W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 2007

Re: New issue: 6.1.1 too vague about parsing requirements

From: Lisa Dusseault <lisa@osafoundation.org>
Date: Thu, 11 Jan 2007 11:17:38 -0800
Message-Id: <9F0A66FD-6BB2-4B6C-9E73-E8F7584F73B0@osafoundation.org>
Cc: HTTP Working Group <ietf-http-wg@w3.org>
To: Henrik Nordstrom <hno@squid-cache.org>

On Jan 11, 2007, at 1:43 AM, Henrik Nordstrom wrote:

>   The client SHOULD NOT examine the Reason Phrase for other
>   purposes than displaying it to the user.
>

Displaying to the user is problematic if you mean the browser end- 
user, because the reason phrase isn't internationalized over the  
wire, nor is it well-suited to be used as a localization lookup.

We can also derive a more enforceable style of requirement by looking  
at who's putting the reason phrase on the wire.

Perhaps something along these lines:  "Servers MUST NOT use the  
Reason Phrase for information needed by end users, because the text  
is not translatable.  Servers MUST NOT use the Reason Phrase for any  
information needed by automated agents, that is, any case that would  
require a client to try to parse or match the reason phrase in order  
to decide what to do.   Reason phrases are for debugging and logging,  
whereas it's the role of status codes to be used both to look up  
localized error messages for display to the user, and for helping  
automated agents decide whether something succeeded or failed or what  
to do next."

thx,
Lisa
Received on Thursday, 11 January 2007 19:17:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:00 GMT