Conflicting command words?

How is Jibo going to handle conflicting activation rules?

Example skill one:

Jibo, play my music presentation.

Where in the skill, play is used with presentation title, “my music presentation”.

Example skill two:

Jibo, play my music

Where in the skill play is used with music to turn on, and “my music” plays that person’s music.

How does JIBO pick which skill to run?

Do we need to reserve words?

3 Likes

Hi Al,

This will in large part be determined by Jibo’s underlying processes for triggering and switching between third party skills.

As mentioned in this earlier post our design team is currently perfecting the construct that handles that and we will provide further detail as soon as it is ready!

I have also categorized this post in the Speech category of our SDK forum so that other developers with this similar question can most easily find it!

Thank you,
-John

1 Like

I ask this for the Voice UX design process considerations. It is conceivable a change could cause a complete rewrite of a skill, given the type of changes to this moving forward. Are there a list of already determined activation words that Jibo will use exclusively? I want to be sure we cover as many cases as possible where we can avoid most probable changes in the future. Also, our skills are running in a sandbox inside Jibo(correct?), but is there a case where Jibo will interrupt and override our skin, while in action, and then return back? If so, what types of cases could cause this?

2 Likes

Hi Al,

There are times when a Jibo skill would get interrupted. We will have further specifics about that closer to when we release details around the system for triggering/switching third party skills.

Also, I can confirm that we will have reserved global words. We are working on making the current list of terms known to developers. When we have those available we will update this thread.

Thank you for your patience on that!
-John

2 Likes

Thank you John,
This is a very important element to understand before begging any real development, so the information is MUCH appreciated! :slight_smile:

This is rather an important question for development of interrupted dialogue that deals with the understanding the user was trying to perform some action before the interruption. So, does the interruption return back to the skill? In either case, a state of interrupted or not must be developed in my skill, so there is an appropriate verbal response. Regardless if the processing is returned where the skill left off, or the skill is restarted; or the skill must be restarted. All create variants needing Voice UX considerations.

I have a Setup routine that checks for the status of the file we store to hold names.

From what I understand, we would all need this and it should include an interruption routine to handle picking up where the interruption left off. Does this make sense?

Should we all be checking for a previous interrupted state in our skills?

1 Like

Hey Al,

We understand this information is important to be able to build fully functioning skills. As John said above, we will have further specifics about this when we release details around the system for triggering/switching third party skills. We will definitely post details when they are available. The skills will work within a sandbox and I would recommend focusing on developing your skills to run on their own without interruption in the meantime.

Thanks for your patience!
-Joe

3 Likes

Thanks @joe.t Joe, that’s very helpful. :smiley:

I’ll mark this as solved when you post the list of reserved words. Is that okey?