ABI CellML Meeting Minutes, 14th July 2010

Present: Dougal Cowan, Randall Britten, Andrew Miller, Tommy Yu, Justin Marsh, David "Andre" Nickerson

Apologies: Peter Hunter, Mike Cooling

Last week's action items:

1) Dougal - ask Catherine whether it is worth bulk-converting the e-notation if the XML formatting is changed significantly.

  • Done by Justin as part of action item 2.

2) Justin - test the script on an example model, send to Catherine to show formatting of output.

  • The script has been tested, and does not change the formatting of models tested so far. It will change the formatting of certain types of whitespace. See tracker item 1703.

3) Justin: make a tracker item about Catherine's email, re: units checking.

This week's agenda:

1) (Catherine) How hard would it be to write a script to search for the following: xml:base="" xml:base="file.

Discussion:

  • Justin: It would be very easy to automatically convert. This could be done at the same time as the e-notation changes.

Planning:

  • Randall: Perform e-notation and xml:base conversions separately. An Hg commit for each change would be useful.

2) (Catherine) Does anyone know the current status of the containment grouping in the specification?  I've been leaving it out recently as I thought it was going to be deprecated and the (spatial) information it contains replaced with a FieldML description or something else...  Am I wrong?

Discussion:

  • Andrew: Containment is deprecated, as it represents metadata that should not be in the model. This should be gone in CellML 1.2.
  • Randall: Is there any metadata that Catherine could use in place of containment?
  • Andrew: SBML metadata may be usable.
  • Randall: Mike may be working on this now that he has begun work on CellML metadata.

Planning:

  • Action item 1: Randall - ask Mike Cooling about containment in metadata.

3) (Catherine) Discussion of a model describing a population of neurons.

Discussion:

  • Randall: There has been discussion of implementing this kind of agent-based modelling in OpenCMISS. Gib Bogle has code which may make it into OpenCMISS, that is the closest we have to this sort of modelling.
  • Andrew: It depends on whether the values selected from normal distributions are parameters or part of the model. Parameters should be part of the simulation metadata.
  • The paper was discussed (Butera et al. 1999).
  • Randall: This might be able to be done using insilicoIDE, which can import CellML.
  • Andrew: There are ways this could be done with ModML.

4) PMR2 Application note

Progress:

  • Tommy: Still waiting on feedback from some authors.
  • Andre: There have been some comments from Alan Garny which should be included.

Discussion:

  • There was a discussion of format standards for papers, TeX, PDF, Word etc.
  • A private repository for author collaboration, such as a Mercurial repository or even using PMR2 was discussed.

5) CellML core specs - update

  • No update.

6) Metadata specifications (general, citations, graphing, simulation, other) - update

Progress:

  • Randall: Mike Cooling has begun work on the metadata spec.

7) CellML API - update

Progress:

  • Andrew: Has been through the tracker and fixed a number of smaller issues.
  • Andrew: Has been working on 'custom' code generation, where the user can choose to generate code to compute a specified set of unknowns from a specified set of known variables. This is something that the OpenCMISS developers and several other people have asked for (tracker item 1952).
  • Randall: Still awaiting a response from Geoff Nunns about the Saucerman/IDA problem.

Discussion:

  • The Saucerman model was discussed.
  • Andre: The API and model version used by Geoff should still be on his machine. It should be possible to run the model there.
  • Andrew: Has seen some correspondence with Peter Hunter as he has talked to a group which would like to simulate many CellML models in parallel using CUDA. Andrew has replied to note that the CellML Code Generation Service might be useful for this purpose.
  • Randall: Andrew should get in touch with Jin who is using CUDA for his Master's project.

Planning:

  • Action item 2: Justin - find out what version of the API and model Geoff was using by looking at his machine/VM.

8) OpenCell development

Progress:

  • Randall: The document recommending Scala/Java has been sent out. No response as yet.

9) PMR2 development - update

Progress:

  • Tommy: Finalising the way to store configuration settings and set curation flags.
  • Randall: No reply yet about CellML Simulator style exposures.
  • Andrew: Probably no further major changes to make to the paper. Still waiting on response from some authors.
  • Dougal: Some PMR2 documentation is now going up on the site, as private documents until they are complete. The first of these was shown.

10) Repository contributions - update

Progress:

  • Catherine has been translating new models and curating existing models under the Neurobiology section.
  • Hanne has been creating SVG sessions for the Cell Cycle, Immunology and Neurobiology sections.