ABI CellML Meeting Minutes, 29th September 2010

Present: Andrew Miller, Randall Britten, Dougal Cowan, Justin Marsh, Tommy Yu

Apologies: David "Andre" Nickerson, Peter Hunter, Mike Cooling

This week's agenda:

1) OpenCell development

Progress:

  • Justin: the release process for 0.8 is underway, emails about the RC were sent on Monday.
  • Justin: put together a spec file for building on 64 bit Linux.

Planning:

  • Justin: options for dealing with Justin leaving include truncating the release process (release on Friday), or sending release emails from a non-university address on Monday.
  • Randall: the release process will be done by Randall and Andrew after Justin leaves.

2) CellML API - update

Progress:

  • Andrew: added a feature to MaLaES which allows an extension element embedded in the MathML in the model to be passed through unmodified into the generated code. This is useful for several things - code which pre-transforms models part of the way towards code generation, and for prototyping changes. Justin noted that it would not be good practice to put a model containing such pass-through elements into the model repository, as the pass-through is specific to a particular target language. Also added a feature to MaLaES to strip passthrough code, and made CCGS use this by default. This ensures that unmodified applications which use CCGS do not become insecure as a result of this change, by not allowing untrusted users to embed executable code in a CellML model. In addition, Andrew has been working on a feature to split the conditions out of piecewise equations into the model, and turn them into roots which change sign as the condition changes. These can be used to restart the model at discontinuities due piecewise conditions changing. He is also working on making CIS support integration of such models.

Discussion:

  • Pass-through usage was discussed - this should not be used in repository models (unless a model is published using pass-through code). It is useful for prototyping.

3) CellML core specs - update

Progress:

  • Andrew noted that has has been working on several things related to ModML over the last week. He has created a CSV auto-solver - a program which takes a ModML model in Haskell, and solves it to get results. He has also nearly finished the Reactions module, which is a module that forms part of ModML models, and can transform descriptions of reactions into descriptions of maths with units, which is ultimately transformed into ModML core. He has also been working on monads for constructing reaction modules and parts of them (such as process descriptions).

Discussion:

  • A breakaway was called to discuss ModML.

4) PMR2 development - update

Progress:

  • Tommy: continuing to refactor code - Mercurial now depends on the workspace, rather than the other way around. This allows for workspaces to specify other VCSs. Exposures and workspaces separated.

Discussion:

  • Tommy: cellml.org is running out of space due to hosting the OpenCell files. A space upgrade is required, or the links changed to Sourceforge downloads.
  • Randall: more space should be obtained.
  • There was a discussion about moving the OpenCell files to another content distribution network/system.
  • Randall: maybe only the latest and previous versions of OpenCell should be hosted on cellml.org.

Action item 1: Tommy to change the cellml.org OpenCell links to latest/previous, plus redirects for older versions.

Action item 2: Justin to upload the previous versions of OpenCell onto sourceforge so Tommy can then link to them.