ABI Meeting Minutes 2008-09-10

Present: Peter Hunter, Mike Cooling, Dougal Cowan, Randall Britten, Justin Marsh, Tommy Yu, Andrew Miller, Jonna Terkildsen, Michael Backhaus, James Lawson, Peter Schmiedeskamp

Apologies: , Poul Nielsen  Catherine Lloyd


Last week's action items

1) Justin will create a tracker item to discuss possible drag and drop functionality within the model tree view in PCEnv

2) Tommy to post this summary to CellML-discussion mailing list
  • Tommy has posted this.
3) James will put up the link the CellO ontology
  • James has done this.
4) Tommy to activate RSS on cellml.org/news
  • Tommy has done this.

This week's agenda

James - Discussion of who has been accepted for presentations and posters for Singapore.
  • Sarala was accepted for a poster only and is not keen to go.
  • James was accepted for an oral presentation and is keen to go, and to visit Andre's lab at the same time.
  • Mike does not think it is necessary to write a paper. James suggested asking the organisers about this, and Mike suggested James Goh.
Action item 1) James will email the conference organisers and confirm whether or not a paper is explicitly required.
  • Confirmation of attendees and who is to be presenting what should be done by the 30th of September.  Catherine to help organize this.

Tommy - Database workshop

  • Tommy attended a workshop on database research organised by the University of Auckland Department of Computer Science. He wrote two slides for Duane Malcom's talk, and received some feedback from the attendees. Christoph Lutteroth, in particular, was interested in Tommy's work on PMR2, and suggested a collaboration
  • Specifically, Tommy discussed the possibility of putting a relational database into future versions of PMR2 or even PMR3.
  • Using a relational database (RDBMS) for the model repository will require the models be decomposed into individual components, with metadata and results from experiment and simulation be associated with the component by using relationship within the database.  Queries can then be constructed that will return the specified components that can be used to construct another model.
  • This decomposition and relational databases raise versioning issues.
  • Randall said there are some benefits to this possibility, but that it is non-trivial to work out which way is better: the current/PMR2 approach where model files are stored intact; or the approach where model files are disassembled, and the sub-parts stored in appropriate database structures.
  • Andrew mentioned that Sarala's work on CellMLOWL has a similar goal, and using an RDF store might achieve the same objective as using an RDBMS.
  • It was agreed that this was a concept to be considered for the future of the repository but was not currently relevant.

James - Update on repository contributions

  • James has curated all but one of the circadian rhythm models.
  • Peter Hunter raised the idea of having an overview article (in this case on circadian rhythm models) written by an outside expert (Goldbeter?), that all related models could link to.
  • The idea of James and Catherine writing a paper with Goldbeter to use as this initial overview article was generally approved.
  • James has also been closing tracker items as he fixes broken links in PMR.

Andrew - CellML 1.2 spec progress update

  • Andrew had no updates.
  • Randall suggested that Andrew could create some mock-ups of CellML 1.2 and type system usage.
  • Andrew pointed out that he is also working on RDF support for the CellML-API, and that we needed to decide which work was the current priority for him.
  • It was generally agreed that support for RDF in PCEnv is a priority.
  • Tommy said that the metadata spec needs to be finalized.
  • There was some more discussion about the CellML Metadata specifications; Peter Hunter suggested a breakaway meeting to discuss these issues.
  • Mike discussed the idea of removing the "hiding" part of encapsulation which he sees no advantage in.  He sees encapsulation as having 2 concepts; the hierarchical tree of components, and the hiding of sub-trees.  He wants the structure without the hiding.
  • There was discussion around the terminology.  Andrew suggested "nesting" instead of virtual containment, to avoid confusion with the already existing containment.  "Packaging"; "public encapsulation" or "assemblies" were also suggested.
  • James asked whether we have a timeframe for the release of CellML 1.2, and whether we should in fact be spending more time working on the more immediate task of the metadata specifications. The general consensus was that CellML 1.2 would not be released for a couple of years

Justin - PCEnv progress update

  • Justin has implemented the start of a preferences service.  At present it allows the user to choose between the startup and shutdown behaviours for session loading.
  • Jonna would like to see a "reset PCEnv" (load blank session) option.  Justin said this should be relatively trivial to implement, and he will do this.
Action item 2) PCEnv users to obtain latest snapshot and try out the prefs service, and post feedback to tracker item 277.
  • Jonna suggested using relative rather than absolute sizing for sessions.  This will give sessions a consistent appearance at different screen resolutions.  Peter Hunter agreed that this was a very important feature.
Action item 3) Justin to create a tracker item for relative session sizing.  However, this would not be included in PCEnv 0.5.

Andrew - CellML API progress update
  • Andrew has implemented the entire object model for RDF.
  • Andrew has written code that translates between relative and absolute URIs.
  • A problem with processing models that have DAEs (was blocking Jonna's use of PCEnv for some of her models) has been fixed.

Tommy - PMR2 update

  • Tommy:  work continues on building the final PMR2 version.

Additional Items

  • Justin suggested moving the meetings to a larger room due to the increasing numbers.  This was generally agreed upon.
  • Michael mentioned that he had looked at CellO, and does not think it is particularly logically grounded.  He will discuss this with James.