Meeting Minutes 23 March 2006

Present: Poul Nielsen, Dong Zhang, Steve McKeever (from Oxford), Shane Blackett, Peter Villiger, Sarala Dissanayake, Andrew Miller

Carey is leaving the institute in a few weeks for another job, but is spending some time finishing loose ends. He will have very limited time for institute work after that.

Carey has migrated the CellML Plone site to use Subversion. Searching and validation works, and searches can be saved.

Peter Hunter has suggested that a way to link to data be added into the CellML documents. There was some discussion about the best form that the reference should take — whether it is a simple URL in a namespace, or in the xlink namespace, or some other existing way to reference URLs.

Presented a proposal for storing simulation metadata. It was decided that this should be sent to the mailing list. Poul raised the possibility that we might be able to use intervals in MathML to do the same thing, and will think about whether or not that is a proper use of MathML. Andrew will send the proposal to the cellml-discussion mailing list.

The CellML API has now been completed, and autotools based configuration has been set up. It builds out of the box on Linux, but Win32 support has yet to be tested. Shane suggested that it be tested on Irix and AIX to make sure the configure script works. Carey also suggested Darwin.

mozCellML 0.3 has been released, essentially identical to 0.3rc3.

Andrew has made overconstrained models cause a warning and ignore the additional constraints, instead of aborting with an error.

Andrew has developed some code for drawing graph axes in Firefox using a Canvas, as discussed last week. Carey asked whether we should be using an existing Canvas graph library, and suggested that we also consider SVG. Andrew raised concerns as to whether having an SVG DOM for all data points will cause problems. Poul suggested that the decisions from last weeks meeting be revisited, as more consideration should be taken as to whether Canvas, or SVG, or cmgui is the best way to draw graphs. It was decided that Carey will produce a table showing the benefits and disadvantages of different approaches.

Poul mentioned that an annual report for various funding agencies needs to be written.

Peter Villiger has reported a bug in the CellML editor in which components differing only in case are treated as only one component when trying to create connections.

Dong has written an IDL for the CellML Event API. It is decoupled from the CellML API, and so implementations can choose whether or not to implement it. Events bubble up from the source element to the model element, and can be caught anywhere in this hierarchy.

Dong asked whether he should put time into fixing CellML editor bugs, given the upcoming changes(such as migrating to the CellML API) will likely make the fixes redundant anyway. The consensus was that as the changes were a long way from being landed, he should still support the current editor.

There was much discussion about how the CellML tools should be integrated, and Poul suggested that we should discuss our approach further before adopting the decisions from last week. Shange suggested that we should make one change at a time, and not change the GUI at the same time as the API. Carey brought up the issue of SVG, and suggested that if we adopt that change, there would be essentially no code left of the old editor.

Poul suggested that a meeting on CellML tools be arranged, in order to make some decisions, and to include Matt in the process. It was eventually decided that this should be at the normal meeting time, next week. Before then, Poul suggested that interested parties prepare statements about what we want to achieve, as well as enumerating the paths to achieve these, so they can all be considered next week.