Meeting Minutes 2008-02-07

Present: Randall Britten, Michael Cooling, Edmund Crampin, Catherine Lloyd, Peter Hunter, James Lawson, Justin Marsh, Andrew Miller, Poul Nielsen, Jonna Terkildsen, Anil Wipat, Tommy Yu

Note: Meeting not held on the day before (2008-02-06) was due to Waitangi Day (NZ public holiday).

Action items from previous meeting:

  • Justin and Randall will review the specification draft.

    • Randall and Justin have spent some time reviewing the specification draft.

    • However, Randall noted it would be good to involve more reviewers, especially reviewers who have had a longer background with CellML. Poul said he would join in the review.

    • Randall also reiterated Alan Garny's suggestion for a way of cross referencing between the CellML 1.1 spec, a version of the draft of 1.2.  There was some discussion about how and whether to do this.

    • Further discussion was deferred to breakaway session.


Agenda

FieldML URL in the CellML specification.

  • Mike had found a broken hyperlink to FieldML in the CellML 1.1 specification.

  • Andrew pointed the error out to Gareth, and wanted Gareth to create a page at the targeted URI. However, Gareth changed the URI of link on the specification instead of creating a redirection page.

    • This could lead to a problem where a person may have a static version of the specification, and it will not lead to the correct URI. Resolution is to have Gareth create the page at the former link also.

      • Action item: Randall to speak with Gareth about creating a redirect to the new FieldML page from the old URI.


Mike: Separation of initial conditions from the model in CellML

  • Separation of structure from data may be desired to enable overriding of initial value of a model, and changing parameters.

    • Andre does this already using CellML 1.1 imports; it works because he built the models with this in mind.

    • Mike suggests some sort of tags that would make this easier.

    • However the suggestion is to follow current best practices.

    • Poul: this would present a major change to the specification, and this is not a good enough reason to add a new tag (when current practices appear to be sufficient).


Mike: CellML specification "what's new" pages.

  • Mike wants a page that briefly states what has changed between different versions of the CellML specification.

    • There is not much between 1.0 and 1.1 as only imports are added.

    • As for the changes to be implemented between 1.1 and 1.2, it's still under discussion.

    • However such a page will be useful, and can be done later as CellML 1.2 specification becomes more solid.


Mike: Standard Unit Files online.

  • Poul created a CellML file that defines all the secondary derived units that are based on SI from the standard units that are part of the CellML specification.

    • It was pointed out that "block importing" of all these units will be desirable to save time typing all of them out.

    • This could be automated by software tools so no need to add the feature that allows block importing of units into CellML (plus it's a major change).


James: SBML to CellML converter

  • James wants to know what kind of priorities this has, how long it will take.

  • A suggestion was API mappings between data structures

  • However the implementation are in different languages (C++ vs Java).

  • Peter will speak with Mike Hucka regarding possible ways of jointly funding this; it similar to a previous grant.


James: CellML imports into CMISS

  • James proposes having the CellML repository to generate Fortran code will help this.

  • There are two tasks. Firstly, the repository integration with the code generation service (CCGS) needs to be upgraded to the CellML API version 1.3, which has significant changes. Secondly, a Fortran definition file (.mal file) needs to be written for the CCGS.


CellML Specification update

  • Discussed in breakaway session.


Peter: Automating the generation of Javascript hooks between SVG diagrams in PCEnv.

  • Peter wants to know if this can be automated, as components of the diagram must map to CellML components.

  • However, the components of the SVG diagram might not use the same identifier (name) as the cmeta:id of the components found in the CellML file, thus render automation of this process difficult or even impossible.

  • Sarala may have done something in this area that will still allow mapping of identical components identified with different identifiers.

  • Action item: James to organize meeting with Justin and Sarala (and others) to learn about this issue.