How to Use a Combadge

Interacting with a Combadge is similar to any other voice agent, such as those offered on computers, phones and Palantiri. You can ask the Combadge questions to acquire information, or you can give it commands.

Shining a lamp on the Genie

The OEM voice agent is officially referred to as the “Genie”, though in practice it gives it’s name as the OEM’s brand. This agent only has a few informational requests -simple concepts like the date and time, the locations of logged-in badges and the membership of groups. While these are useful they fall far short of the functionalities of the voice agent on most Star Trek shows, which was able to access complex and contextual information sources. This is more comparable to consumer voice agents that ship with computers, phones and “home assistants”.

Similarly, the OEM agent has few commands that do not directly relate to sending and receiving communications. Modern consumer voice agents are far better at replicating (pun intended) Star Trek’s varied computer commands. The most well known example from a home assistant is the ability to stream music, usually from the assistant manufacturer’s commercial music subscription. Another good example of fiction becoming fact is the integration of home automation into home assistants – we now have the ability to control lights, window blinds, and physical access controls entirely through our voices.

The latest additions to these agents is the ability to order food and drink from the voice agent through several different technologies. The first is through integration with the public API’s of take-out chains, with Domino’s Pizza being a notable example. The other, “more Star Trek” innovation is through connected coffee makers – high end models of which offer fully automatic “bean to cup” dispensing by voice command, and are designed to be fully integrated into the fittings of the home. This is as close as we’ve come to a fully-automated home food replicator to date.

What does this Combadge Agent do?

The intent for Spin Doctor is to replicate something much closer to consumer voice agents than to the OEM “Genie”, as it’s intended that these devices will be repurposed for home users. The target “baseline” for the Spin Doctor agent’s command capabilities is to handle basic calling at least as well as the OEM Genie, to be able to manage multimedia playback from a source (e.g. a DNLA server) to one or more endpoints, and to integrate with a smart home – most importantly: lighting, physical access control and at least one device capable of providing “Coffee, black” on request.

Similarly, the informational system should be closer to a consumer agent than the OEM Genie. While it’s intended that location querying will be implemented, in practice most home users will only have a single AP in their environment – making this use case impractical. More useful is the ability to query PIM information (Computer, today’s appointment calendar. Computer, what’s my average step count for the last week?) as well as accessing information of public record such as weather and transport information – as well as providing information on the state of the devices it’s commanding. (Computer, what track is playing right now? Computer, is the front door open?)

How will Spin Doctor achieve this?

There’s no expectation that this application will ever reach the levels of market penetration where commercial bodies will be producing coffee makers specifically designed to integrate with it. What would be great to achieve is integration with well-established protocols for multimedia, PIM (especially contacts and calendars) and smart lighting integration.

From there, the best approach would be to pick a small, accessible, well established subset of smart devices (such as a coffee maker and a smart display) and use them as a model to create a generalized model for interacting with them as both an information source and a command sink.