ABI CellML Meeting minutes 2009-08-19

Present: Poul Nielsen, Peter Hunter, Amir Shirkhani, James Lawson, Randall Britten, Mike Cooling, David "Andre" Nickerson, Justin Marsh, Jonna Terkildsen, Andrew Miller.

Apologies: Catherine Lloyd

Last week's action items:
1) James to get redirects set up for the legacy 1.1 model pages that do not create dependencies on www1.cellml.org.
2) James to check with others (Catherine, Jeelean etc) about any repository contributions.
3) James to send an email to IUPS and to attempt to get feedback from tutorial attendees, both CellML and CMGUI.
4) Justin to look at LibSBML and try to work out what it is that we should put in our API for Morgan

This week's agenda:

  1. Proposal for CellML twitter feed

- James suggests using twitter to update members on news and feeds.
- Andrew suggested it should not be considered the official news feed.
- Poul and Peter consider it the same as the website’s news feed and like the idea.
- James suggested manual updates.
- Poul thinks it could be abused but a good idea.
- It was also suggested that the ww.cellml.org RSS feed could be duplicated on the tweeter feed.

2.    Look at some of the major SBML tools (e.g. COPASI) for GUI ideas for OpenCell
- New GUI idea for OpenCell : same as the MIRIAM annotation for an SBML tool.

SBML/CellML interconversion

- Andrew suggested the hardest aspect of the task is the lossless interconversion of metadata.
- Sarala is expected to work on this in the near future.
- There was a query about the timeframe for Java bindings on Windows.
- Randall said that Sarala has about 150 hours left and she would be working on the task 1 day a week.

3.    Campaign to get more externally developed tools (as with VCell and Jsim,) to support CellML import/export

[Sorry, no notes on this discussion.]

4.    Update re: work on CellML <-> SBML translation
- The latest is that Morgan said she requires the same metadata features as libSBML has.  Randall suggested replying to Morgan that we will look into what libSBML has.

5.    PMR2 update

- Jonna would like to know if some of the confusing aspects of the current search results can be addressed: A search for the term "guyton" returned 85 results where there were only about 20 actual guyton models.  Each model is repeated 4 times in the results.  There was much discussion about what exposures are and how they should be used, and how these things would affect this aspect of the search results, only some points from this discussion are noted:

 - Andre suggested displaying everything (result lines for matching workspace as well as exposures).
 - Mike suggested grouping, at the UI level, exposures by workspaces.
 - Andrew suggested that the default result format groups by workspace and automatically removes duplications of the exposures.

There followed a lengthy discussion about PMR2's exposure / workspace architecture and how it supports various functions.

Action Item: Updating the Tracker Item on customizing the search.

- Andrew suggested the changing of the terminology.
   Workspace -> project
   Top level exposure -> Release version
   Sub exposures -> release files

-Andre's view is that we should not change the terminology.

- There was discussion about whether the approach of the "XML patches paper" relates to the issues with variants and versions.  The paper mentions that traditional version control is not sufficient for modelling needs where combinations of different possible changes to a base model need to be explored.  Randall pointed out that the set of files consisting of a base model and its collection of patches is perfectly compatible with a version control system: The entire set is checked in as a revision, and then changes to any part of the set can also be checked in as subsequent versions of the set.

6.    Website issues – update
- "cellml.org" works now.
- Peter wants the CellML website to acknowledge the NoE logo for funding reasons.
- The placement of the logo was discussed.

Action Item: NoE logo and logos of other funding sources to be displayed more prominently on www.cellml.org.

7.    OpenCell development – update
- Justin has worked on the merging the work that he did on his branch with the trunk.

- The work on the other branch, that was done by Andrew, was going to be completed, and merged by Justin, but this has not been completed yet.

- A break-away meeting was called to discuss whether to make a release.
- Still issues on the Mac platform and a need for a new Mac machine.

Action Item: Justin to send specs for a new Mac to Randall

8.    CellML API – update
- Randall reported that Sarala has started working on the Windows build environment for java
- Andrew is making changes to CCGS to prototype event handling.  This is based on the proposal from the CellML spec meetings to use a piecewise for defining which equations hold under which conditions, and infinitesimal delays to refer to state variable values "before the event".

- The proposal does not include any direct handling of finite delays, but Andrew mentioned a way where finitely delayed values of a variable could be obtained using solely the infinitesimal delays.

-  Poul pointed out that convolution is core to linear systems modelling, and by definition of the convolution operator means that finite delays are needed: a(t) * b(t) is defined to be {integral dτ from -inf to inf of a(τ)b(t-τ)}.

- Python wrappers: Amir has written a new integrated test and has the basic structure to continue on his work back at UCSD.