<scribe> scribe:zkis

Prev minutes

<kaz> May 25

Minutes approved

PR https://github.com/w3c/wot-scripting-api/pull/209

Zoltan: reviewing the PR

McCool: should add editor's note to specify how the spec will be evolving later, wrt Protocol Bindings

Zoltan: question: should readMultiple fail is not possible with a single network request?

McCool: yes

Cristiano: yes

Zoltan: some help will be needed on defining how to handle URI variables

Cristiano: there is an RFC 6570

Zoltan: maybe if there is anything in the Fetch Standard

Cristiano: question about arrayBuffer()
... the function reads all the bytes
... does that include decoding?

Zoltan: same as with the Body mixin, just providing the bytes
... value() now fails early if it can tell it has no chance to succeed
... should we allow implementations to duplicate the stream?

Cristiano: it would be overengineering
... we could leave it this was and change later if needed

Zoltan: on "check data schema", if there is an external algorithm available, let's change
... on writing, if schema is a string and source is not, should we use JSON serialization?

Cristiano: need to think about that

Zoltan: let's create an issue

Cristiano: yes and experiment

Zoltan: OK, this was it. Still waiting on Daniel's input.
... will add the editor's notes

should we extend consume() to accept a URL in simple case?

<kaz> D. Full Web IDL

McCool: make sure it is clear it's a URL and not a string TD
... if it needs security info, the runtime needs to be provisioned

Zoltan: right

Cristiano: currently if we use Fetch we are limiting to HTTP, so we cannot use CoAP
... if we include URL then we could use other bindings
... so this is a welcome idea

McCool: actually packet size limitations might be an issue with CoAP
... on short term it is likely that distributing TD's would be HTTP based

Zoltan: that is great information

McCool: we will study distributing via CoAP

Zoltan: CoAP would work well with constrained devices that don't have HTTP stack but have CoAP support

McCool: doubt we'll do distribution with CoAP, but with CoAPS
... CoAP over TCP could be a solution
... anyway, the Scripting API should be independent

Zoltan: will create an issue or revive old issue to discuss this

[meeting adjourned]