ABI CellML Meeting Minutes, 20th July, 2011

Present: Dougal Cowan, Caton Little, Tommy Yu, Andrew Miller, David "Andre" Nickerson, Poul Nielsen, Randall Britten (later)

1) CellML core specs - update

Discussion:

Andrew: last week's breakaway resolved the type issues for the primary spec. See https://tracker.physiomeproject.org/show_bug.cgi?id=47#c16
There was a discussion about last week's breakaway, concerning namespace mixing. It has been agreed that mixed namespaces will not be valid. See https://tracker.physiomeproject.org/show_bug.cgi?id=193#c7
Andre: will 1.2 models be able to import 1.1 models?
There was an extensive discussion about import and compatibility issues. There seem to be several options for dealing with this:

Convert all models in the repository, and provide tools for conversion. Expect that users will need to convert.
Provide a webservice where converted models can be obtained using the URL of an old model appended to a specific URL.
Say nothing about the problem and let tool developers deal with it.

Andre: perhaps 1.2 should be 2.0, because it will break backwards compatibility.
There was a discussion about these various compatibility issues.
Further discusssion was put off until the 1.2 breakaway.

2) CellML API - update

Progress:

Andrew: temporarily put the work on uncertain parameters on hold in order to prioritize the binary releases.
Andrew: CMake now supports running tests. Clean builds are now working on Linux & Mac, which required an overhaul of the buildbot configuration. Working on buildbot for Windows. Working on packaging with CMake, using CPackage.
Andrew: only 32bit Mac available at present. It may be possible to build 64bit Mac versions on 32bit with an updated version of XCode, but this now has to be purchased.

Discussion:

Andrew: it may be possible to build a Mac linker using the open source version from Apple, which would eliminate the need for XCode.
There was a discussion about how much work this would be.

3) PMR2 development - update

Progress:

Tommy: backported the security fix to 0.3 and released it. Work on the security plugin revealed a CSRF vulnerability in Plone, which has been reported to them.
Tommy: cleaning up the JSON API to make it easier to extend.

Discussion:

Tommy: Peter Hunter reported issues accessing cellml.org from overseas. This may be caused by some networking issue.
There was an extensive discussion about problems with the hosting for cellml.org and the repository.
Andre: the site[s] should be hosted somewhere else and tested, to see if the issue arises from the hosting.
It was agreed to test the sites on US hosting services, and potentially move them if the problems are not seen.

4) Repository contributions - update

Discussion:

Poul: Hanne is back encoding liver models & sessions.
Andrew: do we want to make sessions? Should we be making SED-ML files instead?
Dougal: it should be possible to convert sessions, and this would be desirable in any case due to the very large number of sessions already existing.
There was a discussion of summer students working on this project.

Action item 1: Andrew - write a summer student project for the session to SED-ML converter.

5) Time permitting: Updates or additional items from volunteering attendees

Andre: discussion of the COMBINE standards is ongoing. We need to have an elected editorial board to be a part of COMBINE. We should probably adopt the SBML procedures for this. Having an editorial board is a good idea anyway, to expedite approval and publication of specifications.
There was an extensive discussion about editorial board election.
Andrew: should use STV to avoid vote splitting.
Andre: we need to call for nominations for the board to get the ball rolling as soon as possible.
Randall: Dougal will be returning officer for the voting when it occurs.