ABI CellML Meeting minutes 2009-07-01

Present: Dougal Cowan, Amir Shirkhani, David "Andre" Nickerson, Tommy Yu, Justin Marsh, Randall Britten, Catherine Lloyd, David Szekely, James Lawson

Apologies: Peter Hunter, Poul Nielsen

Last week's action items:

1) Tommy to make a list of models that lack "keyword" metadata.
  • Tommy has not yet created the list of models that lack "keyword". Action item 1)
2) Tommy to arrange a meeting on Friday or Monday to discuss PMR2 plans.
  • Done.
3) James to create a session file for the Li, Smith, 2009 model.
  • Not done - a new diagram might be created. James still has doubts as to the usefulness of session files due to the volatility of the OpenCell UI.
4) James to send an email requesting comments and feedback about the new site.
  • Done. There has been some positive and useful feedback.

This week's agenda:

1) IUPS preparation
  • A breakaway meeting has been called for this item.
2) CellML converters
  • Catherine wants to be updated on the progress of any of the converters - MATLAB - CellML.
  • Martin Fink has a converter tool, but it requires hand-editing of the output. It takes two steps, MATLAB to COR text, then COR to CellML. This process takes about 90 minutes of work, and only Martin and Alan can do it at the moment.
  • Catherine asked if it is possible to create an interface for this tool.
  • Randall reported that OpenCell now has a text editor area for entering equations that can take multiple line input.
  • Randall asked if this could wait until December, when a summer student could be assigned to do it.
Action item 2) Randall to get the MATLAB converter onto the summer student list.
  • Catherine asked if any progress has been made on the SBML<->CellML converters.

3) VPH WP3 deliverable 3.5
  • Catherine sent the VPH report to Sharon on Monday, and has not yet heard back.
  • Randall asked whether there should be anything about the Java API in the report.
  • Catherine said that Peter is happy with the report. She asked if everyone is happy to cut and paste the material into the deliverables report for the first iteration.
  • There was a discussion about what the "deliverables" are.
  • Randall will look at the deliverables list for context.

4) PMR2 update
  • Tommy has been working on getting Mercurial updated to the latest version, and has written some tests that were missed from 0.1.
  • Randall suggested a follow-up meeting on PMR2 at 2pm.
  • Randall reported on Alan Garny's interface tracker items, which we will try to address for 0.2.

5) Repository contributions - update
  • James has put about half the CellML 1.1 models from the old repository up, has uploaded a couple of new models, and has been continuing with the mass curation.
  • Catherine has been doing lots of translation.
  • Randall asked if the RSS feeds are working; James reported that they are, but is unsure whether exposure changes will appear in the feed.
  • James says that workspaces don't come up in the RSS feed, only exposures.
  • James can no longer find the "pending review" list - Tommy says that he has blocked the portlet, but will put it back.
  • There was a discussion about version control and reversion.
  • There was a discussion about scripting the e-notation changes for James' mass curation effort.
  • David reported that the Ten Tusscher model is wrong - the gas constant is out by a factor of 1000. This is apparently well known. Randall said that the Ten Tusscher C code has the same error.
  • Catherine suggested contacting the authors to sort this out.
  • There was a discussion of David's particular problem with the Ten Tusscher model.
  • Randall raised the issue of capitalization of the Ten Tusscher model - different conventions for different places causing problems for PMR2.
  • There was a discussion of searches on PMR2 - workspaces vs exposures.
  • James suggested removing the relevance measure from the results listing, as it is often wrong.
  • Tommy said that search listing changes are part of the theme product - this raises the issue of whether we want to keep the theme identical across cellml.org and models.cellml.org.

6) Website
  • James reported that the website group had a meeting yesterday, going through the content and identifying changes that need to be made. Another similar meeting will be held this week.
  • Catherine asked who is contacting ITS.

Action Item 3) Randall agreed to do this since Poul is away.

  • Catherine reported that Peter and Poul are very keen to have the website up by IUPS.

7) OpenCell development - update
  • Justin reported that tracker items 345 and 1767 have been closed. Imported and encapsulated component hierarchies are shown correctly.
  • Andrew will review the design of this solution.
  • Randall asked if there is something we can do in the interim (while Andrew is working on the build system) to get some snapshots built. There was a discussion of how to get some new snapshots on the 3 platforms.

8) CellML API - update
  • Andrew has buildbot working for Linux & Mac - Windows still has issues (?).
  • Andrew is documenting everything that is required for Win32 builds. Windows currently has GSL disabled (which has no consequences for OpenCell users).
  • Randall reported that Amir is here to work on the Python wrappers.
  • Amir has read all the documents, and has run into some problems on Linux.
  • Randall said that Amir & Andrew should discuss these issues after the meeting.
  • Randall reported that Morgan is still waiting on Andrew for the Java bindings.
  • There was a discussion of the Java changes Andrew could make to fix Morgan's problem. Randall said that Andrew should interrupt his build environment work to get this Java problem sorted for Morgan.

9) CellML Specs - update
  • Randall said that there was a meeting on Friday. It was agreed that the next step is to take the proposal for expressing reset rules and create an example, and a proof-of-concept implementation in the CellML-API's CCGS for handling it. The proposal is to use a piecewise statement as the top-level math element when a condition determines which equation applies. This proposal does not require modification to the CellML 1.1 specification.

  • It has been decided to hold off on spec breakaways for one or two weeks (in favour of IUPS prep meetings).
  • Catherine asked what decision has been made about graphing and simulation metadata.
  • Andre proposed we drop our metadata in favour of SED-ML; see if we can use SED-ML, and if not then return to our own metadata.

10) Additional items
  • Catherine reported that Sarala's second paper was published today.
  • There was discussion about the timing of consumption of logo-cake.