Re: Capabilities & Schemas

On 16 May 2018 at 04:26, Michael Koster <michael.koster@smartthings.com>
wrote:

>
> The idea is that for a complete semantic system, we need to account for
> both capabilities of devices (door lock) and the Feature of Interest they
> are bound to (front entrance door of my house). The idea of integrating FoI
> into iotschema is to create relation types that enable the capabilities of
> a device to be semantically bound to a real-world entity.
>

Interesting. I can certainly see the use case for an IoT system to know
that a door lock, a door sensor and a door opener are all devices which
belong to an entity known as "front door". My question would be whether
this metadata should be in the Thing Description or not?

One question is whether the door is one thing, or multiple things? This is
a topic which has come up a few times on our team. For example, is a smart
power strip with four power outlets a single thing, or a collection of four
power outlet things? We've discussed the possibility of a Thing Description
having a "things" member or things link relation which refers to a
collection of sub-things. Another example is a Thing Description for a Web
of Things gateway <https://github.com/mozilla-iot/wot/issues/91> which is
itself a thing with properties, actions and events, but also has a
collection of sub-things which it manages.

Another question is how much user-defined metadata should be contained in a
Thing Description vs. stored alongside the Thing Description in the
database of a things directory/gateway? Perhaps the user wants to provide
their own name for a thing. What about a location of the thing? A room? A
group? An inventory number? The owner's name? The date it was installed?
The date of its next safety check? How much of this is in the scope of the
Thing Description vs. additional metadata a particular thing directory or
gateway may store and use?

Ben