On 1 Mar 2012, at 19:29, David Booth wrote: > I find the document at http://www.w3.org/2001/tag/doc/uddp-20120229/ to be hard to read, overly broad, unproductively pedantic and unnecessarily entangling philosophical questions with simple matters of protocol design. > > As an alternate proposal, I have started a significantly simplified document on the W3C wiki at http://www.w3.org/wiki/UriDefinitionDiscoveryProtocol > > Others are encouraged to contribute. The goals are: > > - to be simpler, clearer and more direct; > > - to limit the scope to http (and https) URIs, because this is where the LOD community is experiencing the problem that this specification addresses; > > - to specify this as a simple protocol between a URI owner that wishes to provide a URI definition, and an agent that wishes to discover that URI definition; In 2002, the Link: header was nonstandard. Now thanks to RFC5988 we have a Link: header and thanks to POWDER we have "describedby". I recommend this as the official mechanism; use of 303 can be documented as existing practice that implementors should know about. That's what I'd like David's document to become. If “describedby” isn't a good fit, authors can and perhaps should mint a more accurate relation name (using an absolute URI, or in RFC5988 terms an “extension relation type”). -- Tim Bannister – isoma@jellybaby.net
This archive was generated by hypermail 2.4.0 : Tuesday, 5 July 2022 08:45:28 UTC