Meeting Minutes 14 November 2007

Present: Randall Britten, Peter Hunter, James Lawson, Catherine Lloyd, Justin Marsh, Andrew Miller, Poul Nielsen, Tommy Yu

Apologies: Vignesh Kumar

Review of last week's action items

  • Poul will gather those material about CellML 1.1.1/1.2

    • Still need the documents and waiting on Sarala's diagrams

    • Action item: Poul will organize meeting with James and Sarala on the materials and diagrams.

  • Andrew will set a public Git repository for the development of this document as a trial.

    • The Git repository has been set up to assist development of the CellML 1.2 specification

  • Poul will send email about starting the drafting of the next CellML specification

    • This has been done.


Agenda

Update on CellML 1.1.1 and 1.2

  • Peter suggests that we notify sites that use CellML about how deprecating reaction elements may impact them

    • James sent the CSML group (http://www.csml.org/) an email mentioning that the repository has changed significantly since the date that they imported files from it (and converted to their format) - 13th Sept 06. He also mentioned that we are updating our repository and since this is likely to change our versioning system, it is probably better if they defer the re-import of models until the next version of the repository is implemented. James will notify them when the new software goes live.

    • Action item: James will add an issue in Bugzilla pertaining to above.
    • Action item: James will also look at/notify other repositories about deprecation of reaction elements.

  • Poul is leaning on going straight towards CellML 1.2, as the time remaining towards the final specification for that is not too far away; having CellML 1.1.1 between now and CellML 1.2 may give the users the impression that the CellML specification is unstable.

    • Creating the CellML 1.1.1 specification will take some time, even if it may not be much.

  • Community input will be needed, and Andrew's Git repository will be able to distribute drafts and facilitate discussion in getting this new specification up and running.

    • Time will be spent gathering feedback

    • Then the specification will be drafted up.

  • Andrew raised the issue about making the normative specification split out from the informative section

    • He is currently working with that while making commits to Git.

    • Using other specifications and RFC's as guidelines to polish the final form of the specification.

    • The format is being discussed on. There has been no major objection on using Docbook as the base format for the specification, but will allow objections/other suggestions to be raised against this choice until next week.

    • The version management using Git is still under discussion, as people may not want to deal with yet another piece of software.

      • It was suggested that we can allow people who would love to contribute to put in suggestions in the mailing list, and delegate specific individuals who will use Git to update for that individual who contributed.

      • The choice to use Git is still in the open; someone will need to send the email to summarize the available choices, and set date to close off any objections

      • Action item: Randall will send out the above email.

Update on PMR for CellML 1.1

  • Tommy noted that he has gotten more Git features integrating into Plone.

  • Randall noted the need to give this product a name.  The consensus was that it will be called PMR version 2, abbreviated PMR2.  It was agreed that this decision did not warrant community discussion.

  • Poul wanted to know how and why any model can be allowed in the repository

    • Rationale was the repository was originally planned to be a place where anybody can contribute their model, but most cases most new models are not release quality.

    • To address this problem there will be multiple views to the repository, where one view will be for models that are high quality (curated view).

    • There will also be guidelines and minimal requirements to allow inclusion of models that are not ready to be published.

  • Peter wanted to know how components will be broken up and how they will be presented.

    • Currently the models are in CellML 1.0, or flattened models that will need to be broken into components. Those components will then be labeled and be published to the component view, which users can see available components that are usable by them.

  • Randall noted that the current process to upload the model will not result in the same model the user uploaded because the process involves addition of metadata on the fly, which will be incompatible with a source code management system.

    • He suggested that PCEnv may include metadata editing capabilities.

    • However Andrew noted that not everyone uses PCEnv to edit metadata

    • To address this, Randall explained the following approach: the PMR2 prototype will commit the submitted version to the repository, and prompt the user to use the web-based metadata editor.  Once the metadata editor changes are saved, this will be committed as the next version of the model file.  This allows the version history to capture the state of the model before and after the use of the metadata editor.

    • Poul noted that the required metadata must satisfy the requirements of MIRIAM.

      • Randall explained that Git or Subversion usage would allow for the following metadata out-of-the-box: author of each version, date of each version, change notice for each version.

Updates on PCEnv 0.3

  • Justin noted that he still need to create a branch for functional testing

  • Imports are broken under certain circumstances (regression), and this is marked as a blocker in the tracker.

  • This release will include Javascript with SVG

    • Justin created a model with SVG with Javascript to demonstrate this feature

    • Action item: James will put this demo up on the CellML site. The demo currently would trigger Javascript errors in PCEnv 0.2, since it is designed for PCEnv 0.3.  These could either be ignored, or the demo modified to detect the necessary API to avoid the 0.2 errors being listed in the Javascript console.

SVG diagram creation

  • James is in the process of fixing the missing arrows caused by importing diagrams into Adobe Illustrator from Inkscape.

    • Justin noted that there is an extended SVG format that Inkscape uses that would cause this incompatibility problem. However, Inkscape can be set to save to generic SVG, so that extensions are not used.