Connect to a running OpenSees process, avoid startup lag?

[mphair]

We've been working on a sketch-based frontend for OpenSees, and because of the relative simplicity of the models most of the time that it takes to perform simulations is simply waiting for OpenSees to load when it is invoked. Because users will often perform multiple runs in the same session, it makes sense to only pay the invocation cost once.

I could call OpenSees asynchronously, but I have not figured out the best way to determine when the script has completed... I could put some sort of print statement at the end of the input script and then spin until OpenSees prints it... but that seems rather on the hackish side.

It makes sense that there should be some facility for remote invocation, because of the networking aspects that are being (have been?) added...

Thanks,
Mark

[mphair]

Is my question above unclear? Is what I am trying to do understandable?

Thanks,
Mark