I was looking at this just the other day. but the other way round. Voice control for the devices. not impossible.
and yup. the idea is that a properly doctored SDL should support ALL of the devices (maybe the DX will run like crap) thus giving us a much better port base.
currently we have: oh yeah, an ancient QT kludge - now a full version behind, Direct to FB code, GTK on the devices that support it, overally it's a bit of a full-on DIY mish-mash. I probably missed half a dozen partially supported projects / widgetsets / implementations.
to just have "something" to throw ports at would be nice. SDL seems a decent bet as the runtime ENVVAR stuff makes it very flexible for our purposes, nice to just have to chuck a few ENV's around and have the next configuration all setup.
Anyways, it's a time sink but meh. I have TSLIB working now, I have SDL building lovely, working on mastering the awesome/awful layers, just a wee bit more graphics code and a spot of Onscreen UI overlay work (maybe) and then we can think about such niceties as the spoken word.
Gorf would be a nice port for that
"Insert Coin", those words still bring a chill down my spine (Did that machine just speak?)
http://en.wikipedia.org/wiki/Gorf