Hi Jon,

Regarding the 'returning only relevant part' vs. 'returning whole

It is not 'wrong' to include more than just what you asked for. I don't
know about any guideline that says how much extra should/could be
included. For example, if you ask for a description of me, I might want to
give you next to my name, and workplace URI, a summary of the organisation
I work for plus a link to a more complete description of my workplace.

However, since we're (well, you are) identifying things (the somewhat
abstract concepts of classes and properties), it does make sense to use
different URIs for the Thing and a Description of the Thing. Two
approaches for this are using hash URIs and using 303-redirects to
descriptions. Returning an HTML document (or XML document as I get) in
response to a request for an RDA property or class is wrong in the Linked
Data sense [note 1]. This is explained in the W3C WG Note that you
referred to in recipe 2 [2].

Are you planning on introducing 303-redirects?

Groeten van Ben

[note 1] I am aware of the continuing discussion about alternative ways of
linking a Thing and description(s) of the Thing, especially for Things
that cannot be retrieved over the internet. And you're not alone: DOI URIs
and ORCID URIs do not return objects (usually publications) or
researchers/contributors but descriptions of the Things the DOI and ORCID
claim to identify (e.g. a tweet to Rob Sanderson + following discussions


On 23-01-14 00:26, "Jon Phipps" <[log in to unmask]> wrote:

>'slash' vs. 'hash' URIs:
>As a matter of design, we coin URIs for retrieval of information about the
>resource identified by the URI by machines, not humans. The most current
>formal rules[1] state that retrieving a 'slash' fragment should return
>that fragment when resolved. We're currently breaking that rule by always
>returning the entire vocabulary, as if it was indeed using hash URIs and
>will fix it in the next few weeks. An example of such a fragment
>by the Open Metadata Registry for
>is here: