I think the problem here is that you're associating "relevance
ranking" with "the way III implemented relevance ranking in their
OPAC". Analogies like this are like saying we should scrap the
automobile because you drove a Yugo once. Maybe we do need to scrap
the automobile, but let's not conflate the Yugo as omega of automotive
engineering.
There is also the "one person got weird results when he tried to use
this tool in a way it's not well equipped to handle" aspect of your
anecdote, as well.
-Ross.
On Jan 4, 2008 3:01 PM, Schupbach ,William <w.schupbach_at_wellcome.ac.uk> wrote:
> A researcher entered a search in our online catalogue yesterday by
> typing in the word "vliet". He got back a "relevance ranking", failed
> to find the item he was looking for, and emailed me today because he was
> surprised we did not have it.
>
> We do actually have it, but the computer told him (or would have done
> had he persevered) that it was the *least* relevant item having the word
> "vliet". In reality of course it was the *most* relevant item, but the
> computer could not know that on the evidence of just one word.
>
> A searcher might well feel aggrieved at the arrogance of the
> programmers in purporting to tell him the items most relevant and least
> relevant to his search. How would they know? I agree with Martha Yee's
> suggestion in one of her recent publications that catalogues should
> eschew claiming such a skill-until they can live up to that claim.
>
> William Schupbach
> Wellcome Library, 183 Euston Road, London NW1 2BE
> E-mail: w dot schupbach at wellcome dot ac dot uk
> Catalogue: http://catalogue.wellcome.ac.uk/search~S5/X&SORT=AX
>
> The Wellcome Trust is a charity registered in England, no. 210183. Its
> sole Trustee is The Wellcome Trust Limited, a company registered in
> England, no. 2711000, whose registered office is 215 Euston Road, London
> NW1 2BE, UK. For information about Wellcome Collection, our exhibitions
> venue go to www.wellcomecollection.org.uk
>
>
>
>
> This message has been scanned for viruses by BlackSpider MailControl - www.blackspider.com
>
Received on Fri Jan 04 2008 - 15:34:43 EST