Meeting Minutes 17 August 2000

Get the PDF!

1  Summary

The 17 August 2000 version of the CellML preview specification was presented to Poul and David by Warren to warm critical acclaim (something shy of the enthusiastic reception that was expected, but at least we're making progress.) Some minor points were made before the meeting was closed due to a rude interruption by John Kernohan.

I had suggested separating the input role of a variable from its output role. If a variable is required it is an input variable, if it is declared in a component it isn't an input variable. The scope of a variable (private or public) then determines its possible role as an output. I had envisaged separating scope from input role so that we could have various levels of scope, adding perhaps something akin to the Java concept of a protected variable where a variable might be visible to components that have a geometric "is-in" relationship to the parent component.

David was of the opinion that these should be combined into a single variable attribute (called role??) with possible values input, output, input+output, and none. This probably makes the specification of variables cleaner, so why not.

David was also confused about what I meant exactly by the term connection. He suggested that it sounded a bit physical, and maybe the word relationship might be more appropriate. I'm going to hold off on incorporating this into any documentation unless someone else is similarly confused — I would tend to say that the connection contains different types of relationships as discussed in the spec. Objection noted.