The majority of my notes were about individual plans for the time remaining on myGrid. I've included some discussion relating to the MIR that was had in the small group that I was in.

Manchester university

Pinar Penpecioglu is being employed as an RA on myGrid from October 2004 until June 2005. She’s going to be working on Feta for most of this time.

Phil Lord

• has produced a document entitled “Feta – the next steps”
• has been writing Perl scripts to help process Hannah’s data produced by her workflows
• is intending to generate a new version of Manchester’s ontology in OBO (Open Biological Ontology) format
• is intending to write a simple public registry with an HTML interface
• is intending to integrate Feta into Taverna

Chris Wroe is employed on myGrid until Easter. He will be carrying on working on provenance, including looking at how to present provenance information to bioinformaticians. He describes his current provenance information as shallow provenance – eg “this output was produced from this input”, but is interested in deep provenance – eg “sequences matching the sequence in this input file were produced by running this blast program”.

Steve Pettifer is going to be working on building visualizatins of Hannah’s data (in Taverna?).

Nedim thinks that that Information Model is fairly stable, and is implemented by the MIR. He is planning work on his eScience Mediator idea.

Sheffield

Both myGrid people at Sheffield have funding to last them to June 2005 (Neil Davis and who else?). They are planning to

• add their text-mining services to the existing Williams-Beuren Syndrome and Grave’s Disease workflows
• work on AQL (Ambit Query Language)
• do some work on mining provenance
• build web interfaces for some of their workflows

Southampton – IT innovations

Justin Ferris is currently working half-time on myGrid. IT Innovations also have a year’s funding from another project for enactor research/development.

Justin has developed code in Freefluo to allow a workflow to be run in a separate process, which could solve some of the enactor’s memory problems. He is planning to work on

• long running workflows – checkpointing for failure recovery
• work arounds for soap web-services which return large amounts of data (apparently there are some bugs in Axis’s SOAP implementation that are causing problems)

If any web-services become available that require SSL security, he will implement it in the enactor, and he confirmed that the basic framework for username/password authentication already exists in Taverna/Freefluo.

Southampton University

Southampton are taking on another researcher within a month (subject to work permit problems) to work on the notification service.

Victor Tan has received replies for the questionnaire that he sent out about security. He is planning to write a document describing the security problems involved in depoying myGrid components. He suggests that basic security for the MIR could be provided by http authentication over ssl.

EBI

Martin Senger is planning to remove SOAPlab’s dependencies on CORBA, to make it more reliable and efficient. He is also planning to write an (HTML-based? Swing-based?) management interface for SOAPlab.

Tom Oinn is continuing development of Taverna. He is looking at a project from the University of Reading which uses Inferno Grid to move large amounts of data around in a workflow environment. He is planning, along with Justin Ferris, to refactor code in Taverna and Freefluo during the next month.

Newcastle

Newcastle plan to write a definitive Grave’s disease paper, and to develop a web-service interface to ENSEMBL.

The GOLD project is planning to enhance the MIR by adding

• role-based security
• an OGSA-DAI interface
• facilities to watch for changes in data items

They are planning an MIR that uses Oracle directly (rather than Hibernate over Oracle) to take advantage of some of Oracle’s data mining capabilities.

Peter Li is working on a more robust implementation of the array expression web-service. He is also interested in mining MIRs for intersting biological data.

Keith (Flanagan?) is working on making the Grave’s disease workflows more biologically useful.

Nottingham

Kevin Glover is starting to work on a new graphical workflow editor for Taverna. The editor is likely to improve visualisations of workflows by allowing the expansion and contraction of sub-graphs.

Stefan Rennick Egglestone was planning to investigate the use of SRB for input and output of data through both Taverna and a web-portal. Carole expressed concern about the amount of time this would take.

MIR

After splitting into groups, the group I was in (me, Carole, Tom, can't remember who else) came up with the following ideas about usage of the MIR (I've paraphrased this from what we said):

"The MIR cannot come close in performance to a database system which is custom-built for a particular problem. There's nothing wrong with this. What the MIR should be good for is as a system roughly equivalent to your lab book. It might allow you, for example, to search it for all workflows you've run in the last few days. If these workflows generate large amounts of data, it should probably be stored in a specialist storage service, and a reference stored in the MIR. If we think of the MIR this way, then security problems become less of an issue, because you rely on the security features of your specialist storage system, rather than the MIR providing security."