W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > May to August 1998

Re: Proxies and gethostbyname

From: David W. Morris <dwm@xpasc.com>
Date: Mon, 11 May 1998 11:59:56 -0700 (PDT)
To: Henrik Frystyk Nielsen <frystyk@w3.org>
Cc: http-wg@cuckoo.hpl.hp.com, jg@w3.org
Message-Id: <Pine.GSO.3.96.980511114941.20100H-100000@shell1.aimnet.com>


On Mon, 11 May 1998, Henrik Frystyk Nielsen wrote:

> I think there is one situation that 504 doesn't cover and that is if the
> proxy doesn't *want* to proxy a certain URI, maybe because it can't

I continue to disagree ... there is a distinct difference from the 
perspective of recovery whether the DNS consulted by proxy is unable
to resolve the host name or whether the host fails to respond. Hiding
all failures under a single code does nothing for usability or
perceived reliability.


> 418 Not Proxying
> 
> The server is not willing to proxy or gateway the Request-URI. No
> indication is given of whether the condition is temporary or permanent. The
> client MAY repeat the request using either a different proxy or no proxy at
> all.
> 
> The 504 (Gateway Timeout) status code SHOULD be used if the server can not
> serve the request due to upstream errors.

Why bother .... there is a whole lot less an end user will be able to do
with this distinction over a simple not authorized response than over
a mistyped host name in a URL ... the suggested 418 may be useful to know
but on a priority scale less important to know than better granularity
on conditions covered by 504.

Dave Morris
Received on Monday, 11 May 1998 12:04:44 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:33:18 EDT