ABI CellML Meeting Minutes, 22nd September 2010

Present: Dougal Cowan, Andrew Miller, David "Andre" Nickerson, Tommy Yu, Justin Marsh, Randall Britten

0) (Andre) Summer student projects

Discussion:

Andre: There is a student for a last minute project on neuronal population modelling, looking at extending CellML to handle these kinds of models.
Andre: The two repository projects have students.
There was a discussion about the projects.

1) Repository contributions - update

2) PMR2 development - update

Progress:

Tommy: will submit the application note today.
Tommy: has been refactoring and writing test cases.
Tommy: has decided to use sequential IDs for exposure IDs, in the absence of discussion on the tracker item (tracker item 2541).

3) OpenCell development

Progress:

Dougal: testing on Linux and Windows is complete, with no errors.
Justin: testing on Mac will begin shortly.
Justin: Jonathan Cooper has requested a 64 bit Linux version of OpenCell. Justin is not planning for this release to be an official OpenCell build, only an unofficial build with less testing done. OpenCell Linux and Windows builds currently ship with a cut down version of gcc, which means that OpenCell works even on a system with no system gcc installed. Ubuntu and several other distributions don't come with gcc installed by default. Working out which files are and are not needed for gcc to work takes a fair amount of effort, and Justin was attempting to upgrade to the latest gcc at the same time. There were several options - provide Jonathan with a version that needs a system gcc (we wouldn't want to do this for an official release), try to build an older version of gcc and use the same set of files, or continue work on building a minimal gcc of the latest version. This work will help with the upgrade of gcc on other platforms as well. Andrew noted that it could be useful to provide this gcc build alongside the API rather than OpenCell in the future, as it is used by CIS, part of the API.

4) CellML API - update

Progress:

Andrew: working on trying to get IDA to work with models that have reset rules. This is currently not working.

Discussion:

A breakaway was called to discuss the IDA/reset rule issues - attendees, Randall, Andrew, Justin.

5) CellML core specs - update

Progress:

Andrew: working on ModML reactions, translation of models that use units back to the unitless basic model.
Andrew: will probably be most useful to hard-code time at the reaction level. Andrew noted that ModML Reactions now hard-code in time as the dimensions of the variable of integration - this makes it simpler to use, but means that it is not possible to make a reaction model using this component using anything other than time as the independent variable (which is unlikely to be a significant limitation, and as ModML Reactions forms part of the model, and not part of ModML core, a variant without it could always be produced if needed).

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