maintainable software practices for e-science - introduction to workshop
DESCRIPTION
TRANSCRIPT
First Workshop on Maintainable So3ware Prac7ces in e-‐Science
Neil Chue Hong So3ware Sustainability Ins7tute
Jennifer Schopf IEEE Computer Society
hDp://www.so3ware.ac.uk/maintainable-‐so3ware-‐prac7ce-‐workshop
Stopping errors before they maDer • Errors unchecked in code cause retrac7ons
• Chang – Protein Structure – Error in data analysis code flipped two
columns of inver7ng electron density map – Science 22 December 2006: Vol. 314 no.
5807 pp. 1856-‐1857 DOI: 10.1126/science.314.5807.1856
• Bertoia – Hypertension – Error in code led to doubling of sample size
and significantly different es7mates – Hypertension. 2012;60:e29, published
online before print August 6 2012, doi:10.1161/HYP.0b013e318269bc7a
Ensuring so3ware is maintained
• OU Knowledge Network – All funding stopped a3er 8 years, with 11,000 ac7ve users – s7ll available
• Atmospheric Model – Upgrades to hardware meant so3ware did not run – took effort to address
So3ware Preserva7on Benefits Framework: hDp://www.so3ware.ac.uk/aDach/So3warePreserva7onBenefitsFramework.pdf
A typical e-‐Science project organisa7on?
Steering Group
InvesJgators
Researchers
Students
Project Managers
Developers
10 partners
Topics of Interest • So3ware engineering and so3ware product management best
prac7ce as applied to e-‐Science and computa7onal science • Community development, collabora7ve development, and
widening adop7on • Licensing, funding, and business models for eScience and research
so3ware • Managing governance and organisa7onal change during the
so3ware lifecycle • Measuring and analysing the impact of so3ware and so3ware
processes • So3ware aDribu7on, cita7on, and credit • Interac7on between researchers, developers and stakeholders • Transferable so3ware prac7ces from industry
Agenda – morning Communi7es and Prac7ces
• 09:00 Introduc7on to the Workshop • 09:30 Sustainable So3ware for Computa7onal Chemistry
and Materials Modeling • 10:00 Break • 10:30 Adop7on of So3ware by a User Community: The
Montage Image Mosaic Engine Example • 11:00 On Realizing the Concept Study ScienceSo3 of the
European Middleware Ini7a7ve -‐ OpenSo3ware for Open Science
• 11:30 The Rela7onship Between Development Problems and Use of So3ware Engineering Prac7ces in Computa7onal Science
• 12:00 Lunch
Agenda – a3ernoon Models and Methods
• 13:30 libHPC: So3ware sustainability and reuse through metadata preserva7on
• 14:00 Overview of the TriBITS Lifecycle Model • 14:30 IPOL: Reviewed Publica7on and Public
Tes7ng of Research So3ware Debate introduc7on
• 15:00 Break • 15:30 Culture Change debate: "What can we do
to change the status quo? Should we do this?” • 17:00 Close
Area of Interest from Audience • Prac7ces and models • Portability across heterogeneous
plalorms • Ways of reducing complexity • Recording informa7on on
so3ware structure • Making so3ware verifiable • Building communi7es around
so3ware • Cultural issues in organisa7ons • Collabora7on between
ins7tu7ons and countries • Meta-‐issues of so3ware across
science and social science
• Mixed industry/academic projects • Integra7on of so3ware
components • Sustaining so3ware on liDle
money • So3ware engineering applied to
computa7onal science • Is there a life beyond public
funding (e.g. for dCache)? • So3ware + data preserva7on for
10 years usage (e.g. Tevatron) • Taking responsibility for orphaned
so3ware (e.g. for OSG) • Understanding what so3ware’s
been funded and what’s happened to it?
Culture Change Discussion
• Goals – Understand what are the key barriers to culture change with respect to so3ware in research
– Understand who and what needs to change, to change the status quo
– Record the studies, examples, prac7ces and models that the workshop par7cipants agree should be highlighted
Ques7ons
1 What are the basic so3ware skills and understanding every scien7st needs? – Beverly Sanders et al.
2 Life beyond public grant funding – what happens next? – Bruce Berriman, Michael McLennan, Jennifer Schopf,
Nigel Ward, Jim Willenbring 3 So3ware reward, cita7on, aDribu7on. Tracking
usage and impact – Neil Chue Hong, Alberto Di Meglio, Josh Greenberg,
Juan Lalinde, Kevin Jorissen
Process
• Brainstorm issues and ideas (45 mins, around break) – Put down ideas without judgement (it doesn’t maDer if
someone else doesn’t think it’ll work) – If you’ve heard of something, write it down
• Understand and group (15 mins) – Do you have the same understanding of all the ideas – Are some of the ideas linked
• Priori7se and record (30 mins) – What are the biggest issues? – What are the ideas that you’ve seen that (don’t) work best? – Are there trends/correla7ons that should be explored?
• Share your findings (15 mins) • Please record digitally so that it can be easily shared