I’ve been working on an agent framework for some agent research I’m doing.  It’s taken all of my time to understand and implement the agents and their messages.  As soon as I finish with it, I’ll start the implementation of PyTables.  I have finished the investigation into how PyTables would integrate into Picalo, and it should go off without a hitch.  I won’t know fully until I get into it.  But assuming it works (and I have every indication that it will), it will be totally seamless to existing scripts, including the entire Picalo application.

What is PyTables? It’s a Python library for holding data. It’s built upon the widely-used numarray framework.  Both projects (PyTables and numarray) are written in C, so they’ll speed up Picalo’s data structures considerably.  PyTables includes a disk-caching system, so Picalo should be able to load any size table without killing memory or the processor.  Overall, this integration should bring Picalo into the “full production” arena.

I’m excited about doing it and will jump on the implementation of it ASAP.