The microformats wiki has the closest thing to a real-world rel value registry of link relation types (or link types).
We have the most comprehensive documentation of existing-rel-values currently which performs quite well as a rel registry, perhaps even better than any other "official" attempts at rel registries (AKA link relation registry or link relations list) out there.
stable rel values
draft rel values
provisional rel values
The brainstorming section of the existing-rel-values page can be used as a list of ideas / simple proposals for rel values that merit further research and development (and possibly documentation as drafts) per the microformats process
additional rel values
The existing-rel-values page contains additional sections/clusterings of rel values that have been documented in the wild. Some of these may merit brainstorming/provisional registration and further development.
There is also documentation of dropped and rejected rel values for the sake of thoroughness and to hopefully avoid having them be unnecessarily re-proposed (without at least following up on why they failed in the first place.)
new rel values
If you wish to develop a new rel value, please follow the microformats process.
- a page for the rel value itself clearly documenting it as a brainstorm proposal, and explaining it in more detail, including documentation sections per the process (real world end user use cases, examples of content in the wild that would benefit, previous attempts at such a rel value or related format)
- email to the microformats-new mailing-list stating your interest in developing the rel value
- IRC (or email) asking help regarding anything in the process or general rel value questions that would help you to develop your new rel value.
Does the microformats rel registry reflect reality
Q: Does the microformats rel registry reflect reality?
A: In practice yes. The list of existing-rel-values very carefully documents which rel values are in well-accepted standards (formats), which have well written microformats drafts (proposals), and which others are in development, sometimes on other sites (brainstorming).
In addition, the microformats existing-rel-values reflects reality better than any other attempt at a registry, because it also explicitly documents which rel values have failed in the real world, or have otherwise been obsoleted or rejected.
Finally, the barrier to entry for adding to and correcting the existing-rel-values is the lowest of any existing registry and thus the microformats community has best track record of keeping it up to date as compared to other attempts at registries.
Can the microformats rel registry document HTML specific details
Q: Can the microformats rel registry document HTML-specific details?
A: Currently the existing-rel-values are documented in a format-agnostic manner, though in practice most of the real world usage of existing rel values is in HTML.
If there are specific real-world documented needs for capturing HTML-specific details, the existing-rel-values tables could be expanded with a column to describe those HTML-specific details.
Can the microformats rel registry be used for provisional registration
Q: Can the microformats rel registry be used for provisionally registering new rel values?
A: Yes. See the provisional rel values section.
How do provisional values achieve full registration
Q: How do provisionally registered values achieve full registration?
A: The microformats process documents a number of steps to advance any microformats effort, including new rel values. In short, documenting specific real world use cases for a new rel value, and research into existing content that could benefit from it, as well as earlier/other attempts at such concepts help greatly.
Also the microformats process has been updated this year (2011) to provide a more explicit path from draft, to specification, to standard. Those take additional steps such as documenting real world uses of a provisional rel value, implementations, test suites, test results.
Is the microformats rel registry discoverable
Q: Is the microformats rel registry discoverable?
A: Yes. If you google for "rel registry" this page currently comes up on the first page of results, and may of the other results point to this page. Similarly a searching "rel values" returns results including existing-rel-values on the first page of results (in fact, the very first result as of 2011-091 on Google and Yahoo).
How does the microformats rel registry compare to RFC and W3C options
Q: How does the microformats rel registry compare to RFC and W3C options?
A: The W3C doesn't currently have a rel registry. The closest thing to a W3C rel registry is the HTML4.01 Recommendation (which defines several rel values), and the HTML5 Working Draft. The latter is only edited/updated by a single editor, and thus all "registrations" must be committed by one individual. Such "registrations" are simultaneously published by the WHATWG in their "HTML" specification (see http://www.whatwg.org/specs/web-apps/current-work/multipage/links.html#linkTypes) which is available under the MIT license.
Also note that the WHATWG spec presently states:
Extensions to the predefined set of link types may be registered in the Microformats wiki existing-rel-values page.
In comparison, the RFC/IETF/IANA method which appears to be at this URL:
Provides no license, copyright or other statement of how it may be (re)used.
Finally the microformats rel-registry, and existing-rel-values as with the rest of the microformats wiki are published with a Creative Commons Public Domain License.
Thus comparing these three (microformats, W3C/WHATWG, RFC/IETF/IANA) :
- microformats has the most liberal (least private) availability as public domain
- W3C HTML5 (via WHATWG's "HTML") has the effectively as liberal MIT license
- RFC/IETF/IANA has no obvious license and thus it is unclear.
Is there a machine processable version of the microformats rel registry
Q: Is there a machine processable version of the microformats rel registry?
A: Yes. Each of the levels of rel values (formats, proposals, brainstorming) are in their own semantic HTML table, in order (1,2,3) on the existing-rel-values page.
Machine processors may pick the table of their choosing by choosing the appropriate ordinal (1,2,3) and then parse it as well-formed markup, using the first column of the table as the rel value, the 2nd column as informative text, and the first hyperlink in the 3rd column of a row as the defining specification/draft/proposal where the rel value is further explained.