Summary from joint W3C/OCF call on 29 April 2016

Last week I organised a joint call between people from OCF and W3C with a view for starting a collaboration on joint demos that explore the integration of OCF based devices as part of the Web of Things. The call went well and we are now planning on next steps, as well as opportunities for exchanging ideas over a longer period.

The participants were: Ravi Subramaniam (Intel), Markus Jung (Samsung), Zoltan Kiss (Intel), Michael Koster (SmartThings), Johannes Hund (Siemens), Matthias Kovatsch (Siemens), Louay Bassbouss (Fraunhofer FOKUS), and Dave Raggett (W3C staff).

My notes from the call:
OCF is adding an introspection capability. Today the behaviour is specified using RAML and JSON Schema. There is a collection of these on http://www.oneiota.org/ <http://www.oneiota.org/>
These take the form of templates which can be specialised to classes and further specialised to instances. This aligns well with our discussions on the Web of Things, where we focus on instances of things and want to refer to generalised definitions and to be able to model collections of things.
There isn’t currently a direct way to get a link to the RAML file from a device, but this is under discussion.
There is an opportunity for adding annotations, e.g. semantic annotations. This would be useful for say declaring how to map a resource to a property, event or stream.
UPnP have decided to adopt the JSON based framework of OCF
OCF current supports CoAP, but is working on other protocols, e.g. HTTP, Bluetooth Smart (GATT), and integration with OMG’s DDS
OCF defines a repository for information about specific devices. This could be a handy way to discover the devices for a demo. Another way would be to use multicast based discovery.
Ravi is interested in a medium to long term discussion on the ideas we are considering for the Web of Things, e.g. semantic constraints, data model integrity constraints, and richer data interfaces.
Markus is keen to collaborate on joint demos and I’ve offered to help him with justifying his involvement in the Beijing plugfest in July.
Dave would like to work on extending his WoT gateway to expose OCF resources as things.
The RAML files can be used to generate skeletal code for developers to flesh out.
It will be interesting to study how to map RAML/JSON schema to a thing description. This may include the addition of information where there is a choice, e.g. between a property update or an event.
It doesn’t look like we can easily map to things and streams as first class types without a lot more work, however, this might be something for OCF to take on board, and as such a discussion area for a longer term collaboration.
Dave will arrange a follow up call to talk about plans for joint demos at the Beijing plugfest. This means agreeing a roadmap for the preparations leading up to Beijing.
Dave will work with the IG on framing the goals for a medium to long term collaboration with OCF.
Ravi notes that OCF is already talking to other IoT orgs/SDOs and we could pool resources on this.
Ravi adds W3C has experience in semantic work and so having W3C describe a meta-schema/language that can be use to capture semantic info in a light weight way then OCF can use that to enhance the current work on data models to include semantics.
OCF will continue to do data models (it is a big part of the charter and how OCF is organized) in addition to abstraction and transport but it would be great to see if there is a light weight method to apply semantics - I have an approach in mind but maybe W3C Things description or something else that WOT is working on, may be better (I have not studied TD yet).
—
   Dave Raggett <dsr@w3.org <mailto:dsr@w3.org>>