Go Go Internet Braintrust

[gfish]

On the long drive home Monday, I had plenty of time to think. I happened to come up with a very interesting idea for an extension to the PhraseServer element of my media infrastructure.

The question I pose to you is: What kind of things would you want to do using the voice interface for a smarthome? Ask when the next bus is, set the thermostat, what? Specifically, I'm trying to get a feel for what kind of content it would be conveying, and what the range of structures of the generated phrases would be like.

Comments

[shadowblue.livejournal.com]

thermostat, "what's the weather like?" (now / rest of today / whatever I might find relevant -- like warnings, or "oh, and by the way, Mount St Helens erupted while you were sleeping. Thought you might like to know." that could be expanded to any supermajor news, but I am not sure how to differentiate between major news and supermajor news.)

I don't think I'd use voice for lights very often. Or, even, security (locks), if I had some better way of doing it. (handprint? I know biometrics are iffy, but I'd give it a try.)

Traffic reports. Bus routing. Voicemail/e-mail (for the latter, a "(x) e-mailed you" is probably quite sufficient). Set/activate/deactivate timers & alarms, e.g., "wake me up in two hours", "remind me to go to (blah blah) tomorrow".

Only in future revisions would I ever dream of being able to say "make me a cup of tea and some eggs". Although when I think about it, those aren't incredibly complex tasks. Perhaps we should work on a kitchenbot.