Re: document node attributes

Mmm.

But we do allow the convention that if there is a link then
there MUST be something on a web server to which that
link resolves; i.e. broken links are Bad. It seems a bit like
that to me to say that there is an implicit link between a
resource of a certain media type and another resource
algorithmically derived from the URL of that resource. I.e.
that assigning a certain media type 'application/fooxml' to
a resource is itself making a link to another, specified
resource (e.g. having the syntax and semantics of atom's
 link model) wose URL is to be algorithmically derived from
the first resource's URL. If there is no resource at the
expected URL or its syntax is not as specified then it is
effectively a broken link and the web server can return the
appropriate code or the calling app (if the syntax is not as
specified) can treat it as if it were a broken link.

----
Stephen D Green


On 3 July 2013 15:46, Rushforth, Peter <Peter.Rushforth@nrcan-rncan.gc.ca>wrote:

> **
> I think it is impossible for a media type to dictate what a server has to
> serve.  It can
> only describe its own semantics.
>
>  So there could be a media type -
>
>
> 'application/*fooxml'*
>
> that required that for
> every URL containing a document of its media type,
>
> http://abc.xyz/foo/bar/123
>
> there MUST be a corresponding document containing its links
> at
>
> http://abc.xyz/foo/bar/123/ <http://abc.xyz/foo/bar/123/links>*fooxml*
>  and, say, state that the links MUST be published using a particular
> XML syntax with particular semantics to state which attributes,
> say, are for the actual links.
>
>
>
>
>