Agenda

1. Security
2. Planning the final 4 months
3. Follow-up projects
4. Commercialization
5. Publications plan
6. Post-project support for components
7. Link-up visitor programme
8. ISMB and other demos
9. myGrid workshop and user group meeting

Present

Matthew Addis, Stef Egglestone, Rob Gaizauskas, Carole Goble, Luc Moreau, Tom Oinn, Nick Sharman, Paul Watson

Apologies

Peter Rice

1. Security

Since most bioinformatics services do not require authentication of users, we felt we only needed to discuss how to provide security for core myGrid services such as the MIR.

We discussed the use of existing security architectures which could be used to quickly provide security for services, in particular the architecture provided by the OMII. However, it was noted that the current OMII security release only works on SUSE Linux, with releases planned for Redhat Linux and then Windows at a later date. This is a problem, as biologists tended to use machines running Windows or MacOS^?.

Without the ability to use an existing security architecture, and with the final major release of myGrid occuring at the end of March, we agreed that there would only be time to implement a basic level of security for services. This would involve passing a user name and password with a service invocation over HTTP, which would then be compared against any Person entities in the MIR. Whilst trying to decide whether HTTP authenticatin should be implemented, we noted that there is some distinction between the following two different groups of potential myGrid users:

• follow on projects to myGrid who would not be interested in the basic level of security that we could provide in the limited time left on the project, but for whom a coherent security architecture design would be useful
• users such as May Tassabehji (Hannah Tipney’s supervisor at Manchester), who would not use myGrid services that are not protected by a basic level of security

Luc argued that it would be better not to implement any security at all, and that we should focus on developing ideas to be implemented by future projects. As part of this process, Luc and Victor are already working on getting a secure version of GRIMOIRES which uses the OMII security infrastructure.

It was eventually decided that the HTTP authentication idea would be implemented, but that it would not be presented as producing a service which was particularly secure.

2. Planning the final four months

We made some small alterations to Nick’s plan for the next few months. Nick agreed to update the section of the plan relation to Kevin’s work on the interactive graph editor. We also agreed that some effort is required to improve documentation of myGrid to help follow-on projects. We noted that Neil is likely to be replaced on myGrid by Ian Roberts at Sheffield. Stef agreed to make available a list of JDBC queries that the portal is making on the MIR to help in the development of a query interface to the MIR. Given that a large number of services are invoked through specialist processors in Taverna rather than as web-services, Tom suggested that code to make a procesor available as a web-service would be useful. Tom also suggested that results rendering components from Taverna might be used via applets in the portal, and mentioned a new data viewer included in Taverna source.

3. Follow-up projects and 6. Post-project support for components

We dealt with these two agenda points together. We agreed that it would be good to avoid forking the myGrid code base as much as possible, and Carole suggested there should be a consortium of people to make decisions about changes to the code base. Ideally, this would be funded through a platform grant, the proposal for which includes both an architect and someone (Katy?) to support follow-on projects. If no money is received by this mechanism, Katy will be moved onto PsyGrid and will support some of the follow on projects from there. Tom indicated that some effort may be made available for Taverna from other internal projects at the EBI.

4. Commercialization

Matthew gave us more details of his contacts with <?>, who has approached myGrid with an interest in commercializing some of the project. We agreed that it seems unlikely that anything will happen in this instance, but that we should prepare for any future commercialization activities. We agreed that the following tasks should be performed:

• an audit of IPR and copyright issues in myGrid, which would establish who own what, the dependencies of myGrid on third-party software and the licensing conditions of these items of third-party software.
• the collection of statistics about how frequently myGrid software is used. Tom indicated this could be estimated using logs from the LSID authority installed on phoebus at Manchester. The number of users of Taverna could also be estimated from the number of people on the mailing list, although this probably underestimates the user base

It was suggested that a discussion group on commercialzation could be set up at the next EPSRC pilot projects meeting. People invited to take part in this group could include academics who have been involved in succesful commercializations, representatives of universities’ commercialization departments and potential venture capitalists.

There seemed to be a general agreement that no one currently working for myGrid would be interested in pushing forward with a commercialization themselves. Tom indicated that Apple have taken an interest in Taverna. Carole agreed to raise commercialization during a meeting between Manchester and other life scientists at Manchester.

5. Publications plan

We talked about publishing a book about myGrid. Tom pointed out that a book has recently been published about the EMBOSS toolset. We agreed that it was unlikely that one person could write the book themselves, and that it was more likely that a number of editors would prepare it using submissions from other members of the myGrid team.

We talked about the need for one or more definitive journal papers about myGrid. We mentioned a number of possible journals that could be targeted, including ACM Transactions on Information Systems, IEEE?/ACM? Transactions on Computational Biology and Bioinformatics and ACM Transactions on Internet Technology. We decided that work on journal papers should start soon in order for them to be published before a ? in 2007. We also considered the possibility of trying to publish in a high-readership publication such as Nature or Science.

We also talked about publishing Taverna’s user manual, first as a technical report and perhaps then as a book.

7. Link-up visitor programme

Carole indicated that there was still a substantial amount of money available (£18,000 - £20,000) in the link-up budget which she would like to spend. She agreed to ask for an extension to period in which this could be spent. We discussed a number of ideas about ways of spending the money, including

• sending myGrid people to spend a period of time working with one of the link-up programme partners (eg ~1 month)
• organizing link-up meetings after conferences, thereby funding people’s trips to conferences
• another large meeting with link-up partners

8. ISMB and other demos

There will definitely be one or more demos of myGrid at ISMB. There may also be demos at eScience All Hands, EPSRC pilots meeting and GGF in Boston.

9. myGrid workshop and user group meeting

Robert Stevens is organizing a myGrid workshop in Manchester, which will be combined with a myGrid/Taverna user-group meeting. We suggested the dates of May 16th-21st for this.