abi-meeting-minutes-2009-04-29

Last week's action items:

• This has been done.
  

• This is tracker item 1727.

• This is tracker item 1728. Both David Nickerson and Jonathan Cooper posted feedback on this; both thought that things would be too complex if we added the addtional lists.
  

This week's agenda:

• Allyson Lister had asked about this so she can support CellML in SAINT.
• When the suggestion of using SBML-style RDF annotations was made, Sarala said she would rather go through OWL and her biophysical ontology as an intermediate to reference the citations.
• Matt noted there were a lot of places in the CellML model where we could use MIRIAM annotations without using OWL; he also suggested that Sarala's OWL annotations could be written back into the model.
• Sarala's work uses instantiated versions of CellML 1.1 models. This is essentially a flattened version where it is possible to reference every component after it has been imported, and make specific statements about it (rather than generic statements about the model being imported).
• Andrew suggested that we allow for properties to be declared on the imported model, and variables to be annotated relative to the properties. The value of the properties would then be annotated against the model.
• Action Item 1: Sarala will send an email to Allyson about this.
• There were also discussions of what was required by Morgan in order to implement this; it seems there is nothing they are waiting for a response from us on.
  

• As part of the process of moving to PMR2, James has downloaded all the models, and attempted to run them all through the validator so they can be annotated in PMR2 to describe whether they validate as valid CellML.
• The consensus was we should wait until after PMR2 is deployed, due to the fact that changes are being made to the repository on a regular basis.
• In particular, there is a problem with models being changed through the 'version replace hack' used in certain special circumstances to correct problems with models without formally creating a new version. James is currently dealing with this problem by asking people who know how to do this to e-mail him if they do it.
• Tommy discussed the migration script. Images are mapped to local URIs, which will result in image duplication. A breakway meeting is to be held on this issue.
• Jeelean updated the Elowitz 2001 model such that the model now recreates the published results.
• Catherine updated the Kongas 2001 and Cortassa 2003 models, and added the van Beek 2007 model.
• Catherine has been manually checking the legacy repository to ensure that every quantitative model has been transferred to PMR1. James noted that models in this repository were coming up in Google. Most models from it are in PMR anyway, except that some qualitative models which cannot be represented as quantitative models, without using reactions, were never moved. After some discussion, it was decided that the best option would be to simply redirect all URIs to the root of the new repository.
• Action Item 2: Tommy will speak to Gareth about getting Apache configured to redirect the legacy repository to the root of PMR. This can happen ASAP.
  

• Randall noted that a model which was being translated into CellML as part of the VPH project had required reset rules to properly represent a valve. Randall suggested that IDA might make it easier to represent this type of model, where the structure of the model changed so that variables could change from being rate controlled to being constant. Andrew thinks that we could also implement this without needing IDA if required.
• The same report mentions the issue we plan to resolve soon, related to the difficulty of identifying imports in the OpenCell user interface.

• Dagmar has circulated a draft paper on MIASE to solicit feedback.

• Andrew has been working on refactoring the tree-view code so that we can more easily add metadata editing into the tree.
• Action Item 3: Andrew will add a tracker item about this.
• Andrew, Alan, and Randall had a discussion last Friday about whether we should move to a platform other than xulrunner. QT and C++ was discussed - Andrew thinks this would be more of an efficiency loss than an efficiency gain. wxWidgets (possibly from some dynamic language) or Eclipse (possibly using Scala, Groovy, Jython, or JRuby) were other options.
• Action Item 4: Andrew will add a tracker item about this.

• Tommy has been refactoring PMR2 so he can more easily support the listings on the front page of the repository. He has also been working on the migration script.
• Mike is keen to have PMR2 ready to use for a paper (and noted that Peter Hunter is too). The software is almost ready, and Tommy was planning on making the software usable stand-alone, to de-couple it from the rest of the Plone site (but see later for how this plan has changed). A subset of the skins are required for this to work.
• Part of Mike's paper is about the infrastructure, so it would be much preferable if he could actually refer to the working PMR2 repository.
• Andrew suggested that we could set up a temporary PMR2 on a VPS, and then eventually point the URL for the temporary VPS to a final repository. Tommy estimated one day's work for this. Another option discussed was to have some URLs at cellml.org temporarily redirect to the new VPS.
  
• The possibility for Tommy to set up such a repository was discussed, but the final conclusion was it wasn't necessary.
• How PMR2 can be tested was discussed. There is a need for both automated and manual testing. Andrew suggested that new models could go up on the staging site, even if this meant that the exposures had to be re-done on the live site (with the actual models being kept in Mercurial and eventually pushed to the live site). Catherine doesn't yet know how to use PMR2, so needs a tutorial for this.
• Andrew suggested that any requests for help with PMR2 go through the mailing lists or preferably the tracker, rather than asking Tommy in person, so people outside of Auckland can benefit from this later. However, the consensus was that instead, the person being asking Tommy has a responsibilty to ensure that they or Dougal document whatever they learn.

• Andrew suggested that, rather than trying to tie the move to Plone 3 with the updating of the site, we instead migrate the content doing the minimum amount of work, so we have a Plone 3 site with PMR2, and then we update the content and the look-and-feel (skins) afterwards. It was agreed that this was the best way to make progress and move to the new site.
• Action Item 5: James & Randall will talk to Gareth about migrating the content to Plone 3 fairly soon.