Talking Data?

[John Graves]

Working with text-to-speech technologies potentially opens up new communications channels.

 

For example, here is a short promotional slide show with voice overs generated with the open source Wiki-to-Speech system (http://wikitospeech.org):

http://dl.dropbox.com/u/12838403/20110901/mixandmash.htm

 

What data has relatively simple keys which could fetch useful verbal responses?

 

In Auckland, we have these signs at bus stops which identify the bus stop by number. If you send that number in an SMS message, you get back the expected arrival times of buses at that stop. There is an Android app and a web app which do the same thing. Some of the signs are big, electronic notice boards with a yellow button which, when pressed, reads the arrival times aloud. (There was a funny bug in that system recently. They had coded the letter "N" to be read aloud as "Nightrider" but then they introduced these new INNer and OUTer Link buses, so the text-to-speech would say "I Nightrider Nightrider in 2 minutes" ...)  The key thing about these signs: they are always there for bored people to look at while waiting for the bus.

 

So what else could be offered that might make for more interesting/useful listening?

 

I'm guessing that there could be:

- location-based keys which yielded verbal responses with simple information (address? description of attractions nearby?

weather? as in http://www.radionz.co.nz/radionz/programmes/long-range-weather/audio/2497079/long-range-weather-forecast-for-2-september-2011.asx)

- council notices, keyed by some council identifier

- podcasts? http://www.nztri.org/podcast?page=7