ABI Meeting Minutes 2008-06-25

Present: Mike Cooling, Randall Britten, Justin Marsh, Tommy Yu, Geoff Nunns, James Lawson, Andrew Miller, Poul Nielsen

Apologies: Edmund Crampin, Peter Hunter, Catherine Lloyd

Action items from last week

1.)  Justin didn't get around to creating the tracker item to discuss the proposed workaround (PCEnv installer needs admin rights when installed in Windows in subdirectory of any directory with spaces in its name.)

Action item 1.)  Item moved forward to this week. "Justin to create a tracker item to discuss implementation of a workaround for this."

2.)  Tommy has made enquiries but the PMR2 prototype is not yet available externally (to the ABI) - this effort is ongoing.

3.)  Justin added the configure script to the latest PCEnv / API release

4.)  Justin used the digital signed installer for the Windows release of PCEnv

5.)  Poul has recorded his decisions re: the non-controversial items to be added to the CellML 1.2 specificaiton

6.)  Randall added "Update on repository contents" as a standing item to the meeting agenda

7.)  Catherine sent Andre an email but is away this week - James followed up

This week's agenda

James - Singapore meeting
  • James emailed Andre to ask him for an update on what the CellML team would need to do in order to get a forum at the ICBME meeting.
  • Andre said that the format would probably be 2 hours, 6 talks
  • We are running out of time if we want to do this: the abstract submission deadline is 30th July 2008
  • James mentioned that the conference registraton is not cheap - ~NZ$1000 per person
Action item 2.)  Poul to look into possible agenda and relevant funding issues

PMR2 Update
  • Andre has been looking at the prototype, and has been in correspondence with Tommy
  • Tommy said that he has looked into getting an external instance of the PMR2 prototype hosted by Amazon, but that it appears difficult
  • Randall is concerned about this, since he was under the impression that the ABI IT group had smoothed this out that that it shouldn't be a problem.
Action item 3.)  Tommy will talk to IT about this to see if they have addressed the issues he was concerned about or if not, will put in writing what he considers to be the problem so that Randall can discuss this with IT

Action item 4.)  Randall would like James, Mike and Catherine to write some notes discussing their thoughts on the PMR2 prototype

PCEnv / CellML API progress update


  • The PCEnv 0.5 roadmap has has had enough time for review by the PCEnv development team

Action item 5.)  Randall will post the PCEnv 0.5 roadmap to the CellML Discussion mailing list for feedback
  • Justin has been working on providing a better interface for visualising connections.
  • The information contained in a connection element - that is, the components connected, the variables mapped and the direction of those mappings - is now displayed in the same pane as the information contained in variable, units and equation elements when the connection element is selected in the model structure tree.
  • There has been some discussion within Tracker Item 415 over whether the icons that Justin plans to use to represent variable mappings between variables with incorrect interfacing are intuitive or not.
    • The current icons are: ->| for a mapping from a variable with an 'out' interface to a variable with no interface, and -> <- for a mapping where both variables have "out" interfaces.
    • James and Alan think that something extra is needed in the interface to indicate that these icons represent broken connections.
    • James suggested that a tool-tip could be used explain what the icons mean.
  • Mike asked if there are plans to include a graphical connection editor in the PCEnv UI.
    • Randall: yes, see Tracker Item 230 for a discussion. This functionality is planned for PCEnv 0.5.
  • Justin has also been working on the "text box entry misaligned" bug discussed in Tracker Item 104. The current proposed solution is to move the input field for editing variable fields to where the equation input field is.

  • Andrew has been working on the validator
  • Randall has created a new test case for the DAE solution approach
  • Code export needs to be adapted to allow code to be generated for models containing DAE systems.
Action item 6.)  Andrew to make a tracker item to discuss this

CellML 1.2 specification progress update
  • Randall asked if Andrew's work describing the proposed 1.2 typing system is on Git - yes it is.
  • James asked if there could be some kind of 'work in progress' link put up on the Specifications page at cellml.org
    • This is not really possible due to the decentralised nature of the Git system.

Update on repository contents

Curation flags
  • Mike has put up some ideas onto Tracker Item 87 (was 1108, but 1108 is duplicate), of which he gave a brief description.
  • Poul replied that we will need to tidy up non-orthogonal criteria for satisfaction of curation requirements. There may also be different degrees of satisfaction of requirements (other than a binary 'satisfies' / 'does not satisfy')
  • James explained his current conception of 'curation flags'
    • The system would be defined within at least a constrained vocabulary, perhaps a full ontology using OWL
    • There would be three main 'levels' (note, not similar concept to current 'curation levels'):
      • Major curation status descriptors would be represented in a glyph / icon format. The purpose of these would be to provide information intuitively and rapidly to users with experience levels from novice to advanced. There would be a small number of these glyphs - on the order of ~10. These glyphs may well be visible from the main repository listing, or may come up in searches.
      • Below these major descriptors would be a text based tagging system, which would be used to indicate whether a model satisfies certain conditions. These tags could be arranged in a hierarchical ontology, for example: the tag 'charge conserved' would be a subclass or instance of the class 'conservation.'
      • The third proposed level would describe the 'wellness of fit' of the tag. Values such as 'yes/unknown/no,' or scales such as 0=does_not_fit, 10=fits_perfectly, ?=unknown could be used. James noted that he has not yet thought out the implications of such a level, but it will be necessary to define how well a tag fits in some manner.
    • James is keen to build the system on a solid base so that is extensible and doesn't break when new terms are added.
  • Andrew noted that once the information structure of the curation flag system is decided upon, the interface should be much easier to develop.

Mac support for PCEnv

  • Mac support for PCEnv is in the PCEnv 0.5 roadmap
  • It was previously supplanted by the DAE solver and the validation service, which are now almost complete
  • The main differences between the platforms are in rendering of the user interface.