- From: Ray Denenberg, Library of Congress <rden@loc.gov>
- Date: Fri, 12 Mar 2004 09:21:44 -0500
- To: "Roy T. Fielding" <fielding@gbiv.com>, <uri@w3.org>
----- Original Message ----- From: "Roy T. Fielding" <fielding@gbiv.com> To: <uri@w3.org> Sent: Thursday, March 11, 2004 10:39 PM Subject: Re: info scheme has no authority component, why? > > > Indeed, the reg-name token in 2396bis seems to be targeting a different > > usage. Whereas the RFC-2396 reg_name was a kind of non-host authority > > that could not have a port number, the 2396bis reg-name is a kind of > > host and can have a port number. Maybe the 2396bis vision is not to > > provide for abstract registered naming authorities as described in > > RFC-2396 and info-uri-01, but merely to allow network entities (hosts, > > services, domains) to be named using more naming systems than just > > RFC-1123 hostnames. Is that the intention? > > No. What the syntax allows and what a specific scheme uses are two > different things. There is nothing wrong with a scheme that only > uses a subset of the available syntax, provided it avoids the > reserved characters that would indicate otherwise. Moving reg-name > under host has the appropriate impact of reserving the ":" and "@" > characters for a specific purpose, but no effect on any URI that > might have been defined to use the old reg_name (which also reserved > those characters). The result is simply less ambiguous to parse. > > > On the other hand, if the intention of 2396bis is that reg-name can > > really be an abstract naming authority, shouldn't info: be using it? > > The "info" scheme proposal misuses almost every single aspect of URI > syntax, philosophy, technology, and accepted best practice. There is > no need for it to exist at all. Yes, it would be better to use either > the generic authority syntax or the URN authority syntax for new URI > schemes that make use of delegated naming authorities. > > ....Roy >
Received on Friday, 12 March 2004 09:21:46 UTC