Meeting Minutes 24 October 2007

Present: Randall Britten, James Lawson, Catherine Lloyd, Justin Marsh, Andrew Miller, Tommy Yu

Apologies: Poul Nielsen


Review of last week's action items:

2007-10-17-1: (James to reschedule the meeting with Matt regarding what metadata can be used as a replacement for reactions).

  • James noted that Matt requested information that he needs before he can fill out the metadata, and to sort out those information before having the meeting.
    • Action Item #1: James will get status update from Matt when he's done with that.

2007-10-17-2: (Catherine to get feedback regarding updates to the 'CellML and Related Efforts Section' from the mailing list).

  • Catherine made changes and clean up the related efforts page.
    • She emailed cellml-discussion list about her intention to update the related efforts page but there were no further response, possibly people listed on the related efforts page are not on the discussion list.

    • James suggested personal emails to the people

    • Action Item #2: Catherine will write emails addressed to those people for updates.


Agenda

Update on the progress towards CellML 1.1.1 and CellML 1.2

  • Reaction Element document

    • Poul got Matt to look at it and pending his response.

    • Poul also is waiting on Sarala's diagrams.

      • Action Item #3: Randall to follow up on the status and location of Sarala's diagrams.

Update on the Physiome Model Repository support for CellML 1.1

  • Randall and Tommy had a discussion on this Tuesday.

    • Points raised by that discussion can be found at http://www.cellml.org/Members/tommy/moving-pmr-to-scm-backend

    • Andrew noted the lack of discussion on permissions, as the focus was on Subversion which does not handle permissions in the way the PMR may require natively.

    • Randall did suggest using multiple Subversion repositories to deal with multiple projects, but Andrew noted that this goes against the original design goals of the repository, and glue code to move components between repositories may very well duplicate what Subversion already does.

    • Git was named as a better choice, and Tommy may go ahead to try out using Git (instead of Subversion) as a test backend for the next PMR.

      • Git will allow private repositories stored on each users' machine, and model authors only need to push their changes to the central repository only when s/he thinks its ready.

      • Action Item #4: Tommy will look into using Git as new backend.

    • Naming convention was discussed. It will be best to allow users to name models within a loose set of guidelines rather than a rigid one like the current PMR.

      • Naming convention within the current PMR is highly confusing to use, as noted by Catherine.

The status of the PCEnv 0.3 release

  • James noted that PCEnv 0.2 came out prior to the CellML API re-factoring which added support for model features such as using derivatives multiple times as you can with variables. However, as he is currently curating models against PCEnv 0.2 (the latest release), he expressed an interest in seeing PCEnv 0.3 released soon.
  • PCEnv has supported changes to graph traces from Javascript in sessions (for example, handling events on SVG diagrams) for some time, but no one has created a session which actually uses this yet, which would be nice to have prior to 0.3.
  • Built-in XML Editor

    • James noted that the XML editor in PCEnv can lose data, because if you type malformed XML, and then switch away from the editor view, PCEnv will try to parse the model, fail, and then you will be left with the version prior to changes being made.

  • Andrew has now partially implemented model validation support in the CellML API. So far, the code only checks for extraneous elements and attributes (or missing mandatory elements / attributes), as well as a few other representational issues. It doesn't yet check the MathML, or perform higher level checks on the CellML model.
  • Andre reported a bug on code generation, which was fixed.

  • Randall suggested creating an issue on the tracker for PCEnv 0.3, and other issues such as requested features or bugs can be marked to block that release to allow better notification. Andrew suggested using flags.

    • A breakaway meeting was held to discuss other pending issues blocking the release of PCEnv 0.3.