Contributing new intents? #212
petercwallis
started this conversation in
_General
Replies: 3 comments 2 replies
-
Right now we want to limit to basic interactions because we're still working on building out the very basic foundation of the platform. We might explore such features, but I don't see that happening in 2023. If people are interested in this, Home Assistant's APIs are open and our data sets available, so people can pursue that. |
Beta Was this translation helpful? Give feedback.
2 replies
-
Yep. So adding a row is easy; and what needs a process is for us to add a
column. That is tricky.
…On Thu, 5 Jan 2023, 09:02 Filip Krane Adamsen, ***@***.***> wrote:
Hi @petercwallis <https://github.com/petercwallis>. Regarding the 2D
matrix, is this <https://home-assistant.github.io/intents/> along the
lines of what you are looking for?
—
Reply to this email directly, view it on GitHub
<#212 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ALY3LWMFFN7XFEP63AI6BG3WQXXQ3ANCNFSM6AAAAAATNJBE5Y>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
Beta Was this translation helpful? Give feedback.
0 replies
-
Marathi language leader? |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi all. I am excited by the possibilities of HomeAssistant's year of voice. I was Principle Investigator on a EU project that put "talking rabbits" (nabaztag) in older peoples homes to encourage exercise. We had PIR sensors and knew where the house keys were. This was back in 2009-2012 so before the Echo and we had major issues with the speech recognition I must say. Things have moved on in the last 10 years but the "understanding" side of voice interfaces does appear to be stuck. I can help.. I hope.
The intents for home assistant are very much about "control" of the home, which is fine when the user a) knows what he/she is doing, and b) does only good things. An alternative is that a voice interface is, and acts as, a social actor, and I think this would be interesting to many in the community. To this end I was wondering how I should go about suggesting two more (sets of) "intents".
The first is based on being helpful. The gist is that "turn on the bedroom light" should not simply say that "bedroom light" was not found, but should say what it can turn on, and do that not as a long list, but as part of an ongoing dialog. For instance listing the 3 most popular lights. An "intent" HassListOptions (say) would have a field specifying the HassTurnOn intent, and the action would be complete when the user turn(ed)on something or cancelled somehow.
A second is based around authority. For cloud-based solutions identifying individuals is problematic but for a local service the technology is the limit. Assuming the audio signal processing automatically assigns a probable id to a voice, Questions such as "Are you allowed to open the garrage door?" make sense. Any child worth his or her salt will say "Yes", but the system is then justified in asking a "favourite colour" quesiton or indeed ask for a password. The intent might be HassCheckAuth or something.
The big picture is that language in use is actually an action in a social space, and hence could do with an "intent" to make that explicit. In 2023 I intend to set up home Assistant with a voice interface so that I can play with the voice dimension. I am new to git hub and am not confident about how this whole community process works. I get the impression I should fork the project, and work with that. Please let me know if that is not right - I imagine I need to lodge pull requests in incremental bits but am not sure how that would work.
Beta Was this translation helpful? Give feedback.
All reactions