- From: Roy T. Fielding <fielding@gbiv.com>
- Date: Wed, 11 Feb 2009 12:19:02 -0800
- To: Henri Sivonen <hsivonen@iki.fi>
- Cc: Mark Nottingham <mnot@mnot.net>, "www-tag@w3.org WG" <www-tag@w3.org>, Anne van Kesteren <annevk@opera.com>
On Feb 11, 2009, at 1:19 AM, Henri Sivonen wrote: > Requiring colonless rel tokens to expand to URI before comparison > would not work for me, because allowing "http://iana.org/whatever- > the-registry-url-is/stylesheet" and "stylesheet" as synonyms in > future UAs would be gratuitously incompatible with existing UAs > that look for "stylesheet" only. > > In short, my opinion is what I said at TPAC: If people who want to > mine HTML documents for data in order to put it into an RDF process > want to see rel values through URI-tinted glasses, that's fine. > They can check if the rel value contains a colon and if not prepend > a fixed string. However, requiring everyone else to process rel > values that way is not fine. In other words, if a short name is in the registry then it MUST be used short regardless of its corresponding URI. Parsers that are only interested in standard relations need only check for the short names. I don't have a problem with that, since it is good practice to encourage standard names and to avoid wasting bits on the wire. ....Roy
Received on Wednesday, 11 February 2009 20:20:06 UTC