As mentioned in another thread I am working on a game, and I want to add an optional audio interface. The game is a card game, two-player Canasta human vs. AI to be exact. Canasta has an interesting feature, card suits are not important (except for threes). This makes it possible to link any card in Canasta to a single letter or digit, so you can actually implement a text-based canasta interface without problems.

The text-based UI is already accessible because you could use a screenreader. However, the game play experience would not be very smooth I think. Basically you would need to tell the screenreader to read line X of the screen if you want to know the cards in your hand, line Y if you want to know the cards on the table etc. And then you would get output like: Hand: JJJAA

My current plan for the audio interface is to use SAPI (i.e. speech synthesis) to

a.) read the relevant parts of the screen on demand
b.) produce nicer output

For example pressing 'h' (for 'read hand') would produce output like:
Three jacks, two aces
i.e. you have three jacks and two aces in your hand.

I could also use SAPI to inform the player about the actions of his opponent. For example:
Che discards a king.

What do you think, would such an UI work? Is it worth the effort?