W3C home > Mailing lists > Public > www-tag@w3.org > November 2006

Re: [metadataInURI-31] Update on Use of Metadata in URIs Draft

From: Karl Dubost <karl@w3.org>
Date: Wed, 1 Nov 2006 13:43:45 +0900
Message-Id: <3124FFE0-69D7-4EDD-A0C8-C2E4301F7A69@w3.org>
Cc: www-tag <www-tag@w3.org>
To: "Rice, Ed (ProCurve)" <ed.rice@hp.com>

Hi Ed,

Le 1 nov. 2006 à 03:28, Rice, Ed (ProCurve) a écrit :
> http://www.hixie.ch/tests/adhoc/http/content-type/
> The issue is many of these files are are supposed to be text.html  
> but are in fact other file types.  The browser ignores the file  
> type and attempts to interpret the file by extension.  We should  
> either eliminate the file type (Ignore it) and enforce file  
> extensions or we should enforce file-type and ignore file  
> extensions.  Seems like we're inconsistent.
> My take on this is that users are becoming better educated in their  
> use of the web and when the click on a .rss file they expect to see  
> an RSS feed instead of an xml file or an html file.

Taking into account your comment, how do we deal with these URIs?

  * http://example.org/foo/
	which can be the listing of a directory or any kind of content in fact.
  * http://example.org/foo/bar
	same here (happen very often in content negotiation)
  * http://example.org/foo.cgi
	a CGI program which can send anykind of content.

An URI doesn't end with the name of a file on a file system. It is an  
identifier. Sometimes the server deliver a file contained on a file  
system, but sometimes not (request to database for example).

If we choose one solution or the other, we will always loose, because  
I guess there will be always a camp.
   * eliminate the file type (Ignore it) and enforce file extensions
   * enforce file-type and ignore file extensions.

That will not work. We have to acknowledge that the error will  
happen. More interesting would be to find the workflow in client- 
server communication which makes it possible to detect the error, to  
report it and then to fix it, or help to fix it when possible, and  
this, for each HTTP verb.

If I had time that would be a good opportunity to revise CUAP[1] and  
CHIPs[2] as practical recommendations to solve errors in  
implementations (human or machine).

[1] http://www.w3.org/TR/cuap
[2] http://www.w3.org/TR/chips

Karl Dubost - http://www.w3.org/People/karl/
W3C Conformance Manager, QA Activity Lead
   QA Weblog - http://www.w3.org/QA/
      *** Be Strict To Be Cool ***
Received on Wednesday, 1 November 2006 04:44:01 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:50 UTC