ABI CellML Meeting Minutes, 11th August 2010

Present: Dougal Cowan, David "Andre" Nickerson, Tommy Yu, Mike Cooling, Justin Marsh, Andrew Miller

Apologies: Randall Britten, Peter Hunter

Last week's action items:

1) Tommy: Look into a script to find models that require fixing.

  • Not yet done.

1) Tommy: give push access to the Saucerman workspace to Justin.

  • Done. Justin has put the model up on PMR2. The model has been shown to work with IDA, but it still does not contain a stimulus. The output matches the MATLAB with stimulus switched off.

This week's agenda:

0) (Justin) OpenCell 0.8 release process - final release of current codebase

Discussion:

  • Justin: Can anyone think of anything which is still blocking the release? Examples might be COR text view or further IDA validation.
  • Andre: COR text is probably too much work to include in this release. Would like to see further IDA testing.
  • There was discussion of IDA validation.
  • Mike: Does releasing without COR text support delay COR text support being available. There was a discussion about development of COR text support, and its usefulness to students starting in November.
  • Andrew: Does the renaming of base_uri to baseURI need to be done before release?
  • There was a discussion about this renaming.

Progress:

  • Andrew: IDA support has been tested on some test models with analytic solutions, as well as a number of the existing repository models.

1) (Catherine) Licence terms - how do we plan to handle 1.1 model building? For example, if I am creating a large heart cell model made up of 20 smaller models do I have to cite each of them individually? Or at this point is the model sufficiently different that I can just licence it as one new model?

Discussion:

  • Andre: suggest that the emails from Catherine about this are copied or summarised in a new tracker item.

Action item 1: Dougal: Get in touch with Catherine about this issue.

2) (Alan) COR view in OpenCell - is this going to be fully tested and made available in 0.8 or not?

Discussion:

  • Justin: It might be possible to make the COR text view usable in 0.8; the major problem that was stopping testing was the incredible brokenness of the XULRunner editor component. If that is not an issue with our current version of XULRunner, then we should be able to properly test the COR view, and hopefully include it in 0.8.

Planning:

  • Justin: Look at the new XULRunner text editor and other issues. If the COR text view will be relatively usable, it might be able to be included in the 0.8 release.

3) Repository contributions - update

Progress:

  • Justin: updated the Saucerman model with the version from Geoff's VM that matches output with the MATLAB. Merged the working version with Catherine's properly annotated version. Also put up the scripts used to generate the comparative graphs.
  • Catherine: Coded up Hai and Murphy 1988 from new and curated Mijailovich, Butler and Fredburg, 2000, and Westermark and Lansner 2003.  Catherine has also added category keywords to about 40 models which weren't coming up under any category (Tommy found them by running a script).  Note there are still a handful of models with no category keywords.  These include Poul's units model and all the rest are 1.1 models which form parts of an importing model (which does have a category keyword).  Question - is this going to be a problem when we build the 1.1 model library - i.e. lots of small models with no metadata!
  • Hanne: Has created many OpenCell session files and is down to the last 3 sections of the repository.

4) PMR2 development - update

Progress:

  • Tommy: Has been caught up with upgrade issues to do with Plone 4 and its dependencies. The CellML skin has been fixed up for Plone 4.
  • Tommy: Some progress made on the curation module.
  • Justin: Has tested the converter script on the a1 workspace. It seems to work well, with the caveat that all unescaped entities that need to be escaped, will be in the output. Most of the models in the a1 workspace contain some unescaped entities that do need to be escaped.

Planning:

  • Andre: We need to find out how many models need fixing for e-notation and xml:base before we do anything else.

Action item 2: Justin: Write script to identify all models that require e-notation or xml:base fixes, and then run the script on the entire repository.

  • Andrew: No major changes are required to either the DVCS or application note papers, and these should be able to be submitted soon.

Action item 3: Tommy and Andrew: Send out final versions of papers for approval for submission.

5) CellML API - update

Progress:

  • Andrew: Has been working on the Python bindings for the CellML API. Good progress has been made.

6) OpenCell development

  • No update.

7) CellML core specs - update

Progress:

  • Andrew: has been working on ModML reaction mapping, a module to model reactions. This will involve a database of processes.

Discussion:

  • Mike: Will there be a version of ModML that can be used by students by November?
  • Andrew: Probably not, although there may be enough written up that students could work from that.

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

Progress:

  • Mike: Currently looking through a large number of ideas and issues

Discussion:

  • Andrew: Is it still the plan to have a core spec and write extensions to this spec as required?
  • Mike: This will probably be the way it is done.
  • There was a discussion about RDF and the problems it might cause.

Action item 4: Justin: Email Mike a link to James' work on the metadata spec.