- From: Reto Bachmann-Gmür <reto@gmuer.ch>
- Date: Mon, 30 Jan 2006 14:38:53 -0500 (EST)
- To: "Tim Bray" <tbray@textuality.com>
- Cc: "Larry Masinter" <lmm@acm.org>, =?ISO-8859-1?Q? 'Reto_Bachmann-Gm=FCr' ?= <reto@gmuer.ch>, "'Windows-world'" <windows-world@wanadoo.fr>, uri@w3.org
> On Jan 30, 2006, at 10:34 AM, Larry Masinter wrote: > >> http://www.w3.org and http://www.w3.org/ are two different URIs >> that identify the same resource. They identify the same resource >> because the two different URIs specify the same effective procedure >> for connecting to the resource (whether via GET, POST, or some >> other HTTP method). > > For an example from the other direction, http://www.tbray.org/ongoing/ > misc/Tim and http://www.tbray.org/ongoing/misc/Tim/ do not identify > the same resource (in fact, there is no resource corresponding to the > latter) Yep - my URI styling question was in fact about this case, the best practice I would suggest (new paragraph for Timbl's "Cool URIs" document) would be: "Cool HTTP-URLs don't end with a slash unless they point to the root of a webserver or point to an index of subordinate resources." See my previous posts for reasons/arguments. I didn't mention the second exception before, but I think this is coherent with the meaning such an URL has in the context of webdav. I realize there is still some grey between a "rich index" (with slash) and a "superodinate resource" (without slash), still if this practice would be adopted for the w3c-website, chances to guess if an URL ends with slash or not would be much higher. Or did anyone succeed in the puzzle in my second but last mail? > This breaks poorly-written web robots that foolishly assume > that they can append "/" to URIs where the last path component > doesn't have a "." -Tim I wouldn't use the attribute "web" for such pieces of software :-) Reto
Received on Monday, 30 January 2006 19:39:03 UTC