Re: [core] Links, hosts, and resource directories

[Klaus Hartke <[e-mail]>]

Christian Amsüss wrote:
> >     <coap://example.com/foo>;rt=a,<coap://example.com/bar>;rt=b,<coap://example.com/baz>;rt=a
> >
> > This means, there are three links [6]:
> >
> >    <coap://example.com/.well-known/core>  hosts
> >           <coap://example.com/foo>  which has type "a" .
>
> Actually it is `<coap://example.com> hosts ...` due to the wording of
> RFC6690 Section 2.1 b).

Good point! For some reason, I was convinced that only the "hosts"
relation type would do weird things to the link context. I missed that
RFC 6690 defines the anchor for *all* links without an anchor
attribute to be the root resource of the origin.

So, for example, the links in this Link Format representation of
<coap://example.org/.well-known/core>...

    <foo>;rel=hosts,<bar>;rel=icon

...are resolved to...

    <coap://example.org/> hosts <coap://example.org/foo> .
    <coap://example.org/> icon <coap://example.org/bar> .

...which, to me, have the following meaning:

    (coap, example.org, 5683) has a resource <coap://example.org/foo> .
    The shortcut icon for <coap://example.org/> - but not immediately
        any other resource hosted by (coap, example.org, 5683) - is
        <coap://example.org/bar> .

RFC 6690 is really interesting.

> I'm not sure that this interpretation of 'hosts' necessarily follows
> from the definition ("hosted by the server indicated by the link
> context").

Isn't that exactly what "hosted by the server indicated by the link
context" means? See also the text in
<https://tools.ietf.org/html/rfc6690#section-5>:

    Thus,
    the links in this example tell that the Origin server from which
    /.well-known/core was requested (the context) hosts the resources
    /sensors/temp and /sensors/light (each a target).

It's probably not worth discussing the exact semantics, though, as
those have little relevance in practice.

Klaus