AltaVista Strange Results

[Helen R Lloyd]

> >Really though, what possible difference can it make to a user if there are
> >30,000 or 50,000 hits? I defy *anyone* to show me a student who tracked
> >down even 2,000 hits, let alone these larger numbers. What matters here is
> >relevancy, not numerical sameness....
> 
> Agreed, but what good is "relevancy" if 20,000 possibilities are left out
> of the determination of what's relevant?
> 
Exactly!  When I analysed the first 20 hits of the "date rape" query at 
both the original AltaVista site and the Australian mirror, there were 
only two matches.  The first and fourth items from the Australian mirror 
appeared in the first 10 hits from the original US site.  That's rather a 
lot of items not turning up in a search of the Australian site.  

Also a search that finds only 20 or so items at one site may find over 
100 at the other.  This is an important difference for a researcher.  

When is a mirror site not a mirror site?

What concerns me is that we have always recommended that our clients use 
the Australian site, and only list this site on our pages.  But now I'm 
thinking I should be including a link to the US site as well.

Helen Lloyd


Helen Lloyd                         | email: ulhrl at dewey.newcastle.edu.au
Faculty Librarian (Art & Design)    |
    and Web Manager                 |
Huxley Library                      | Ph  (intl+61+2) 49216455
University of Newcastle, AUSTRALIA  | Fax (intl+61+2) 49216904