Meeting Minutes 16 February 2006
Meeting Present: Carey Stevens, Andrew Millar, Poul Neilsen, Shane Blackett, David Nickerson, Matt Halstead, Sarala Dissanayake, Peter Hunter, Shane Blackett
apologies: Edmond Crampin, Peter Villiger
Sarala and Paul went to a systems biology graphical notation meeting in Japan over the weekend. Very good meeting, with representation from people using different notations. The effort was to develop a common notation that is more consistant and everyone uses. More info at www.sbgn.org, look at meeting papers, there are pdfs of all the talks and some discusions.
Sarala: Working on a comparison matrix of the different grapical noatations.
Got a grant a to spend a couple of months in Kitano's group in Japan.
Andrew: Released new version of mozCellML.
Need to track down the problem with the slow Noble98 model.
Some output to to find problems in models might be useful...how about a wall time variable?
Been working on the cellml API, making progress, but have a few issues...
How to asynchronously load models, current IDL is for synchronous load
What about thread saftey? DOM API is not thread safe. Make cellml API threadsafe and tell people not to use the DOM at the same time. This needs to be well documented, but not resolved at this point. Some examples might be helpful.
Need to write a comprehensive test suite for the cellml API. There are frameworks that could be used such as xunit.
Estimated time to usable - more than a few weeks.
Peter: Discussion of other sorts of graphical interfaces that could/should be used in mozcellml. eg the canvas element.
Seems many of the problems are related to the virtual windows implementation, will re-acess this interface design. Some time could be put into looking at using the canvas. Not a high priority at the moment.
Been using mozcellml to see how it could be made easier for a biologist to use. Problem with getting the right integration parameters and graphing paramenters. Suggest we keep the saved workspaces with the model or put metadata in the model. What do people think?
Need to be able to link data to models in the repository. Could possibly be done using a link in the metadata.
ie a way of us handling the data aspect of models without storing/curating it ourselves. This isn't part of the MIRIAM reference description.