Meeting Minutes 6 April 2006

Meeting Minutes 6th April 2006
Present: David (Andre) Nickerson, Sarala Dissanayake, Dong Zhang, Peter Villiger, Carey Stevens, Shane Blackett, Matt Halstead, Poul Nielsen, Steve Niederer, Andrew Miller

Shane: Has added a new paragraph to the Planning document about support for other direct language bindings.
Andrew has merged it into the section about XPCOM, and extended that section to make it clear that XPCOM is a language binding and Shane's point fits into the same general case.

Steve: Added some points about some features he would like in the CellML tools.
The consensus was that they were useful points, but that they were not relevant to the choice of framework and so could be stored for later use.

Poul: Personally believes that the Mozilla framework is the best option for our tools, and asks for objections(there were none, although Dong noted that the time taken for the Mozilla option could be greater).
There was discussion about how long it will take to get back to the current editor state.
Shane: Believes that 6 months is a rough estimate.
Matt: Believes that the U/I needs to be reworked at the same time, e.g. believes that the view showing connections isn't widely used, but agrees with Shane's estimate. Suggests feedback from potential users be solicited.

Matt: Noted that Sarala would probably be able to use the code via JavaXPCOM, as her code currently generates SVG and uses DOM events on it, but notes JavaXPCOM may not be complete. He believes that if Mozilla SVG supports <users>

Poul: Raised the issue of how we will provide graphics, and stated that he believes we should use a single technology.
Andrew: Disagreed with Poul, as he believes some technologies are suitable for some purposes, and others are suitable for others.
There was discussion as to whether SVG would be suitable for large graphs, and agreement was reached that it was not. This left cmgui, OpenGL, and Canvas as the remaining options.
Shane: Believes OpenGL is better because it will allow draw lists to be created and so transformations like zooming can occur without redrawing everything.
Andrew: Stated that he is not convinced that these make a big difference to performance, when compared with the portability issues that OpenGL will raise, and so prefers canvas.
Shane: Is happy for the implementors to make this decision.

Andrew: Raised the issue of what to do with his CellML Events specification.
Matt: Asked whether it should be in /wiki
Carey: Suggested that /wiki should be reserved for thoughts on issues, while specifications can go on other pages.
Andrew: Noted that wiki pages can be added anywhere in the tree.
Poul: Believes that we should wait longer for feedback before endorsing the CellML Events specification.

Andrew: Raised the issue of whether we should mix CellML namespaces in future CellML specifications to achieve backwards compatibility.
Shane: Believed that we should not, as he isn't convinced we can achieve backwards compatibility anyway, as old tools will not have enough information to make sense of the model.
The general consensus was that mixing namespaces for different versions was not a good idea.
Andrew: Asked whether we then should assign different MIME types for different CellML types.
The was much discussion, and eventually the conclusion that we should apply for distinct CellML 1.0 and CellML 1.1 MIME types, as this would make it easier to remove old features like reactions.