Hey Deanne and Jeff; Since you guys are planning on doing both fronts in python, what I was thinking about doing was taking the front to middle XML protocol and extracting it into python function calls, so that a lot of the XML generating code could be reusable in both of your front ends. For instance, instead of having to generate XML like this to connect two nodes: <front> <connect> <workspace id = 'id for the workspace containing the loci to connect.'/> <input id = 'workspace1/workspace2/container1.xml' connector_id = '1'/> <output id = 'workspace1/workspace2/viewer2.xml' connector_id = '1'/> </connect> </front> Instead, I could just add a function call like: connect_loci(self, workspace_id, input_id, in_connector_id, output_id, out_connector_id), and you could just call this instead, and all of the xml building could happen inside of this. I think this would make more efficient re-use of code, and could help pave the way towards a corbafied front to middle communication protocol. What do you guys think about this? If you like it, I can put it together on my local copy and make a new tarball of it, so you can work off of this (since we don't have cvs ready for the new directory structure yet). Brad