Re: Robots

From: Chris Roadknight (roadknic@drake.bt.co.uk)
Date: Fri, Dec 11 1998


From: Chris Roadknight <roadknic@drake.bt.co.uk>
To: www-wca <www-wca@w3.org>
Date: Fri, 11 Dec 1998 16:02:47 -0000
Message-ID: <01be251f$b22537b0$2c819284@buccaneer.drake.bt.co.uk>
Subject: Re: Robots


>
>GUI browsing
>line mode browsing
>high bandwidth vs. low bandwidth browsing
>voice dictated browsing
>link checking and content validation
>mirror generating
>indexing (breath or depth)
>cache validation
>...
>
>all of which I think are valid ways of using the Web but are likely to
result in different browsing patterns. I also think that it is likely that
the same application can cover several modes and hence splitting it by
user-agent may not capture the various modes of browsing.
>
>Do you know of any mechanisms attempting to characterize browsing patterns?


(I'm not sure this is what you mean but....)
I have used request metrics (hit, requests, throughput rates etc) to try and
identify user clusters, ie. high/low bandwidth users.
My look-see experiments seemed to suggest that the search space was too
continuous for clusters to appear..I'm working more on this, with more data
so might get further soon.


>Henrik
>--
>Henrik Frystyk Nielsen,
>World Wide Web Consortium
>http://www.w3.org/People/Frystyk

Chris Roadknight
BT Research