ABI CellML Meeting Minutes, 3rd November 2010

Present: Dougal Cowan, Randall Britten, David "Andre" Nickerson, Mike Cooling, Catherine Lloyd, Poul Nielsen, Andrew Miller, Tommy Yu

1) CellML Workshop and HARMONY attendance. (Andre)

Discussion:

Andre: It would be good to make sure we have a strong presence at HARMONY. COMBINE is the replacement for the SBML forum, and HARMONY for the hackathon. It would be a good place to provide help for people who want to add CellML support to their tools. It is to be held some time in April or May 2011 in New York. We have a few projects we could focus on for HARMONY - annotation, metadata, the repository, and SBGN.
There was a discussion about who wants to or is able to go, and where to obtain funding.
Catherine: the CellML workshop could be in April, and is unlikely to clash with HARMONY, or it could be in February. If in Feb, it could be tied in with the new recruit CellML tutorial.

Planning:

Action item 1: Mike to check out funding sources to get people to HARMONY.

2) API priorities survey (Andre)

Discussion:

Andrew: there was a discussion via email about the survey. This discussion could be sent to the CellML discussion mailing list, and to users who are not using the API.
Randall: it is not that obvious from the website that there is an API and source code available.
Andre: it would be useful to find out why people have not even considered using the API.
Randall: we need an ongoing process of input and feedback on the API. Can we promote and provide information about the API on the website?
There was a discussion about the importance of documentation for the API, and the goals of the survey.
There was a discussion about problems with using the API - installation of the build environment, lack of documentation, etc.

Planning:

Poul: the survey could identify people who want to use the API, and what problems they perceive with it.
Randall: releasing binary versions of the API could make the process of using it easier (by bypassing the build step).
There was a discussion about the release process.
Poul: when the survey is sent out, ask for feedback on how to increase the API's usability.

Action item 2: Dougal to put "Developing for CellML" on the home page of cellml.org somewhere.

Action item 3: Randall to revive survey planning email thread, and add statement on the current state of API resources.

3) CellML API - update

Progress:

Andrew: trying to get the Shi, Hose, et al. model running. This is currently failing to converge, due to problems with the difference between <= and <, etc.
There was a discussion about how to solve this problem.
Andrew: continuing work on TeLICeMS.

Discussion:

4) CellML core specs - update

Progress:

Andrew: ModML version of Rice 2008 is now almost entirely coded up, but is having problems converging using IDA.
Andrew: new debugging facilities have been added. The ModML Core now supports a no-op function which allows string identifiers to be attached to parts of mathematical expressions to help identify the source of incorrect mathematics (which may have gone through several transformations to get to the final form, making it hard to identify the source of the problem otherwise). A similar operation is supported by the units module, and passed through to ModML Core. In addition, the units module now has support for 'units assertions' which will give an error if the units of an expression are not equal to a particular unit - making it easier to isolate units problems in complex models.

Discussion:

Randall: Alan was wondering if anyone else is up to speed with ModML. Randall, Caton and Richard have already tried to make sense of the publicly available ModML code, but have found it unclear.
Andrew: is writing a chapter on ModML for thesis, which might help. There was discussion about the possibility of Andrew circulating the chapter, which may provide him with useful feedback on ModML, but he has some reservations.

4) Repository contributions - update

Progress:

Discussion:

Catherine: Alan commented on the problems with heart models recently curated by Penny, which COR has stripped the c:meta ids out of. Hanne has fixed the session files broken by this.
There was a discussion about the c:meta id stripping problem.
Andrew: a small tool could be provided for adding c:meta ids back in.
There was an extensive discussion about RDF/metadata editing, tools and user interfaces for this, and editing models in text editors. There is a definite need for more direct metadata support in OpenCell.
Randall: Alan has said that metadata editing is going to be a key feature of the new OpenCell. We will thus have to wait for Alan's OpenCell.
There was a discussion about metadata editing features required in the API.

Action item 4: Mike to create a tracker item about metadata editing features for the API.

5) PMR2 development - update

Progress:

Tommy: finished the foundation for dynamic file rendering. Now writing plugins to render file types separately. Exposure/workspace merging. Also working on a sequential ID product for exposures.

Discussion:

Randall: the application note might need someone else to take over the changes and re-submission process.
Catherine: happy to take over the rebuttal/changes and re-submission of the application note.
Mike: PMR2 needs documentation
Dougal: has published some basic PMR2 documentation on the repository site.