[Scilab-users] Persistent data in scicos computational functions

[john fulgor]

Hello all,
I'm trying to create a new scicos block, based on a scilab computational function. The computational function creates some constant matrices at start, derived from the parameters, that it uses these matrices all through the simulation. My question is this: where can I store these matrices? Am I supposed to re-compute them all times the function is called? Or: should I store them someway in the discrete state vector, reshaping them someway? And what are the "Discrete state objects" mentioned in the model documentation? 

Thank you very much!
John