ABI CellML meeting minutes 2010-02-17

Present: Catherine Lloyd, Randall Britten, Mark Hanna, Ethan Cho, Tessa Paris, David "Andre" Nickerson, Tommy Yu, Justin Marsh, Poul Nielsen, Andrew Miller

Last week's action items

  1. Tommy to add a "how to cite" link to the PMR2 list. This depends on method for linking to models, and is part of the plan for the next version of PMR.
  2. Justin and Mark to discuss where the work Mark did on improving the sessions to make them zoomable was kept. Discussed in a breakaway after the meeting.

This week's agenda

Scripting the replacement of incorrect e-notation constants

  • Catherine requested that someone apply the script to replace e-notation in the incorrect 1E10 style form with proper MathML e-notation. She noted that we said we would do this in May 2009 in our announcement to tool developers. At the moment, this process is done manually. Catherine would be happy even if the workspaces were updated, but they needed to be manually rolled over. The script Tommy has at the moment doesn't commit to Mercurial or do the roll over, but it could be updated to do this.
  • Action Item 1: Tommy will look at the tracker item to see if any the reasons that we did not run a script over the whole repository earlier still apply.

PMR2 development

  • There was a discussion about how to deal with user workspaces, and what URL they should have. Several options were discussed:
    • Andrew suggested having no prefix - for example http://models.cellml.org/username/workspacename . Tommy was concerned that this would mean someone would take a username that used up valuable words in our namespace that we might want in the future, for example, a username like cellml.
    • Justin suggested using tilde usernames - for example http://models.cellml.org/~username/workspacename. Tommy and Randall were concerned that this looked old-fashioned.
    • Randall also suggested workspace_2 - for example http://models.cellml.org/workspace_2/username/workspacename. Andrew didn't like the length of this, Tommy didn't like the fact that it contained a number.
    • Andre suggested putting the item in Members as is common for user content in Plone - for example http://models.cellml.org/Members/username/workspacename. Randall pointed out that the users didn't really have full control over the namespace - the main point is to avoid collisions of workspace names, not to give users full control over their namespace (such as the ability to delete published models). Andre suggested that putting models in Members would give users control to create their own exposures. Tommy confirmed that users will be able to make their own exposures anyway. Andrew suggested that, rather than putting it in Members, users should be able to annotate any exposure, but it should be clear whether these are annotations by the officially endorsed curators, or by another user. Everyone agreed that this was a separate issue, and that there was no need to put content in Members.
    • Tommy noted that Alan Garny has suggested using http://models.cellml.org/w/username/workspacename. Consensus was reached that this was the best approach.
  • Tommy has made some changes to PMR2 [Tommy might want to expand this... I missed the details]

Repository contributions - update

  • Some models have missing RDF in the session files; Catherine suggested this may have been caused by PCEnv 0.6, although it isn't an issue with OpenCell 0.7pre snapshots. Many of these have been fixed, and Tommy is working on a script to find models with this problem.
  • Mark has added the ability to reset the view (i.e. undo all changes like zoom to the original state) in the SVG viewer in the session template.
  • Action Item 2: Ethan to push his remaining models, even if they don't work, and notify Catherine of which ones need fixing.
  • Ethan also discussed an issue he had with piecewise functions. To express that a value is between an upper and lower limit, he discovered that he needed to use two inequalities joined together with a binary and operator. In addition, in one place, he is relying on the fact that the CellML API treats true as 1, and false as 0 in an equation. Andrew suggested using a piecewise for this instead, as otherwise, it is invalid CellML.
  • Randall noted that someone had told him that the need to reimplement models from a paper, rather than use a working CellML model, was a good thing, as it allowed people to better understand the model. Consensus was that CellML was still a good thing, because the paper might not contain enough information to correctly reproduce the model, however good one's understanding, that people can still understand a model from the CellML, that if someone uses another model and doesn't understand it, reviewers will likely pick up on it, and that some users already understand the biology and just want a working model, and that some users don't need to fully understand the model to use it.
  • Catherine noted she had been working on the endocrine models.
  • There was consensus that the talks by the summer students on Monday were excellent.

CellML core specs - update

  • Andrew has been working on the ModML solver - it now generates C code, using IDA, but this is not yet fully working.
  • The SBML workshop in Seattle in coming up, which Andre is attending.
  • Dagmar may be going to present on SBML at the CellML workshop, if Evo can work; otherwise Andre is presenting on her behalf.
  • Action Item 3: Catherine to test out Evo with Dagmar to see how well it works from the US to NZ

OpenCell update

  • Justin has been working on fixing keyboard navigation issues with the treeview, and has made another snapshot to test to see if it is suitable to be RC2. Further testing on Windows is still required.

CellML API update

  • Andrew has been working on the NSIS based installer for setting up the build environment. At present, there are issues with getting it to work through proxy servers, but Andrew is working on this.
  • Andrew noted that he now has the go-ahead from Matt Halstead to submit the revised CellML API paper, and submitted it this morning.