Re: please advise....

[Nelson Minar]

Thanks for the thoughtful and important message, Ginger. Fulfilling
this need is high on our priorities. I hope that the new Swarm control
mechanisms will go a long way towards providing experiment management:
there is now an obvious place (the ModelSwarm and the ObserverSwarm)
where the parameters for an experiment need to be filled in.

The main question left is, how to get data from a file into the
running sytem? My basic plan is to create and use file probes, much
like we currently use graphical probes to fill in the GUI version of
the application. Ie, your data control file would be fed directly to
an object that reads files and uses probes to set variables. The
format of the control file then just becomes a text format for probes.

That's the basic idea, but it will take some fleshing out. How does it
sound?