Hi, I'm glad you brought this up. I've been wrestling with a Speakup related UTF8 issue for some time. Fortunately, there has been progress. 1.) The old situation where screen review would suddenly only see null chars appears resolved. I have been defaulting to LANG=en_US.UTF-8 on my Fedora 9 laptop without problems. 2.) However, I can still trigger that problem with null chars by suspending and then resuming. Since that pretty much destroys the usability of Speakup for me, I spent my entire week at the recent Linux Foundation Collaboration Summit working exclusively with Orca which suspended and resumed without incident. To the other question of mixed language pronunciation--I've not noted English Speakup being able to correctly pronounce foreign language chars. Would be nice, and may be possible, I suppose. In rich text environments (like web or OpenOffice), however, we should expect our synthesizer synthesizers to switch pronunciation on the fly, if content is correctly coded with <lang=> tags. This may require loading more than one language, but that could be indicated by configurations. I believe there has been movement in this direction, but I expect it's mostly evident where rich text is properly parsed, which would mainly mean Orca. Janina Zachary Kline writes: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Hi, > I've been curious for a while now about Unicode, UTF-8, and related > issues. I imagine that whether or not Speakup works with Unicode > characters would depend in large part on the speech synthesizer used. > Add to that, of course, the fact that seeing a character on screen and > pronouncing it are two different things. So, all told, I'm just curious > about this topic in general: what works and what doesn't? I'm not going > to be doing anything foreign language related myself, but do like the > idea of being able to put in an accented letter now and again. > Best, > Zack.
