- From: Julian Reschke <julian.reschke@gmx.de>
- Date: Thu, 23 Jun 2011 20:59:18 +0200
- To: Boris Zbarsky <bzbarsky@MIT.EDU>
- CC: public-iri@w3.org
On 2011-06-23 20:11, Boris Zbarsky wrote: > On 6/23/11 2:05 PM, Boris Zbarsky wrote: > ... > On its face, either a URI is just a string (in which case it doesn't > matter where it came _from_ and what happens when you do things with it > should just be covered by the spec that defines said things and has no > place in the URI RFC) or a URI has some sort of additional metadata > attached to it about being a same-document reference (or not) as 4.4 > seems to imply. Which is it? > ... A URI is a string and doesn't have any hidden metadata attached to it. The check for "same document" is made when you resolve the URI, and you determine the base URI at that moment of time based on the rules for the media type in which it occurs. No need to attach something to the URI. Or, alternatively: if you parse the reference early, and then need to attach data, that's an aspect of your implementation. The URI spec does not care. Now, whether the whole section is important or not is an interesting question; maybe Roy can enlighten us a bit about the history. If this is the *only* RFC3986 problem browser vendors have, I'm more than happy :-)
Received on Thursday, 23 June 2011 19:00:06 UTC