- From: Danny Ayers <danny.ayers@gmail.com>
- Date: Sat, 2 Apr 2011 12:17:44 +0200
- To: "Edward O'Connor" <eoconnor@apple.com>
- Cc: HTML WG <public-html@w3.org>, tantek@cs.stanford.edu
Both yourself and Tantek have made the point that the bar to creating a new rel value is low with the microformats-wiki-as-registry approach, yet I note on the registry page [1] : [[ If you wish to develop a new rel value, please follow the microformats process. ]] The microformats process [2] is long (creating 9 pages, moving through 3 stages) and geared towards creating a whole new vocabulary, there seems a whole lot of baggage here that doesn't seem appropriate for the creation of individual terms. Personally I still prefer the RFC 5988-based proposal, but in case the microformats community is to become the registrar, I strongly suggest that a dedicated process is created for rel terms, and the way they are documented be restructured. There is an unavoidable mismatch between the requirements for microformats and the requirements for the rel registry. In principle there are a vast number of conceptual relations that could potentially apply between any two resources (and the door to these should be kept open in the use of URIs for rel values). Most vocabularies defined for Web semantics are compiled according to their domain. But here the domain is wide open : "document" - and we are looking at terms that may be used in a specific syntactic construct. It's probably leading out of scope here, but whatever the decision on registry, I hope some work is done to sensibly categorise a vocabulary that includes terms such as "stylesheet" and "sweetheart". Cheers, Danny. [1] http://microformats.org/wiki/rel-registry [2] http://microformats.org/wiki/process -- http://danny.ayers.name
Received on Saturday, 2 April 2011 10:18:17 UTC