more on "Why I won't be a Google IPO Investor"

[Dave Farber <dave () farber net>]

Delivered-To: dfarber+ () ux13 sp cs cmu edu
Date: Sat, 01 May 2004 12:28:11 -0700
From: Andreas Ramos <andreas () andreas com>
Subject: Re: [IP] more on "Why I won't be a Google IPO Investor"
To: dave () farber net

> Today's search engines are not evolved enough to guess what we mean when
> we type in a single-word search query while looking for answers to complex
> questions.

Here's a bit of text from my upcoming book on search engines.

"These various (search engines) can be tested by using the same dataset,
making a search, and comparing the results. This uncovers problems that
arise from the nature of the data.

"For example, if a group of pages are highly-interlinked within a "tight
knit community" (TKC), the pages can appear to be significant although they
have no real value. For example, several years ago a bunch of bloggers
linked the words "dismal failure" to Bush's personal page at the White
House. If you search for dismal failure, you got George W. Bush.

"Link analysis work well when the topic is clearly defined, there are
significant articles about it, and it has an interconnected community. But
if the topic is vague, there aren't good webpages about it, or there aren't
interconnected communities that discuss the issue, then the search engines
produce wrong results. This is known as the topic drift problem (TDP). For
example, search for net gain. This produces 6.8 million hits, but the
results are random pages.

yrs,
andreas
www.andreas.com

-------------------------------------
You are subscribed as interesting-people () lists elistx com
To manage your subscription, go to
 http://v2.listbox.com/member/?listname=ip

Archives at: http://www.interesting-people.org/archives/interesting-people/