Hi Colin, Thx for you last mail on this topic. You gave me interesting thing ideas :-). When I have understood your mail correctly speech-dispatcher should splitted to two componennts: 1. speech-server running as system service 2. speech-dispatcher-agent, which only connects to sd, pull's the audiosamples and plays it to currently available audiosystem. Now your question why the consolescreenreader can't be run in usersession: 1. The sr needs to insert keys using the /dev/input/* devices for the mentioned cursorrouting functionality. This seems not to be possible in userspace. 2. Screenreading accross different textconsoles with multiuseracces and the loginprompt needs access to /dev/tty0 /dev/console /dev/vcs /dev/vcsaX HTH. Halim
