Re: Associating URI-based identities with HTTP requests

[Manu Sporny <[e-mail]>]

On 05/10/2013 07:42 PM, James M Snell wrote:
> Re: when to use a link header, the rule I follow is fairly basic :
> if the value consists of a uri that is optionally typed and
> potentially requires additional metadata,  and if you think the link
> relation will be generally useful in other scenarios (e.g.  Atom
> links,  json documents, html tags), then use a link relation.

The link could be typed, or it could just be a plain 'ol URI. If the
link is typed, it could have multiple different types (since we're
dealing with Linked Data). The purpose of the link is to express an
identity, but the same link can be used for describing ownership, access
control, etc.

> If, on the other hand it is something that only requires a uri with
> no additional metadata, and you don't care about non-http use cases,
> use a new header.

Yeah, I think we're in this scenario.

> Based on an initial review of your note, a new header is likely 
> appropriate.

Thanks - your basic rule was very helpful... would be good if we could
put it in an RFC somewhere. :)

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny, G+: +Manu Sporny)
Founder/CEO - Digital Bazaar, Inc.
blog: Meritora - Web payments commercial launch
http://blog.meritora.com/launch/