Other services: Tangible infrastructure support services
- Discovery registration system. Includes:
- Software review and certification
- Use and quality metrics
- Ticket tracking
- Seed funding to join the ecosystem
- user defined registry
Support software lifecycle
- Function: Assist in full lifecycle software development: e.g. (and order notwithstanding) initiation, licensing, maintenance, hardening, extensible, release, decommissioning
- Call for Proposals - to solicit consulting help (or to nominate a project to receive help) (e.g. Developer needing help supporting software. Community rescue of software. Developer wanting to archive obsolete software.)
- Selection mechanisms for projects to assist - community-based steering group
- Dissemination of best practices and training materials
- ISEES: review successful proposals, identification of likely problems, selection of appropriate consulting expertise
- ISEES: factiliation of execution consulting service - on-site, reverse-site, web-based, …
Consultant “hotline” - quick answers to funnel questions to appropriate commmunity-based solutions
- Resources required:
- Network/Registry of consultants / developers / analysts / etc. The experts may be full-time staff, part-time staff and ad hoc groups of experts to help with a well defined problem.
- \($\) for consultants
- Website for submissions
- People to review submissions and identify resources
- Facilities for teleconferencing, online working space (e.g. wikis, repositories for data, software, cook books)
Physical meeting/collaboration spaces
Reproduciblity Editor service, a la Journal of Biostats
- Take minimal advocacy approach. Roadmap best practices.
- Advocate NSF adopt software management plan: not gonna fly. Outline best practices
- Build brand, swag
- managing sponsors for bootcamps?
The challenge with registries: Quickly become useless. Successful ones need automated updating. Does a few things things well.
- How would ISEES uniquely provide value to enable science
Intergrative science through software collaboration.
- What 1 thing is most critical to pursue?
via Hackathon / working group collaboration of researchers and expert developers. Building a sustainable software model emboding modularity and best practices.
Registry model as incentive problem.
Education and Code review Value is science. Pursue illustrating what can be achieved through open science.
Right Scale. Why 1 institute, not 6 regional institutes?
Efficient, Scalable, Reproducible, Impactful
- Ian Truslove
- Christine White
- Annie Bryant
- Patricia Cruse
- Chris Mattmann
- Bryan Heidorn
- Greg Ederer
- Karthik Ram
- Bill Michener
- Bruce Caron
- Matt Jones
Science via Integrated Software.
Working group / hackathon on Software Lifecycle challenges that are impeding grand challenge science.
2 - 4 meetings over 1-2 years. 8-10 people.
Improvements to software lifecycle. (documentation, hardening, maintance, interface, usability, efficency, scalability)
E,S,R,I (efficency, scalability, reproducibility, impact) science advances. New science that can’t be done now. Overall software quality advances.
Number, quality, use of software products. Publications. Citations. Co-author networks (proxy of community)
Workshop model. Address the other generic challenges (code review/registry, etc, see ‘other services’ at top)
What is the metadata for scientific software? The DataCite Schema for software? Avoid duplicating listing efforts of other repositories (CPAN, CRAN, etc), while providing academically relevant elements like citations (and potentially other use metrics, a la Github), code review scoring, link to permanent repository and bug tracking.