AW: Seeking WebSocket proposal team

Dear all

> > I'm talking about the actual data, not metadata (e.g. continuous temperature
> measurement). JSON-LD (triples/quads) must be used, otherwise, the schema
> information is lost.
> 
> I totally agree on that point, and on that we could collaborate with JSON-LD /
> RDF-JS groups.

We should keep in mind that the main goal is to counter fragmentation in the IoT space. The WoT group acknowledged that there will be multiple IoT platforms and our approach is to complement and describe those platforms, i.e., attach additional information. The goal is not to become yet another standard that prescribes who exactly people must do things. Dictating how data needs to be sent is going into this direction.

> > But if you don't want to use it for streaming, what would be the benefit of
> WebSocket at all?

This is exactly my question! People keep mentioning WebSockets in the sense of a proper application protocol (or even put them in the same bucket as REST), yet it is simply a stream-based transport (yes, with framing) that happens to be accessible from the Web browser. Everyone has been designing their own protocol very specific to their own application. I don't see how this helps with interoperability. Thus, I want to learn more about the approaches in mind.

> A few years ago, we used WebSockets to control a robot through ROS and
> Rosbridge from a Web browser. I think WebSocket can be convenient for Web
> standard-compliant pub/sub across firewalls.

Firewall traversal is one of the bizarre things where WebSockets have a practical benefit. What you are describing would mean that you have a standard-compliant and globally accepted pub/sub protocol that you speak over WebSockets. That will then be the actual application protocol to which we need to map.

BTW: There is also CoAP-over-WebSockets in the making.

Best regards
Matthias

Received on Monday, 24 October 2016 00:07:35 UTC