Content negotiation and service qualifiers

95 views
Skip to first unread message

berrand...@gmail.com

unread,
Mar 4, 2024, 10:23:40 AMMar 4
to ARKs
Hi everyone,

At BnF, we are thinking about serving human-readable pages (example), RDF/XML (example), RDF/N3 and RDF/NT data by using both service qualifiers and content negotiation for authority records identified by an ARK.

Has someone used both content negotiation and service qualifiers? Any advice or feedback on this topic?

Kind regards,

Bertrand Caron

Donny Winston

unread,
Mar 4, 2024, 10:31:36 AMMar 4
to ARKs
What do you mean by "service qualifier"? The qualifier part of an ARK (https://www.ietf.org/id/draft-kunze-ark-38.html#section-2.5)?
--
You received this message because you are subscribed to the Google Groups ARKs group. To post to this group, send email to arks-...@googlegroups.com. To unsubscribe from this group, send email to arks-forum+...@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/arks-forum?hl=en
---
You received this message because you are subscribed to the Google Groups "ARKs" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arks-forum+...@googlegroups.com.

--
Donny Winston, PhD (he/him/his)
Polyneme LLC
New York, NY

berrand...@gmail.com

unread,
Mar 4, 2024, 11:55:11 AMMar 4
to ARKs
Hi Donny,

By "service qualifiers", I mean "qualifiers that reveal object variants" (https://www.ietf.org/id/draft-kunze-ark-38.html#name-arks-that-reveal-object-var).

Donny Winston

unread,
Mar 4, 2024, 12:32:04 PMMar 4
to ARKs
Cool, thanks for clarifying. I am yet to systematically use both, and I too am interested in this topic.

Dave Vieglais

unread,
Mar 4, 2024, 1:49:37 PMMar 4
to ARKs

Hi Bertrand,

I’m curious how a conflict between resource type requested via content negotiation vs. object variant in the ark might be reconciled. Which would take precedence, the provided Accept header or the variant hint?

Another issue may be the lack of a well-defined vocabulary for variants when expressed in an ark versus IANA registered media types. Perhaps some review of similar systems might lead to some consistency for variant labels.

One approach to assist technical users may be to advertise the availability of alternate representations of the identified resource by providing a link header in the server response. For example, a request to https://data.bnf.fr/ark:/12148/cb11893627z may return as part of the response a link header (formatted for readability):

link:
  <https://data.bnf.fr/fr/11893627/nicolas_bouvier/rdf.xml>;
    rel=“alternate”;
    type=“application/rdf+xml”,
  <https://data.bnf.fr/fr/11893627/nicolas_bouvier/rdf.nt>;
    rel=“alternate”;
    type=“application/n-triples”

or perhaps:

link:
  <https://data.bnf.fr/ark:/12148/cb11893627z/rdf.xml>;
    rel=“alternate”;
    type=“application/rdf+xml”,
  <https://data.bnf.fr/ark:/12148/cb11893627z/rdf.nt>;
    rel=“alternate”;
    type=“application/n-triples”

Then it is clear to a consumer application the association between resource links and the type of resource available. With a little javascript, a browser client could generate the alternate links in the UI by issuing a HEAD request against the target and examining the link header values in the response.

Another mechanism is agent-driven negotiation but that process is not so well defined.

regards,
Dave Vieglais

Reply all
Reply to author
Forward
0 new messages