Transcript Software Lifetime

OMII-UK collaborations workshop
Exploiting and sustaining research code
30/04/2009
Web: www.omii.ac.uk
Email: [email protected]
Who
•
•
•
•
•
Pascal Ekin – University of Manchester
Mike Jackson – University of Edinburgh
Tobias Schiebeck – University of Manchester
Claire Sloggett – Intersect
Aaron Turner – University of York
Web: www.omii.ac.uk
Email: [email protected]
Issues
• Researchers don’t productise code
o
o
o
It’s not research – code is not citeable
Rapidly changes and may be discarded and redone
But even API doc and comments are neglected
• Researchers don’t publish code
o
o
“It’s just a proof-of-concept”
“No-one will use it”
• Research councils don’t fund productisation/hardening
• Lack of comments/doc
o
o
Productisation blocker – what does it do?
Reuse blocker – what does it do? how do I use it?
Web: www.omii.ac.uk
Email: [email protected]
Issues
• Lack of tools to enable researchers to reuse
components
o
o
o
Drag-and drop e-science components
Glue together
Focus on business logic
• Research code decay
o
o
Seems faster as code often isn’t stable in first place
Layerered on other unstable code
• e.g. e-science middlewares
Web: www.omii.ac.uk
Email: [email protected]
Combat decay
• Virtualisation
• Pros and cons
o
o
Access to the particular OS and tools that are needed
Reluctance to risk changing a black box that works
• Miss possible gains in efficiency or power
• Expose as services or provide another layer
• Pros and cons
o
o
o
o
o
Buffer from changes in middleware
Validate inputs and improve robustness
Who hosts?
Who maintains?
Another layer is another source of bugs or can decay
Web: www.omii.ac.uk
Email: [email protected]
Encourage improved coding
• Comments, APIs, what it does
• How?
o
o
o
Very short cheap courses provided by unis
Online materials
Cite benefits to researchers
• Pros and cons
o
o
o
o
o
Productisation by third parties
Maintainable
Reusable
Who pays?
Who ensures that researchers continue to apply what has been
learned?
Web: www.omii.ac.uk
Email: [email protected]
Software consultants
• Pool of trained software developers
• Available to all researchers in a uni
• Pros and cons
o
o
o
Expertise on tap
Can ingest, review, harden, support
Who funds?
• Uni
• Portion of a research grant
o
What if there aren’t enough projects?
• Cut down the team and lose expertise?
Web: www.omii.ac.uk
Email: [email protected]
Code as a recognised research artefact
• By researchers, unis, research councils
• Citeable code
o
o
A paper describing an experiment and results is valid
So why not code?
• Reproduce results
• Validate implementation
• Research councils insist that code is published in
a repository
• Repository evaluates code
• How to enable such a fundamental shift?!
Web: www.omii.ac.uk
Email: [email protected]