Current list of action items from the third e-Science Sisters Link Up meeting

Workflow planning for Shim services

1. Forward types of Shim services used in Graves’ disease workflows to Duncan. Action by Peter Li - done.
2. Contribute to a detailed survey of Shim operations in bioinformatics being done by Duncan. Decide on a suitable classification scheme for Shims taking into consideration of attributes such as experimental-neutrality. Action led by Duncan Hull working with Chris Wroe, Phil Lord and Newcastle?
3. Investigate the usefulness of Kepler data transformations/conversion actors in myGrid and implement them as local Java widgets. A list of Kepler data transformations is initially required. Action led by Shawn Bowers and Peter Li working with Tom Oinn and Manchester?
4. Chris Wroe to send information on requirements, service descriptions and ontologies in myGrid to Jim Blythe for comment with regards to Shims and workflow planning. Action led by Chris Wroe.
5. Arrange a follow-up Shim workflow planning meeting on Access Grid between Manchester, Southampton, San Diego and ISI. Action led by Martin Szomszor and Duncan Hull working with Shawn Bowers and Jim Blythe.
6. Organise visit to Manchester in 2005 by Jim Blythe to talk about AI planning in workflows and combine with CAT-S work. Action led by Manchester with Jim Blythe.
7. Target a paper on using AI planning with Shims. Action led by Duncan Hull and Jim Blythe?

Registry and semantic annotations for workflow components

1. Descriptions of the current myGrid registry/annotation/search architecture and the planned GRIMOIRES registries are required. GRIMORES team are currently writing an overview. Action led by Pinar Alper and Phil Lord.
2. Kepler/Triana/Hunter teams to annotate and discover their services and workflows with Pedro and Feta. Kepler have dropped this action. Hunter team to take a look at using Pedro and Feta.

Service and data semantics

1. Future myGrid work may need this more detailed conceptualisation, for example in Shim selection and deep provenance. Duncan Hull and Martin Szomszor will work with Shawn on the description of structured records and relating them to syntactic structure for shim management. Action led by Manchester and Martin Szomszor working with Shawn.

Data intensive workflows and services, and third party data transfer approaches

1. Build data streaming services using the Styx protocol and test third party data transfer approaches in workflows with Taverna and Kepler. Also look into problems with data streaming in workflows. Action led by Tom Oinn and Peter Li, working with Kepler team.

Compute-intensive and long running services and workflows

1. Justin Ferris will begin work on modifications to the myGrid workflow system towards the end of 2004, and will decide what concepts are useful from Triana, Pegasus, Kepler, for the initial use cases we identify. Action led by myGrid with Triana, Pegasus and Kepler teams.
2. Finalize timetable for work on compute-intensive and long running workflows, interaction with SRB, etc for Integrative Biology project in April. Action led by Justin Ferris and Tom Oinn.
3. Organise one week visit to the UK by Ilkay Altintas to work with Justin Ferris and Tom Oinn on compute-intensive and long running services and workflows. Coincide Ilkay’s visit with HealthGrid^? and Integrative Biology project meetings. Action led by myGrid team with Ilkay.

Workflow annotation and provenance models

1. A future deliverable should be educational material on these two areas for scientific workflow endeavours. Action led by Chris Wroe.

Workflow interoperability between workflow engines

1. Need to decide on concrete actions for workflow interoperability section. Ilkay suggested the following ideas:

• The execution of Kepler processes in Taverna can be extended by implementing further functionality to a Scufl processor so it that takes an actor, and puts it in a workflow that has multiple components for input and output ports of that actor.
• The execution of Soaplab services in Kepler and Triana can be made more generic by making the URL for the Soaplab service customizable and mapping the inputs and outputs of actors/tool to a Soaplab client generically.
• The workflow languages for each system can be examined if they can be converted to each other (at least at a basic pipeline level). This can then use the already accomplished component execution to execute each other's workflows.

Action led by Justin Ferris, Ilkay Altintas and Ian Wang.

Microarray data analysis breakout group

1. Need to perform a survey on the types of microarray data analyses performed by biologists. Action led by Peter Li and Cheryl Hornbaker.
2. Write up statement of intent for a Web Services framework for distributed microarray data analysis. Action led by Peter Li.

myGrid information model and mediator service

1. Keep other workflow projects updated with status of the myGrid information model and mediator service work. In particular, advise Yolanda Gil of the issues relating to why the myGrid information model had to be recently changed and updated. Action led by Nedim Alpdemir.

Validation of workflows using CAT-S

1. Send documentation of CAT-S validation service APIs to workflow editor developers. Action led by Jihie Kim.
2. Provide a bioinformatics workflow as a use case for validation by CAT-S.

Portals

1. Involve Efrat Jaeger in meeting with Gridsphere people and Stefan Egglestone (Nottingham) in 2005. Action led by Manchester.

-- PeterLi - 15 Dec 2004