[
Lists Home |
Date Index |
Thread Index
]
It means measurement is not independent of intent/purpose operationally.
Somewhat like the ad hoc query problem, one can get a simple answer or a
precise answer, or a simple interface to use or a complex interface
to use.
What is fascinating about Google is how well a very simple interface
enables reasonable precision and that most attempts to best it or
tune it are inevitably more complicated and less used. So a
measurement framework could be reasonable and simple, not
complete or accurate as the examples you provide demonstrate.
But again, the measure of the goodness of the answer depends
on the question (back to Fisher Information, it seems).
I would expect each audience to find a ratings provider tuned
to its questions. AARP for some; Vogue for others. There is
a part of a business there.
On the other hand, can one really standardize what the UDDI
descriptors should be, or does it have to be OWL and a trained
agent (human or software)?
len
-----Original Message-----
From: Joseph Mayer [mailto:jmayer@kathrein.com]
It just means
that even within one "industry", there may be a plethora of
metrics/descriptors
that need to be looked at. Perhaps too many for an accurate rating system.
|
