Hello, I'm trying to look more into the problems I'm having with the speech support on Dapper (see my mail with subject "Fixing brittle speech support on Dapper" from June, 10th, which strangely isn't showing up in the archives at http://mail.gnome.org/archives/gnome-accessibility-list/2006-June/thread.html). I started reading through gnome-speech source code. I noticed that it runs festival as a server and talks to it. This makes us have a screen reader that talks CORBA to a server that talks TCP/IP to another server who then does the synthesis. That's too many passages in which something can go wrong. Instinctively, I'm considering rewriting the festival driver to using the C API rather than the festival server. The C API of Festival is just as simple as this: http://rafb.net/paste/results/I7trk068.html I'll look into it a bit more, writing some test code to talk to the CORBA festival driver as well as test code for the festival C/C++ API, so that I can gain familiarity with both things. Any reasons why this hasn't been done yet? In the meantime, however, before I hurt my brain too much with this, what's the overall situation? Is it worth the effort of fixing gnome-speech, or is the effort better spend on making something else work? I'm already quite frustrated of not getting any sort of answer on the list for this problem that is getting me totally stuck (and thank Luke Yelavich for mora support on IRC), and I don't know how I would cope if I spent time and effort on this just to hear as soon as I've finished that everyone's moving to speech-dispatcher or some other kind of a totally different technology. If it's not worth spending efforts on gnome-speech, please let me know what I can use to replace it, since it doesn't work for me. Ciao, Enrico -- GPG key: 1024D/797EBFAB 2000-12-05 Enrico Zini <enrico debian org>
Attachment:
signature.asc
Description: Digital signature