W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1997

Re: HTTP question.

From: David W. Morris <dwm@xpasc.com>
Date: Mon, 24 Nov 1997 09:12:17 -0800 (PST)
To: "Choi, Hyoung-Kee" <hkchoi@cc.gatech.edu>
Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Message-Id: <Pine.GSO.3.96.971124085624.14815B-100000@shell1.aimnet.com>

What you have encountered is a Netscape extension which is also
implemented MSIE ... the REFRESH header. This is typically generated
using the HTML <META> http-equiv markup rather than real HTTP headers.

The feature is not part of 'official' HTTP.

The really onerous aspect of refresh is the usage you have noticed where
the refresh continues to loop long after the user has gone to lunch, 
etc. I discovered this usage of refresh when I went to track down
why my ondemand ISDN connection was remaining active hours after anyone
had been knowingly using the Internet.  The page in question had
an advertising banner which was being refreshed every 20 minutes
thereby generating many false hits against the advertisement (my believe
is that it was fraud but the banner company apparently didn't care as
they ignored my mail).

By luck, I noticed the ISDN activity, otherwise even the usage pennies
would have added up.

If this, or a future IETF group, should consider this header for
standarization, I will be quite vocal in my crusade to have limits
set on the number of refreshes between human intervention or perhaps
the amount of time it continues.

The refresh header provides a very useful facility but it is also 
subject to the potential for serious abuse in terms of useless network
traffic and usage cost to end users who no clue what is happening.

Dave Morris

On Mon, 24 Nov 1997, Choi, Hyoung-Kee wrote:

> Hi Everybody!
> 
> 	I am doing research on Web traffic modeling at G.I.T.
> With certain Web page(www.cnnfn.com), the client continues to re-connect
> with that page in exact every 180 second automatically. This situaion was
> found on
> tcpdump of dial-up Web traffic. 
> 
> The user in a client watches the
> time and press the button ??  It is not a possible scenario.
> 
> If anyone is able to answer to the question, it would help me a lot.
> Thank you in advance.
> 
> PGP public key is available upon request.
> 
> CHK
> URL: http://www.ee.gatech.edu/users/hkchoi
> 
> 
Received on Monday, 24 November 1997 09:14:31 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:33:04 EDT