ABI CellML Meeting Minutes, 15th of June, 2011

Present:

Dougal Cowan, David "Andre" Nickerson, Andrew Miller, Randall Britten, Tommy Yu

This week's agenda:

1) Should we rename the CellML API implementation to reflect a broader scope for the project? Tracker item 2955 (Andrew)

Discussion:

  • Andrew: it makes sense to first decide the scope of the API, and then decide on any name change.
  • Andre: the CellML API should contain everything people need to work with CellML.
  • There was a discussion about which services should be "inside" or "outside" the API.
  • Andrew: one precedent to look at might be the MathML IDL from W3C.
  • There was a very extensive discussion about who is using the API, the implementation, and why people are or are not using it.
  • There was a discussion about how to potentially improve the language bindings.
  • Andre: would like to see current API kept separate, and any additions made on a branch.
  • Randall: this is essentially already how the development occurs.
  • There was a discussion about how to separate out parts of the API/IDL/implementation.
  • Andre: can we decide not to add anything to the interfaces unless a tracker item is made, a discussion occurs and agreement is reached?
  • Andrew: a VPH share branch may be required for features required by VPH share.

2) CellML API - update

Progress:

  • Andrew: confirmed that Alan is using a recent version of GCC that comes with Qt Creator.
  • Andrew: have been going through tracker items and fixing bugs. Fixed a bug in the integration service that helps the Shi Hose model integrate a little further.
  • Andrew: for VPH share - completed xpath implementation. Could not find any test examples, so created some test cases. These all pass - the xpath support is complete. Working on a SED-ML running service.

Discussion:

  • Andre: is there any chance of getting units inference?
  • Andrew: the prototype is written in Ruby, could be possible to add it.
  • There was a discussion about SED-ML support.
  • Andre: SED-ML support shouldn't be part of the CellML library/API.
  • There was a discussion about reasons for the use of xpath and why we have our own DOM implementation. Xerces was discussed as a defacto standard XML library.
  • Andre: tracker item 2954 - inconsistencies in typedefs.
  • Andrew: this issue will be documented, but not fixed.

3) CellML core specs - update

Progress:

  • Andrew: have been through the comments on the uncertainty draft spec, made various changes according to these comments, and pushed the new version to github. What should we do with the spec now?
  • Working on SED-ML support to enable uncertainty implementation.

Planning:

  • Randall: try to get a working implementation prototype of the uncertainty spec.

4) Repository contributions - update

Progress:

  • Dougal: made a new exposure for Alberto and Alan of the Corrias et al. 2011 model. Also working on a simpler RDF template for adding basic metadata to models.

5) PMR2 development - update

Progress:

  • Tommy: working on cleaning up the UI. Added bigger better download buttons. Added direct link from workspace to exposure.
  • Tommy: the 0.4 release is waiting on Plone 4.1.

Discussion:

  • Randall: is the CellML simulator view going to be in 0.4, or wait for the next release?
  • Tommy: it is in 0.4 in a basic form, and should work automatically. Currently only using pre-generated output data.
  • Randall: VPH share is focused on web-services, so these will be an important priority for future development.
  • Randall: there was a long meeting yesterday about features PMR2 would need in order for ABI modellers to use it. A document of requirements will be written, and this may lead to another developer working on PMR2.
  • There was a discussion about the future of PMR2.