Reporting back slides

Back to the Collaborations Workshop page.

Themes and thoughts from the final session

Final agenda

The final agenda lists the topics that were discussed during the workshop.

Themes

• Researchers and developers want to work together, but they’re not sure how to do it.
• Exchange of information between disciplines is important
• What is it that encourages people to replicate work and tools?

Thoughts

• Want more researchers next time
• Ring fence some cash – get researchers that have problems to come to the CW and pitch their idea in which they bid for some of the cash (could even do it like a Dragon’s den style of presentation)
• Role of the research councils needs to be improved – must get them to attend next time
• Don’t want to separate researchers and developers onto different days
• Could run short sessions on the first day (and more of them, I guess) where ideas get laid down, and longer sessions on the second day where stuff get sorted out
• Could we get the funding councils involved so that users get involved too, because they see the CW as a way to access funding council money

Report back from Wednesday 10 March, 12.00-13.00 session

Software sustainability: what are the issues?

• Download slides

Flipcharts	Transcript
Overview p1	Overview p1
Overview p2	Overview p2
Overview p3	Overview p3
Group 1 p1	Group 1 p1
Group 1 p2	Group 1 p2
Group 2	Group 2
Group 3	Group 3

Secure Data Sharing

• Download slides

This session was facilitated by Pawel Plaszczak, who has provided the following information:

1. Manifesto for secure data sharing: http://bigdatamatters.com/bigdatamatters/2009/09/manifesto-for-the-secure-data-sharing.html
2. A three-minute introduction movie: http://gridwisetech.com/adhoc#adhocmovie
3. A technical overview: http://gridwisetech.com/sites/default/files/file/AdHoc_1.1.0_Data_Sheet.pdf
4. An online demonstration: https://adhoc.gridwisetech.pl/gridsphere/gridsphere/loggedin/home/r/ (It is not currently available to the general public, but people can write to contact@gridwisetech.com to request access.)

Long-term researcher-driven collaborations

• Download slides
• Mind map

Report back from Wednesday 10 March, 15.45-16.45 session

Data copy docs

• Download slides

OSSW-SSMM & software sustainability

• The OSS Watch self-assessment sustainability form and the results summary
• Steve Lee's presentation
• Mind map

Assistance with publicity, outreach and dissemination

• Download slides

Report back from Thursday 11 March, 9.15-10.15 session

Use of infrastructure at different levels for researchers and developers

• Download slides

SAGA

• Download slides
• Download support materials

e-Labs

• Download slides
• Mind map 1 and Mind map 2

Sustainability

Flipcharts	Transcripts
Part 1	Part 1
Part 2	Part 2
Part 3	Part 3
Part 4	Part 4
Part 5	Part 5

Google summer of code

• Download slides

Report back from Thursday 11 March, 11.35-12.35 session

Commercialisation and IPR

• Download slides

Collaboration 2

• Download slides

Photos

Transcripts

Sustainability 1 overview p1

ISSUES – Main issues preventing continued development of s/w

• Code Sharing
• New developers
• Run code but not view **
• IPR issues
• A good repository
• Well built vs succession of quick and dirty
• Miss matched timeline expectations
• Cutting edge vs productions
• Clashing management styles
• Software dependencies

Sustainability 1 overview p2

ISSUES

• Group cohesion
• Politics
• Differing research agendas/priorities
• Development methodologies
• Not everyone buys into coding guidelines
• Broken promises, no cont
• Dying developers – bus factor – redundancy, team size for encapsulating *knowledge sharing
• Must be funded to work together to produce deliverables in the UK
• Deviations from consortium plan
• Running off with an idea
• Building community is not difficult
• Getting concrete contributions is
• Difficult to get people to use others code where appropriate
• People often wear multiple hats
• Getting good real sample data ** ** in design process from actual users
• Scalability
• Trusting other peoples code when integrating different components

Sustainability 1 Group 1 p1

ISSUES

• Trust – e.g. HPC
• Research software in production - QA – Support
• Pushing to users
• Why – collaboration *** INSTITUTION
• Extra effort to make reproduction
• Current practice is to publish not reproduce
• Can we trust results – sustainability
• Good tools aid capture, tracability
• Resources – code quality – long term support – cost/funding
• Open source – rubbish code is useful
• Secrets
• Monitization
• Licensing
• Patents
• Great Tools
• Collaboration

Sustainability 1 Group 1 p2

SHORT TERM ACTIONS

• More collaboration – start between cs and research
• Training in collab
• Review research process – recommendation
• Paper reviews look reprodoc

LONG TERM GOALS

• Change culture make reproducible results

Sustainability 1 Group 2

Verification of results from software?

• “Throw-away software” problem - training?
• Move towards open pub of s/w as *** requirements
• Research computation expen** -v- “properly designed” s/w – balance?
• Sustaining “
• Training PhD students to build better s/w tests

Sustainability 1 Group 3

TOP ISSUES

• Differing research agendas/priorities
• Bus factor, team change impact and size
• Mismatched timeline expectations
• Cutting edge vs productions
• Well built vs quick and *** series
• IPR Issues
• Short term actions
• More F2F meetings
• Employ more people who can improve team cohesion
• Long term goals – culture change
• Better set of adopted processes
• Change scoring systems to individual s/w deliverables

Sustainability 2 p1

BETTER RESEARCH SOFTWARE DEVELOPMENT TEAMS FOR SUSTAINABILITY

RISK FACTORS

• People Leaving
• End of funding
• Schisms
• Divergent groups
• Different goals/vision
• Code fidelity/trust in code
• Using obscure technology
• Everyone using different tools
• Divergence from agreed

Sustainability 2 p2

CAPTURING KNOWLEDGE

• Use backed up electronic records
• Mailing list & archives (searchable)
• Use of ticket/issue tracker
• How to explain benefits to small team
• Not enough to have a “single” process manager – all must buy into process
• Introduce competition?
• Show benefit/value
• Address human side

WHY DO SOME TEAMS MANAGE TO RETAIN MEMBERS

• Investigate and publish case studies

OPEN DEVELOPMENT ALLOWS SEPARATION OF FUNDING AND COMMUNITY CONSENSUS

Sustainability 2 p3

GETTING INSTITUTIONAL BUY-IN

• Over project success
• Over external community growth

USE OF COMMON INFRASTRUCTURE

• Are branches bad

COMMON GOALS AND VISION

• Agreed roadmap
• But this takes time – show “failure” + “benefits” – case studies
• Shared ownership
• Involve entire team
• Take people to a neutral isolated location

Sustainability 2 p4

REPRODUCING RESULTS FROM SOFTWARE

• Software metadata
• Component defn
• Annotate software
• Test driven dev, coverage, reports

PROMOTE TOOLS LIKE HUDSON

• Education in benefits of design
• Summer schools
• Best 1 day SE for research curriculum?
• Recommended bibliography
• Standard Models for work
• Framework “Scaffolding” should be shared

Sustainability 2 p4

DIFFERENCE BETWEEN DIFFERENT METHODS AND FORKED CODE

EDUCATION

• On line materials on methodology
• Case studies c.f TextGen
• Low overhead, low barrier techniques better
• OSS-Watch strategic projects
• Mentoring

PERSUADE FUNDERS TO IMPLEMENT

• Minimal documentation format
• What it does
• What it runs on
• How you use it
• Minimum information to allow code to be run by someone else