 likes | 202 Vues
This report summarizes findings from a workshop addressing sustainability in eHumanities software and tools. With insights from experts Peter Wittenburg and Walter Lioen, the compilation by Mark Dupuis discusses various software types, including algorithmic modules, community software, and core services, each with distinct sustainability challenges. It highlights issues such as hidden maintenance costs and the reliance on amateur developers, suggesting a need for improved funding models and better documentation. The report advocates for a collaborative approach to bolster software usability and persistence in the field.
E N D
Workshop reports • Abstracts of the findings from the two parallel tracks, during eHumanities address software and tools, provided by the rapporteurs, • Peter Wittenburg (Policy and Organization) • Walter Lioen (Technology and services) • Compiled by Mark Dupuis (SURF) Software and Tools Sustainability – Technology and Services
Problems • Different types of software: • algorithmicalmodules • widely used community software (PRAAT, ELAN, etc.) • infrastructural software (repository software, meta data aggregation, portals) • core services (persistent identifiers) • Have different sustainability requirements: • algorithmically: very dynamic • community software: both stability of existing features and extensions • infrastructural software: well-maintained • core services: proven technology • For some types: • working in niche markets • special: researcher workflows built in software • no formal verification, lack of test suitessometimes testing by students Software and Tools Sustainability – Technology and Services
Problems • Funding model isn’t right • development costs, about 20%, are funded via proposals • maintenance costs are not made explicit and remain “hidden” in the proposals • this is not a sustainable model Software and Tools Sustainability – Technology and Services
A lot of software is developed by engaged “amateurs” (PhDs, etc.) • This leaves the software unsustainable although its needed for follow-up research • There always remains a need for the Darwin model • It isn’t necessarily wrong to re-invent the wheel Software and Tools Sustainability – Technology and Services
Are there in technology and or criteria ehumanities’ specifics? • use varies from theatre studies to neuro cognition modeling • more traditional fields: more tool usage • upcoming paradigms: more software development like Software and Tools Sustainability – Technology and Services
Conclusions / Actions • Some hundreds of tools • Abstraction of commonalities and based on this produce more robust software • Quality: skeptical about yet another institute but we all like to start with a light-weight quality seal (similar to DSA) • team stability • funding stream • responsiveness (bugs, support, feature requests) • documentation • No big committee • “Registry” (1 single/unified national eHumanities portal) • Trusted group • Simplicity for the users • Light-weight seal can be integrated • Have software maintenance costs incorporated in proposals • commissioned services model • Using valorization grants • Training, consultancy, fellowship programs (Neil) Software and Tools Sustainability – Technology and Services