SIGIA-L Mail Archives: RE: SIGIA-L: Popularity in Relevance Ran
RE: SIGIA-L: Popularity in Relevance Ranking Was: understanding metadata
From: James L Weinheimer (jamesw_at_princeton.edu)
Date: Thu Sep 06 2001 - 05:58:55 EDT
>Besides, Google also measures the popularity of sites that link to >you,
so a few links from popular sites would push your site higher in >the
rankings, while a lot of links from less popular sites wouldn't.
>To continue with the burger analogy: if Google were a burger searcher, >it
would rate the few links from gourmet magazine sites higher than >the many
links from personal home pages.
There's certainly nothing wrong with arrangement by popularity. My own
disagreement is with the argument that arrangement by popularity gives the
"best" sites. If this popularity comes from home pages, or from
profit-driven magazines is beside the point.
Arrangement by popularity is one method of finding items that may be
excellent, but it could also be bad and certainly can be misused. I don't
think any of us would be overly shocked to discover that
Google-arrangements are being misused right now, but I have no evidence of
this. A few years ago, metadata was misused so seriously that it was
disallowed in Yahoo and Excite--I don't know if they've allowed it again.
In other areas, when university administrators began to use citation
indexes for tenure purposes, there was a lot of: "you cite me and I'll cite
you" going on. I'm sure this is still done.
I can't blame anyone for this "misuse"--each one just want to get their own
site higher in the rankings. One person's "misuse" is another's "brilliant
business practice". After all, if a CEO sees that his $2 million site
doesn't come up any higher than 500th place, what is he going to think of
his web people?
But, there should be more than one way of arranging results. I confess that
I still get tons of irrelevant material I have to wade through when I go to
Google, but it's currently my favorite search site.
This archive was generated by hypermail 2.1.2
: Sun Nov 23 2003 - 22:54:48 EST