- From: <hallam@etna.ai.mit.edu>
- Date: Thu, 11 Jul 96 15:40:51 -0400
- To: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
- Cc: hallam@etna.ai.mit.edu
Perhaps we need to address the following questions separately:- 1) What is the information that the Advertisers/Auditors want? I see this as beeing at two levels, 1) What is the information they want to obtain? 2) What measurements do they want to capture to achieve (1) Referer comes under (2) in my view, as do most protocol level hacks. The important question under (1) is "what financial value will an advertiser recive from advertising at a site?" 2) What mechansims are needed to transport information captured under (1) Note that in my Log file format and Proxy notification specs I worked mainly on the second issue and tried to make the transport mechanism as separate as possible from the issue of what must be transported. A second scenario I think that the Newspapers in particular should consider. Say I am a comuter who wants to download the Times each night and pick up a loaded laptop on the way out to work. In that case it would be very usefull for the laptop to inform the originator what material was actually read the next time a download was taken. This is a material benefit to the user - the newspaper can discover what articles it publishes are important to readers and which they don't bother reading. [Yep Privacy issues up the Wazzo, privacy is only one ethical dimension however] Phill
Received on Thursday, 11 July 1996 12:39:31 UTC