Meeting Minutes 11th December 2006

Meeting Minutes 27th November 2006
Present: Poul Nielsen, Alan Garny, Jonna Terkildsen, Shane Blackett, Peter Hunter, David (Andre) Nickerson, Andrew Miller

There was significant discussion of the meeting on Wednesday. It was decided that there would be two distinct meetings:
Alan Garny would give a talk, which is promoted to the wider institute, on COR, from 1 pm until about 2pm. There will be no demonstration of PCEnv at that meeting, instead this will be deferred until a Friday forum at some point in the future.

The meeting would then continue for those with an interest in the details of CellML, and discuss various items which need to be decided while Andre is here.

Peter was keen for there to be four main items:
=> User requirements for PCEnv, and other tools.
=> Workflow
=> CellML API
=> Model re-use.

Andre was keen for the CellML 1.1 support in the repository to be specifically included in this.

Shane thought that Michael Dunstan might be working on this, but no one was sure. The group agreed that we need to clarify exactly what he is working on.

Shane was also concerned that we were excluding the wider physiome project goals from the discussion. Peter thought that we should stay focused on CellML,because there is currently insufficient time to also discuss broader physiome issues, and other specifications like FieldML.

Andre also thinks that the workflow item being discussed on Wednesday should include the model ontologies.

Andre noted that the Kyoto group had developed a UI which uses ontological information to connect variables when components were connected. He was keen to know where the ontology work that Mike Lovell-Smith did was available, as the database of information is probably larger than what the Kyoto group have developed. Shane thinks that Matt has the database.

Andre also noted that Andrew McCulloch wanted a 'lightweight' code generation service. It is unclear how the current CCGS can be any more lightweight, as it makes use of the CellML API. He currently generates object files suitable for the target platform at the server side, and distributes them to the client. The group agreed that we shouldn't provide binary code, and so it would be better for him to be able to set this up using the CCGS. This requires that we provide some more documentation to allow people to create their own customised code generation services. Andrew Miller noted that we could provide access to the currently internal API, and so allow people to access a series of 'procedural steps' to be computed, rather than the code itself.

Andrew asked if we could release the branch versions of PCEnv + CellML API and so on. Shane noted he had an issue with the version of autoconf. Andrew noted that the packages come with a pre-generated configure, so this doesn't really matter, except to people developing on the API. The group agreed that Andrew could release the packages if Shane's issue appears not to be serious enough to hold back a release.