ABI Meeting minutes 2009-03-11

Present: Randall Britten, Dougal Cowan, Justin Marsh, Julian Eden, Tommy Yu, James Lawson, Andrew Miller, Mike Cooling, Jonna Terkildsen
Apologies: Catherine Lloyd

Last week's action items:

1.) Catherine to inform other people who may be interested in attending OpenCell 0.7 CellML API 1.7 planning meeting.
  • Partially done.  There will be another planning meeting.
2.) James to create a past contributors/team members page, and update the team and past contributors pages. This will involve checking with Jonna as to whether she is part of the team.
  • Done.
3.) Catherine to send James the new profile pictures.
  • Done.

This week's agenda:

1) Discussion of the new CellML website.
  • James reported that Gareth wants to know the features the software pages will need.
  • Andrew said he thinks it is a bad idea to radically change everything at once.
  • James thinks the current software page is less than ideal.
  • Randall suggested that the working group could come up with a layout for the tools page.  He prefers an all on one page layout.
Action item 1.) Website working group to create a tools page layout template.

2) Website redesign - update
Action item 2.) URLs for the new templates to be distributed.
  • Randall suggested that Tommy put the templates up on the staging site for the benefit of external people.
  • Randall asked how easy it is to create a skin for the actual site from the templates.
  • Dougal reported that Gareth is working on this; templates to skin.

3) Planning discussion for Auckland contribution to OpenCell 0.7 and CellML API 1.7
  • Randall reported that the following items were made interim priorities at the meeting;
        • Renaming to OpenCell
        • KINSOL
        • Unambiguous component naming for graphing
        • COR-like text language
        • Simplify development environment using Mozilla builds of XULRunner for Win32 and Linux (already doing this for Mac).
        • Rudimentary IDA support (DAE solving)
        • Add automated regression testing of numerical parts of CellML-API i.e. CCGS/CIS
        • Reduce trajectory jumping from random guesses (done)
        • 2D layout is seen as a relatively large task, and likely to be deferred due to its relative priority.
  • Planning will continue at another meeting when Alan is in Auckland.

4) CellML API - update
  • Randall made the random guesses change such that CIS now accepts the first solution from the non-linear solver that is within the tolerance.  Since this is hopefully the solution obtained by using the previous timestep's values as the initial guess, this addresses trajectory jumping in some cases.  Jonna suggested that it might also speed up solving.  Randall agreed that in cases where the trajectory doesn't jump, it would be faster, since 1 non-linear solve would be being used per timestep rather than 100 non-linear solves.

5) OpenCell development - update
  • Justin has been working on multiple equation input; almost finished.
  • James reported that Glenn Ramsey was interested in whether he could use OpenCell to generate content MathML.
  • Randall asked whether it would be worth making the infix-to-MathML converter available as a stand-alone tool.
  • Andrew has been making further improvements to the text editor; now has reasonably efficient scrolling.
  • Randall suggested that a fixed-width font should be used, and that it would be nice to be able to choose the font.
  • Andrew has been working on renaming, PCEnv to OpenCell.

6) CellML Specs - update
  • James has been investigating restructured text and DocBook, as well as XML editing tools.
  • Randall said that it is worth going into DocBook sooner rather than later.
  • Randall suggested James use Mercurial instead of git; now has free hosting, and is much easier to use on Windows.
  • There was a discussion on version control systems for use with the specs documents.

7) Repository contributions - update
  • James reported that Jeelean and Linda are still making contributions.

8) PMR2 update
  • Tommy has got the CellML metadata library going - to render the metadata in a CellML file.
  • Metadata will be indexed and searchable, with a clean URI able to be used as a search on indexed fields.
  • An advanced search interface will be available for searching on non-indexed fields.
  • Randall asked if there will be push access to the Mercurial repository.
  • Tommy said that there are security issues with this - Hg would have to be wrapped in Apache for security.
  • Randall said it is worth spending time on this ability, as it is required for the launch of PMR2.
Action item 3.) Tommy to setup Mercurial wrapped in Apache for push access on staging server. 

This will need to be test-driven by model authors such as James, Catherine, and possibly Mike, Jonna.  This has been given a high priority, since it will be the only method for modifying repository models for the first version of PMR2.

9) Updates or additional items
  • Poul highlighted Alan's request for prompt minutes.
  • It was suggested that comments or corrections be given 24 hours to speed up the process.