[
Lists Home |
Date Index |
Thread Index
]
I think Len is arguing that if we know more about how google searches, we
can interpret the results more.. such as in germany, I get german pages
etc..
I think several techniques used by google are kind of known..?? for
example, we know that it uses page rank, text in anchor tags, tf-idf etc..
but we do not know the exact numbers, and we do not know other metrics
used etc..
I think the question is: has google crossed the line with respect to what
is proprietary and what is disclosed. I think Len argues that it has..
To me, actually, I do not like too much non-determinism in the results
that I see.. I think google has crossed the line with the non-determinism
seen by the user, and user-based adaptivity.. they are using wrong
metrics..?? Irrespective of where I am, I need the same results.. I think
language etc should be handled differently..???
After hearing that google uses user-based adaptivity, and that they are
detecting user characteristics automatically.. I think that is not a
correct approach.. google always maintained they wanted to keep things
simple.. I think this is not keeping things simple... but what is
alternative approach??
as someone pointed out, we can go out google.com/ncr to get the same
results irrespective of our location??
Also Diddier, thanks for the pointer..
best, murali.
On Tue, 9 Dec 2003, Bullard, Claude L (Len) wrote:
> One doesn't know. That is precisely the problem.
> To layer a service over a service, one must know.
> QoS matters.
>
> len
>
>
> From: Murali Mani [mailto:mani@CS.UCLA.EDU]
>
> That is surprising; if they use a deterministic and same algorithm
> irrespective of the server used by google, then we should always get the
> same result unless data set really changes (for example, if ranks change
> between the queries etc)..
>
> Am I not right? Or is it really that different servers will give different
> results..?? I am slightly sceptical..
>
> -----------------------------------------------------------------
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
>
> The list archives are at http://lists.xml.org/archives/xml-dev/
>
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://lists.xml.org/ob/adm.pl>
>
|
