----- Original Message -----
From: "Irmen de Jong" <irmen.nos...@xs4all.nl> Eric, if you're concerned about performance, Pyro4 (the source distribution) comes with several examples that do simple performance related tests. You could run these and see what figures you get on your setup to see if it's anywhere acceptable, before even building anything with Pyro yourself. sounds like a plan. I managed to get around the initial problem I had because apparently the natlink extension for NaturallySpeaking is semi-persistent. Sometimes the statically created RPC connection would work, other times it wouldn't. The current workaround is to place the RPC connection initialization in the code that activates the grammar. <blockquote> Interesting project btw. </blockquote> it has been a long time since a project like this has made me smile ear-to-ear. For example, yesterday I had to create twelve open VPN configuration files and key pairs. With broken hands like mine, it would be an extremely painful hour to two hours to entering the data over and over again even with easy RSA. I was able to complete this task in about 10 to 15 minutes. Now that's what accessibility is all about. I believe my experiment shows that a two dimensional grid with names for both rows and columns can allow a speech recognition dependent user much faster data entry than one could have with straight speech recognition. With some enhancements, it should be possible to use this technique to remember something on the fly. ideally I'd like to take a tool like treesheets , put some Python power underneath the grid, and explore how a grid tool can help accessibility but I'd need a volunteer to make that happen.
-- https://mail.python.org/mailman/listinfo/python-list
