OMIIUK ISSGC09

1
OMII-UK _at_ ISSGC09

• 13 July 2009, ISSGC09, Sophia Antipolis, Nice,
  France
• Steve Brewer / Neil Chue Hong

2
OMII-UK Software Solutions for e-Research

• Expertise
• Open Source
• E-Research
• Software
• Services

OMII-UK is an open-source organisation set up by
the EPSRC to provide software and services to
help the UK research community adopt e-Research
practices and technology.
The role of OMII-UK is evolving into one of
cultivation rather than purely development and
custodian the aim is to sustain community
software through various channels of support
Currently funded by EPSRC, JISC and others
3
Mission sustainability

• OMII-Uks mission is to cultivate and sustain
  community software important to research

Facilitating mutual benefit between all
participants
4
Software sustainability

• Cultivation utopian goal for software
  sustainability

AHE Application hosting environment
SPAM-GP - Shibboleth-based portal presentation
manager
WSRFLite Perl-based API for SOAP
RAPID Portlet generation tool (JSDL)
CPOSS Crystal Energy Landscape Application
5
Services for eResearch software
Design /Code Evaluation
Innovators
Testing / Dev Infrastructure
Governance
Integration
Documentation and Training
Packaging/porting
Software Deployment
Promotion Exploitation
Providers
Deployment Analysis
Software Support
Software Improvement
Community Development
Users
Requirements Gathering
Information Provision
Software Contributions
6
OMII-UK Cycle for Software Sustainability
Domain expertise
Best practice Process
Software products Community services
Requirements gathering
Standards Information
Market research
Software expertise
Specialist Development
7
Engage Initiative Science through software

• How do scientists use computers?
• What do scientists want from computers?
• Where can scientists go for support?

8
JISC-funded ENGAGE Initiative achievements

• talked to 50 researchers
• Detailed picture emerged of the UK e-Research
  community
• Computationally intensive research is being
  conducted across all domains papyrology, climate
  modelling, choreography
• review of best practices and lessons learned is
  underway
• Interviews transcribed and analysed
• Transciptions written-up as public articles for
  further dissemination
• a collection of focussed projects launched that
  take forward the most promising technologies
  identified in interviews
• Projects developed where a significant and
  visible contribution to the community could be
  achieved in a relatively short time-frame.

9
Engage Initiative Science through software

• How do scientists use computers?

50 interviews across many domains
Triage process to guide project funding
Transcripts reviewed and analysed
Best practices extracted
10
Engage Initiative Science through software

• What do scientists want from computers?

11
Engaging Research with e-Infrastructure
Interviews
Wider deployment
Projects
Dissemination
Adoption
New requirements
12
ENGAGE Projects (phase 2)

• 7 projects funded across many domains
• CPOSS Crystal Energy Landscape Application
  upgrade and NGS port
• MCTP Monte Carlo Treatment Planning portal
  improvements
• RMCS Remote My Condor Submit improved support
• Aladdin2 new interface for GENIE Climate
  Modelling application
• eLab - Integrating field work with the e-Lab
  Notebook with centralized services and archives
• eSAD - Integration of image processing tools for
  epigraphy and papyrology (eSAD) within the
  Documents and Manuscripts (SDM) VRE
• OSCAR Chemistry-specific textmining application
  developed at Cambridge over last 5 years

13
ENGAGE Projects (phase 2) outcome
7 projects results (1-4) Monte Carlo
radiotherapy treatment planning (Cardiff) Now
complete new users at Swansea, Galway and
Liverpool making use of the updated
system Crystal Energy Landscape application-CPOSS
(UCL) New DMACRYS system now working
re-engineered workflows being evaluated by Sally
Prices research team at UCL RMCS Remote job
submission for molecular simulation
(Cambridge) Project complete and good progress
achieved Integration of image processing tools
within the VRE-SDM (Oxford) New integrated system
previewed at recent Image, Text, Interpretation
workshop in Oxford user i/f well received by
papyrologists and epigraphers
14

• 7 projects underway (5-7)
• Aladdin 2 a launchpad for the GENIE Earth-System
  Model (UEA/OU)
• Ported GENIE simulator now operational
  configurable parameters can be rendered MatLab
  logic has been ported from GENIELab
• Integrating field work with the e-Lab Notebook
  with centralized services and archives (Bangor,
  Southampton)
• Lab Blog book project linked to users at Bangor
  Southampton who are engaged with projects based
  on the analysis of molecular structure and
  function both via experimental crystallography
  and molecular dynamics simulation updated system
  now on NGS
• Refactoring of OSCAR3, migration to NaCTeM,
  modularisation and code hardening (Cambridge,
  Southampton, NaCTeM)
• Workshop in Cambridge (July 09 EPO, RSC, NaCTeM
  and others attended
• UIMA demo underway refactoring process guided by
  modular architecture plan

15
Monte Carlo treatment Planning
Web portal
NGS resources
Web services
User requests

• Remote access via portal
• Web services requests
• Experiment manager for simulation jobs
• Computational resources where jobs are done
• RTGrid database provides persistent storage of
  info

RTGrid Database
Condor pool
Experiment Manager

• Jobs took 159.45 hours on single machine
• 18.1 hours on Condor cluster
• 7.2 hours on NGS

16
CPOSS Crystal Energy Landscape Application

• MOLPAKDMAREL Centre for Theoretical and
  Computational Chemistry at UCL and the department
  of computing at UCL
• five interacting workflows along with six Web
  Services
• BPEL Web Services shaded grey
• replace dmarel with dmacrys
• Replication of the execution stages on Legion
  and the NGS as well as to still allow execution
  on Condor

17
Aladdin2 a cross-platform GENIE launchpad for
Research and Teaching

• GENIE project was established to deliver
  simplified and faster-running models of the
  Earths climate system
• GENIE users are thus able to model longer
  timeframes stretching over many thousands of
  years
• modular construction of this system has enabled
  Open University students to acquire a stand-alone
  simulator that allows them to create real-time
  models of various earth projections with a small
  number of variables to control

• configure and execute GENIE with real-time
  visualisation of key variables
• XML extension for tutorial examples
• GENIElab s/w enables multiple runs on remote
  resources eg. NGS
• ease of use reduces barriers to uptake
• access for students, teachers and policy makers

18
Integration of image processing tools within the
VRE-SDM (Oxford)

• Project management characteristics
• bi-weekly and impromptu team meetings
• Online collaboration application
• close interaction between researcher and
  developers
• issues easily and swiftly resolved
• understanding of the algorithms,
• correct operation and of the NGS,
• workflow of the image processing tasks,
• design of the data model
• GUI design of the portlet.

• Design choice criteria
• development restricted to the two projects
• VRE-SDM user interfaces exist for researchers
• NGS uniform execution environment
• researcher developing the algorithms in the team
• rapid turn around from algorithm to installation
• compiled binary accessed with GridSAM WS i/f

19
Integrating field work with the e-Lab Notebook

• Partners
• People
• Processes
• Services
• Interactions
• UNSW is the University of New South Wales
• STFC is the Science and Technology Facilities
  Council

20
OSCAR3 text mining for chemistry

• Project objectives
• Refactored OSCAR software
• OSCAR-UIMA integration
• Improved OSCAR software developer-user
  documentation
• OSCAR unit test infrastructure
• Improved SourceForge community support
• Successfully completed test process report

OSCAR3 architecture at outset
21
User engagement techniques and practices

• How do we manage the engagement process in
  practice?
• Should this be handled at the project level or
  the domain level?
• In practice the following points are relevant
• - user engagement is different for each
  project- similar techniques exist
  workshops/training, web, forums, video -
  important to identify the user community you wish
  to have at the end of your project and work
  towards that as goal(sustainability)
• Dissemination sources/subjects can be found by
  targeting either
• - passive users in order to "promote"
  themselves
• - creators/champions at top to bring in more
  overall users

22
OMII-UK sustainable community software

• You know, when you collaborate with people,
  there's no rules about anything.
• Therman Statom