The AB may very well be horrible for general purpose use. Honestly, I rarely use the AB, but use minibrowser, voice control and handheld URC MX-900 remotes that are custom programmed where hard and soft buttons send RS232 strings to Premise to control lighting, A/V, HVAC, etc...
Of course, most other things are automated as I use a fully two-way Elk M1G security module for occupancy tracking, displaying weather updates on alarm keypads, two-way RF keyfobs/dongles, etc...
What would be handy is a WP app that passes a voice control string via a url (in addition to having the features your app has). I've already built a Premise module that generically parses such a voice command based on a sentence containing a location name, object name or type, property name and/or value.
For example, the url may look like this:
http://192.168.1.2/sys/%7B340C8A31-D762-416C-933D-C722A8090093%7D?d??mbNaturalLanguage(turn_on_entry_sconce)
Currently, the module also returns an http response that can be read back using tts. The response for the example above would be html text such as: "entry sconce is now on" or "entry scone was already on" and the phone would read this back. However, with subscriptions working, I can envision the app reading back status updates automatically. These updates could be reminders to close the front door, mailbox open alerts, etc...
You'd be very surprised at how accurate the speech parsing is. It's much easier to tell Premise to set a light to 42% than to use a slider or buttons after browsing to the room and clicking on the light. Same goes for setting the thermostat too.
I can send you the Premise module portion if you're interested in including it in your setup (PM your email). It's built to be generic, but you can also use regex strings under each home object if you want to customize it. Of course, your app would have to use some sort of api for the speech-to-text and text-to-speech. I'm sure windows phone has something available...
Charlie Kindel said:
Right now the app is custom, as is my original WP app. It does SOME dynamic loading (e.g. it queries the server for a flat list of lighting scenes).
I don't believe in automatically generated UI. I think the AutomationBrowser is fine for rudamentary interaction, but I think it is horrible for general purpose use. I have 5 or 6 specific scenarios that me & my family care about and I want the UI for those to be perfect.
That said, there's no reason why my framework couldn't be used to dynamically build a generic UI; I just am not working on that.