I feel like a couple years ago a librarian(s?) created a Google Custom Search Engine that did exactly what you describe as "focused searching," but I can't find a link any more. You can search the CSEs by scrolling down on this page (and there are a couple of links to directories, too): http://www.lib.berkeley.edu/find/types/websites.html
Also, Mike Eisenberg over at the University of Washington was working on that kind of problem with some other groups...A quick search reveals that it's now called "Reference Extract" and it's being done in conjunction with Syracuse University (and OCLC is somehow involved). http://chronicle.com/blogPost/Librarians-Want-to-Out-Google/4365
From: Cindy Harper [[log in to unmask]]
Sent: Tuesday, September 29, 2009 10:53 AM
To: [log in to unmask]
Subject: [CODE4LIB] Bookmarking web links - authoritativeness or focused searching
I've been thinking about the role of libraries as promoter of authoritative
works - helping to select and sort the plethora of information out there.
And I heard another presentation about social media this morning. So I
though I'd bring up for discussion here some of the ideas I've been mulling
Last week I sent this message to the "Suggestions and Ideas" forum at
The basic idea is to develop a delicious network of librarians. Or a network
of faculty members. Then have one login whose network included those users,
and share that login so that lots of people could share that network.
Delicious responded that we could have a wiki where people posted their
delicious names so that others could add them to their personal networks,
but that doesn't scale up very well.
Or another project I've toyed with, involving focused searching: I started
with Robert Teeter's index to Great Books lists.
I've almost completed pulling them into a MySQL database so that I could
sort the titles by the number of Great Books lists that mention each title.
Then I thought about how one could do focused searching of the web,
collecting pages with a title containing (best and books) or (great and
books), and screen scraping title lists (you'd have to have some heuristic
method of identifying the data, of course, and I'm aware what problems might
arise there). But my test searches in that idea showed that one runs into a
lot of commercial ephemeral lists and spurious lists. Now, you could rely
on crowd-sourcing to filter out the consensus by ranking by the number of
sites/cites. But I thought you might want to differentiate between the
source - .edus, librarys, etc.
So that led me to speculate about a search engine that ranked just by links
from .edu's, libraries sites, and a librarian-vetted list of .orgs,
scholarly publishers, etc. I think you can limit by .edu in the linked-from
in Google - I haven't tried that much. if anyone here has experience at
using tha technique, I'd like to hear about it. But I'm thinking now about
the possibility of a search engine limited to sites cooperatively vetted by
librarians, that would incorporate ranking by # links. Something more
responsive than cataloging websites in our catalogs.
Is anyone else thinking about these ideas? or do you know of projects that
approach this goal of leveraging librarian's vetting of authoritative
Cindy Harper, Systems Librarian
Colgate University Libraries
[log in to unmask]