[WEB4LIB] Re: In defense of stupid users

[K.G. Schneider]

\
> > The reality is that when you have been to Paree--the world of full-text
> > searching--going back to the farm of searching limited metadata is
> > frustrating and limiting, no matter who tells you otherwise. You can
> > poke in
> > a few extra terms here or there, add a few thingamabobs, but you can't
> > possibly match the richness of full-text retrieval.
> 
> But for my money the best solution is full-text searching "sweetened"
> with metadata to improve relevancy. Let's call it the Chicago solution
> --  a farm-influenced Paris.
> Roy

Yes, exactly, that is what I was trying to say when I said, "structured
information has its advantages, but ... we have to leverage it against
full-text discovery." 

Or you could call it metadata-searching made plump and juicy with full-text.
I think of it that way first only because of where I am in the library
world. I maintain a retrieval-challenged database (Librarians' Index) where
the sparseness of metadata presents a major obstacle to user queries
crossing what usability specialist Donald Norman termed the gulf of
execution. 16,000 items may not seem like a lot, but we have many core
resources from the Web. We would vastly improve retrieval if user queries
were passed through LII (find me everything from LII) then out to the Web
(find me everything from the Web) and then back into LII with an item match
(find me everything from this user query where the Web-in-the-Wild
identifier matches LII's identifier), and then display results deduped and
tiered: LII, LII+Web; then (draw a line and say "continue this search
outside of LII") Web-only. 

I think that would have been hugely ponderous even five years ago and quite
doable today. 

Karen G. Schneider
kgs at bluehighways.com