search

rit-sse-mycroft / app-projector

0 stars 2 forks source link

Voice command to turn on projector #2

Open ZMYaro opened 10 years ago

ZMYaro commented 10 years ago

The projector should listen for a command, such as “Main screen turn on.” I (@zmyaro) can write the grammar XML if you want.

deriachai commented 10 years ago

I disagree, That is not the responsibility of the projector app,

The projector is responsible for displaying given media. Any service requesting projector jobs will cause it to be turned on, The user should not have to give a projector on command.

derekerdmann commented 10 years ago

What about when we manually hook up to the projector? Shouldn't we be able to instruct the projector to be turned on even though some other machine is the input source?

deriachai commented 10 years ago

I see your point, but still disagree that this is the projectors responsibility. It's scope is solely to control the projector, and to display output, not listen for input and react upon it.

derekerdmann commented 10 years ago

At a minimum, there needs to be a query that you can send to the projector that turns it on. As for the voice commands, whose responsibility should it be? The Speech-To-Text service would be doing the listening, and notify the projector when it hears that input.

deriachai commented 10 years ago

The speech-to-text actually doing the grammer parsing too? I don't know whose responsibility it would be, but i feel it is outside the scope.

The projector should be able to recieve a message to turn on, i feel it is not in scope to actually do text processing to determine if the user has requested it.

derekerdmann commented 10 years ago

The projector app defines the grammar that it expects to hear, in XML form. It gets sent to the speech-to-text app, which parses the grammar, then notifies the projector app when it gets a match.

The only thing the projector app needs to do is define the commands it can handle, then handle them when they're triggered. It doesn't parse the grammar, listen for audio, or anything else.

deriachai commented 10 years ago

In which case I was misunderstanding the way that worked, I was understanding that as it having to parse the grammer.

Also, optimally we should not have to plug anything manually, there should be functionality for anything we want to stream.

derekerdmann commented 10 years ago

People still like having their own control over the presenter view, and may not want to add the processing load to their machine to handle the streaming. This is something we should be supporting.