ABI CellML meeting minutes 2009-10-14

Present: Dougal Cowan, Catherine Lloyd, David "Andre" Nickerson, Tommy Yu, Andrew Miller, Justin Marsh, Peter Hunter, Geoff Nunns, James Lawson, Randall Britten

This week's agenda:

1) Should we enable commenting on the weekly meeting minutes? (Dougal)
  • Dougal asked if we should enable comments on the meeting minutes, so that applicable comments can be directly linked to the minutes in question.
  • Andrew said that we don't want to host these discussions on the website - this is the purpose of the mailing list.
  • Andre said that we could link to the mailing list from the minutes, linking to particular threads if appropriate.
  • It was suggested that it may be appropriate for the minutes editor to add comments as long as they are clearly separate from the main record section of the minutes.
2) IB2010 - Integrative Bioinformatics Symposium - Cambridge 22-24 March, (Catherine)
  • Catherine made everyone aware of this symposium which may be of interest to some people.
3) Updating the CellML Wikipedia page - compare CellML with SBML (Catherine)
  • Catherine raised the issue of our outdated and thin Wikipedia page - compare with the SBML page.
Action item 1) Dougal to update the CellML Wikipedia page.

4) Review of the curation flags (tracker item 87) (Catherine)
  • Catherine reported that attempts were made by herself and James to flag a cohort of six quite different models using a proposed set of flags. There are spreadsheets attached to the tracker item.
  • There was a discussion about curation, what the purposes of flags are, and what flags people want to see.
  • Peter asked what the summary flags are going to be, that is the ones visible in the main model listing and search results.
  • Peter said that we definitely need some indication of important curation features on the model list.
  • Andre said that we need some basic indicators as to whether a model is useful or not.
  • Andrew said that an advanced search  could be filtered according to flags.
  • Tommy said that he can put curation flag icons next to model names in the listing, but he needs the icons and categories.
  • Geoff suggested that there might be two levels of information required; basic user info (simple usefulness flags) and modeller information (detailed flags).
  • Andrew said that the use of an "unknown" flag status is useful because it shows that a curator has made an attempt to ascertain the status.
  • Andre said that curators should put comments (cmeta comment) in to detail changes made, problems with the model, etc.
  • There was a discussion about how to display curation flags on exposure pages.
  • David suggested three key flags: 1) does it run in OpenCell, 2) does it reproduce published results, 3) does it have associated "extra information" such as a session, diagram, SED-ML, etc.
  • Catherine said that another iteration of the exercise will be carried out.
5) CellML specs - update
  • Andrew has made proof-of-concept changes to the API, using top-level piecewise element and infinitesimal delay for reset rules, and these changes have been committed to the repository, for CCGS.  Part of this work requires changes to CIS, still to be done. {Clarification, was not stated at meeting: The goal is to follow through on an approach for handling reset rules that was proposed during some of the CellML core spec meetings a few months ago.}
  • A breakaway was called to discuss this.
6) Metadata specifications (general, citations, graphing, simulation, other)
  • James is refamiliarizing himself with the issues in the tracker.
7) OpenCell development update
  • Justin continues to work on the build environment installer.
  • Justin reported that Alan seems to have been working on the COR-view, adding error messages and refactoring code.
  • Catherine reported that Abhishek has asked if it is planned for OpenCell to ever support SBML models.
  • Justin said that it will always remain best to translate from SBML to CellML due to the way OpenCell works, and that something will probably always be lost in the process.
  • There was a discussion about how to use SBML models.
  • James reported that COR may not be definitive for unit checking CellML models - he has found that COR has problems with exponentials.
8) CellML API - update
  • Andrew has been fixing memory leaks.
  • Andrew reported that there are still some changes to be discussed for the API paper.
  • Randall suggested that we include Valgrind for testing in BuildBot tests (although the idea has been suggested by others before).  Cmgui (another ABI project) already has this.
9) cellml.org update
  • Catherine reported the changes to the community section.
  • Catherine asked whether the website meetings will become Physiome site meetings - the cellml.org group still want to meet. Peter suggested alternative weeks.
  • Tommy fixed some bugs in the Plone site which were only present once the live site was upgraded.  Dev site was not similar enough to live site to expose bugs.
  • Catherine again said that we are waiting on a development site.
10) Repository contributions update
  • Catherine said that Alan and Penny sent a list of models that have been checked; 42 are OK, around 12 need fixing.
  • Catherine has now got the Purkinje fibre model from Manchester (Aslanidi 2009) working.
  • Geoff has added some cardiac signalling and cyclic AMP models.  He has also completed a SERCA pump and cardiac cell model as requested by Jussi Koivumaki.
  • Catherine reported that Pulasthi will be coming in tomorrow to provide training in SVG diagram creation.
11) PMR2 development - update
  • Tommy has been thinking about how to refactor the way exposure pages are implemented, in order to have a rollover-like function in place. This is targeted to be part of PMR2 version 0.2.
  • Existing exposures will be converted.
  • There was a discussion about what these new exposures make possible, custom content etc.
  • Peter asked when we will see other languages available for download from the exposures (Python, etc).
  • Tommy said that he needs executables to wrap, or a completed Python wrapper of the API.
  • There was some discussion.  Andrew said that another option is to use CORBA.
  • Andrew reminded us that Adam Reeve had created "testCeLEDS", which is like "CellML2C", but allows a CeLEDS language definition XML file to be provided.
  • Randall said that other options would be to use CORBA (which is problematic) or wait for the API to be completed.
  • Randall summarised as follows: Using CORBA is too big a task to be included in PMR2 v0.2.  Finishing the Python wrapper of the API is also too big a task for v0.2. If "testCeLEDS" can be used, it is the only option that is viable for inclusion in 0.2, and is also desirable.
Action item 2) Tommy to estimate the time to get code downloads working using Adam Reeve's code.

12) Time permitting: Updates or additional items from volunteering attendees
  • There was a discussion about summer student projects.
  • Peter asked about the status of a 64bit Linux version of OpenCell.
  • There was a discussion about what the problems surrounding 64bit OpenCell are, and their solutions.
  • Andre said that 64bit Linux will be an increasingly important platform that will be worth testing on.
  • It was agreed that ABI-IT would be asked to install the 32 bit libraries on Jonna's machine.