Bug 350999

Description Miles Parker 2011-07-02 16:04:23 EDT

As I mentioned in the other bug just now, I couldn't get the SIR model to work, but I think I understand what it is doing.  I think it is possible to get done the exact same thing in a more integrated way, at least I'm pretty certain I can get the same semantics that are needed for both cases, which is of course the whole point of putting these things together.. But I might be missing something obvious. Can you guys me feedback on the following proposal? 

1) I'm not sure why we need to differentiate between "dynamic" variables and "attributes"? Attributes can change, so in that sense they are dynamic. Can we not simply extend attributes to give them "dynamic" behavior?

2) Similarly, is there a reason that we need SdNamedElement? Why can't these be using SNamed?

3) We should be able to fully specify the stocks and flows component with analogous actions on the components. These can be managed by the AMF editor, please take a look at how Derived works -- that creates 

4) It looks like semantics right now are defined simply as strings. By using the actions framework we can integrate all of that.

With these changes it seems that we probably don't need to have an extension to the metaabm.acore, we can include everything needed in the core model itself. I know we've been reluctant to make major changes to .metaabm itself, but I think this is a good time to tackle that, I think we'll get a much more integrated solution out of this.