Tue Jan 10 15:48:52 EST 2012
The main problem seems to be in the configuration. Two conflicting
properties of the current code base:
- Everything is generated automatically from the type of the update
method. Synchronous sampling is assumed. This is good if all the
types are the same (float*), but not so good if then they need to
be individually specified (float or float*).
- Type-directed specification probably assumes the structure of the
inputs. Is this always going to be straightforward? It better
be, because in Pd it needs to be..
From this it looks like the problem here is a tricky one. The good
thing is that it doesn't exist on the SSM level.
EDIT: see next entry