[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Fwd: New Version Notification for draft-nottingham-http-link-header-03




[1] http://tinyurl.com/5vueqt
[1] http://www.opensearch.org/Specifications/OpenSearch/1.1#Autodiscovery_in_HTML.2FXHTML

OpenSearch defines a "search" relation name. Should your document include an entry for "search" in section 6.2 for the initial contents of the Link Relation Type Registry?

- Relation Name:  search
- Description:  Refers to a search description document
- Reference: [2]

[2] "this document" i.e., http://tools.ietf.org/html/draft-nottingham-http-link-header-04

(assuming this idea makes it into rev 04).

Note that I did not say "OpenSearch" description document because I can envision other content types (besides application/opensearchdescription+xml) that may become applicable to the "search" relation name.

One reason for putting "search" in your internet-draft rather than wait to register it after the registry is established is to make it clear that "search" should not be limited to OpenSearch descriptions.

When the <link> element appears in HTML/XHTML in accordance with [1], it also provides a @type attribute. Your document also supports "type" as a link-param.

So perhaps an additional appendix can be written to essentially say what I said above; the appendix can reference [1] as an example of a search description document, but make it clear that the "search" relation name should not be limited to OpenSearch.

This would also provide the benefit of pointing people to the OpenSearch spec, which may encourage people to consider using OpenSearch (a good thing, IMHO).

The name "search" is just too generic to limit it to OpenSearch even though I like OpenSearch and would like to see it used more.

I think opensearch v4 is looking at some other relation names also, like "suggestions" [3].

[3] http://tinyurl.com/6xnbuy

[3] http://groups.google.com/group/opensearch/browse_thread/thread/b92db46be8cdc52d/f00d21f3287b1a91

Perhaps opensearch v4 should say something about autodiscovery using the HTTP Link Header.

Paul



On 2008-11-30 20:11, Mark Nottingham wrote:
This is a fairly substantial rewrite of the spec, based upon the observation that the link header really isn't the central concept here; it's link relations themselves.

...