Re: What to do with WoT Profiles 1.0

On Wed, 5 Jun 2024 at 10:15, Kaebisch, Sebastian <
sebastian.kaebisch@siemens.com> wrote:

> “A WoT Thing Description MUST accurately describe the data returned and
> accepted by each interaction”
>
> The TDs examples definitely does not match the answer behavior as shown Example
> 15 <https://w3c.github.io/wot-profile/#example-15>.
>

Please note that PR #408 <https://github.com/w3c/wot-profile/pull/408> made
the synchronous member of ActionAffordance mandatory in the HTTP Basic
Profile, which means the output in example 15 would only be ever returned
if "synchronous" was set to false.

In the case of an asynchronous action, a Consumer should not have any
expectations about the response to an invokeaction request in HTTP, since
as far as I can tell there is no default specified anywhere. There is
actually a known bug
<https://github.com/eclipse-thingweb/node-wot/issues/1278> in node-wot
regarding this behaviour which acknowledges this to be the case.

Since the response to an asynchronous invokeaction request is unspecified
in the HTTP Binding Template, I would argue this is in fact entirely
compliant with TD 1.1. If a Consumer expects a particular response then
that's a bug in the implementation.

Today, I will reserve some time in the main call (~20min) for this kind of
> discussion. However, I need to ask the group if this is ok.


I'm afraid 20 minutes will not be enough to time to address these
misunderstandings, but I would be grateful for any opportunity to discuss
this topic and clarify Siemens' position in order to unblock the Profiles
work.

Thank you

Ben