CellML Meeting Minutes, 24th of August 2011

Present: Andrew Miller, Tommy Yu, David "Andre" Nickerson, Caton Little, Randall Britten, Poul Nielsen

Apologies: Dougal Cowan

Action items from last week:

  • Last week's action item 1: Dougal to e-mail CellML API nominees - this has been done.
  • Last week's action item 2: Andrew has e-mailed Lucian about the problems he is having with the CellML API on a Redhat platform; Lucian e-mailed to say he has been away and so hasn't had a chance to reply yet.
  • Last week's action item 3: Minus vs - in TeLICeM: Andrew has now fixed this issue. It was noted that Andre replied on the tracker to say he wasn't happy with this outcome. There was a discussion following on from this around why Andre wasn't using it; Andre is trying to generate a tool to automatically convert from MATLAB into CellML, which is outside the original scope of what TeLICeMS was intended to do, although it was intended to make manually doing such conversions from in a tool using TeLICeMS easier. It was noted that it might make sense to use a general parsing framework or a MATLAB parsing framework, but it wouldn't be very easy / make much sense to have a reasonable parsing framework for converting to CellML. Andre considers TeLICeMS to be unreliable at even converting small fragments of MATLAB to MathML, but doesn't have time to identify specific problems. Randall suggested that as the Saucerman-McCulloch model has both MATLAB and CellML variants, it might provide an opportunity to test TeLICeMS parsing.

Andre noted that he was going to COMBINE; there was a brief discussion of what he should present on the current direction of CellML. Some discussion of uncertainty would be useful. Action item 1: Andrew will send some slides on the uncertainty work in CellML to Andre.

  • Item 1: Herb Sauro has asked why we don't follow libSBML's development model.
    • Discussion:
      • Lucian has also sent comments about this; his most recent comments seem to suggest that he thinks the API has improved since we started providing binaries and a CMake based build system. Some of his original concerns related to the API itself, and not the ease of use - we should ask him for comments on this.
  • Item 2: PMR2 Development.
    • Progress:
      • Tommy noted PMR2 0.4 has been publicly staged.
      • Tommy has done some work to address an issue causing Mercurial push to not work properly.
      • Tommy noted that work on the OAuth plugin is continuing.
      • Tommy has done some work on packaging.
  • Core Specifications:
  • CellML API:
    • Progress:
      • Andrew has fixed the cause of a number of code generation warnings, particularly in generated Java and Python bindings; one of these changes fixed a problem with out parameter support in the Java bindings.
      • Andrew also changed minus() notation to -() in TeLICeMS.
      • Andrew has started looking into setting up automated tests for Java language bindings.
    • Discussion:
      • The reason the CellML API documentation isn't updating properly on cellml-api.sourceforge.net still needs to be looked into.