[
Lists Home |
Date Index |
Thread Index
]
I've been enjoying Tim Bray's articles on searching
at his Ongoing site (http://www.tbray.org/ongoing).
While walking the dog in the rain last night, I realized
I don't understand how some of the popular search
engines handle topicality, that is, do they understand
it at all? My thoughts were drifting between the
attractors of user interfaces and search mechanisms,
my main noodle being how GUIs and presentation of
results affect search.
Of all the different GUI interfaces, Google is the simplest.
Type in a list of words and hit submit at a minimum. Others
do that, but Google is spectacularly unbaubled. Is that
a good thing or a bad thing (good for using it, possibly
bad for interpreting results).
1. If a list without qualifiers of any kind is put in,
how does Google know, or does Google know which one
represents the topic, the dominant search term?
It can't, but does it try to guess and how?
2. Is Google searching for all occurrences or only
referenced occurrences (eg, PageRank)? See Question 1.
3. Are topics (dominant terms) and PageRank related?
4. Should ranking filters be reflected in the user
interface so the user can adjust their strategies
to meet their expectations, in other words, how much
should they come to understand the filtering effects of
page ranking? (Should they interpret Google Numbers to
in terms of the algorithm? How would a GUI help
them do that? Visualization of clusters?)
len
|
