ABI CellML Meeting Minutes 15th December 2010

Present: Randall Britten, Shorojeet Dasgupta, Amy You, Lukas Endler, Tommy Yu, Peter Hunter, Poul Nielsen, Andrew Miller

Last week's action items

Last Week's Action Item 1: Andrew to look at the Antimony code for CellML handling and determine how hard it would be to help Lucian move it off XPCOM. Andrew thinks there are a few hundred lines of code involved, and that it would probably be less work long term than maintaining an XPCOM binary. Some of the changes should just be search and replace. The next step is for Andrew to start looking at changing some of the code over to use the CellML API without C++ to see how long it will take.

This week's agenda

CellML Repository model tracker items

Discussion

Randall gave a summary of the situation. Peter was initially in favour of having fine-grained components on the tracker; Poul suggested an alternative of having a single person as the port of call for repository issues who will delegate changes on to the appropriate curator.

Planning

The consensus was to keep the fine grained categories for now, but to later consolidate them so all curators get all repository issues, perhaps with one person responsible for ensuring every issue gets followed up.

Repository contributions

Progress

Shorojeet has been working on the Butera et. al 1999 model, which includes normally distributed random variables for neural network connectivity. He connected the model up using hard-coded random numbers, and has it working.

Physiome Model Repository

Progress

Tommy has made more progress on the refactoring, and has now refactored the code for merging exposure views into the workspace, and also the model storage support. These changes should simplify testing and future changes.

Planning

Tommy is now working on the archival code.

Randall made a suggestion that we could display thumbnails of models (for example, of a FieldML model).

Tommy suggested that exposures could be auto-generated rather than being a copy of the documentation for one file (in the case where there is a top-level container which groups several related models from the same workspace).

Peter asked about tracker item 2799, which relates to how PMR2 displays MathML which is inside an imported model (currently not at all). Consensus was to make it easier to access the MathML, but not to provide the MathML from a flattened view of the model.

Discussion

Peter noted that it is hard to access FieldML. Randall noted that it is possible to link to FieldML manually from the documentation. Once the repository gets FieldML support, the situation will improve.

CellML API

Progress

Andrew has been continuing work on TeLICeMS; the memory leak mentioned last week has been fixed. Most models successfully round-trip from TeLICeM back to CellML, but there still appear to be some problems with imports.

Amy has created a wrapper around RunCellML using C# as part of the effort to allow CellML models to be used from SEDML. Andrew noted that RunCellML was not intended as a stand-alone tool and so this approach may not be robust, and that it is more reliable to use the CellML API directly.

Discussion

There was a brief discussion of Alan Garny's progress. Alan has set up a new Sourceforge project under the name OpenCOR, which now contains a README file in the text repository. Peter noted that Alan is working full-time on OpenCOR now, but does not have anyone else helping him at this stage.

Core specifications - ModML

Progress

Andrew has been working on supporting sensitivity analyses on ModML models as part from the autosolver.

Core specifications - CellML 1.2

Discussion

The progress made on CellML 1.2 over the last year was discussed. The CellML API now experimentally supports reset rules. A breakaway meeting was held to discuss the numerical problems currently being encountered with this approach.

Other issues

Progress

Andrew has now resubmitted the paper on Distributed Revision Control for model repositories.

Andrew sent out another draft of the Springer Systems Biology essay on CellML which addresses David Nickerson's comments. Andrew hasn't been given a firm deadline, but expects that it will be soon if the encyclopedia is going out next year; suggests we should submit it this year. Poul is meeting with David later today, and will check he is happy with the version; unless Andrew hears otherwise, he will submit it to the section editor today around 5pm, with the caveat that some details are still under discussion.