WAGging Google Scholar (and beyond!):

Localizing and Contextualizing the Web

Ross Singer

Georgia Tech Library

GUGM 2005

Enter the library (or else)

A history of collection development & access in under 3 minutes

We buy or "lease" items that our core users would want
We categorize and index them in our own arcane way
We make them available for the users -- if the users come to us

All access is through the "Front Door"...

Enter the proxy server

This changes everything!

No longer need to be on campus
Still requires the "front door"
Renders other libraries' webpages worthless
Prairie View A&M Database Page

...except it doesn't really change anything.

Enter the problem

Maybe the front door isn't the best way in

From LibQUAL+:

via: Lorcan Dempsey's weblog On libraries, services and networks.

Lorcan Dempsey's Weblog: http://orweblog.oclc.org/

More "bad" news

From John J. Regazzi: Managing Director, Market Development, Elsevier

In a survey for this lecture, librarians and scientists were asked to name the top scientific and medical search resources that they use or are aware of. The difference is startling. Librarians named Science Direct, ISI Web of Science, and Medline, while scientists named Google, Yahoo, and PubMed (librarians also named PubMed).

_{via: nfais}

Also in mind blowing "chart-form" (from the above)

Full transcript: http://www.nfais.org/publications/mc_lecture_2004.htm
John Regazzi's presentation slides: http://www.nfais.org/RegazziNFAISFeb04.ppt

Google changed everything. Again.

Google Scholar, Elsevier's Scirus, PubMed
Instead of being controlled indexes of licensed resources, just freely available indexes of all resources available to index
Not their problem or mission to make sure appropriate copy is available. (more on this later)
Presented copy is available for a price ingenta.com | Science Direct
FindArticles.com, Goliath, HighBeam Research, Questia

Google Scholar: http://scholar.google.com
Scirus: http://www.scirus.com
PubMed: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi
FindArticles.com: http://findarticles.com
Goliath: http://goliath.ecnext.com/
HighBeam Research: http://www.highbeam.com/library/
Questia: http://www.questia.com/

Google Scholar: Catalyst for Change

Google Scholar: debuted November 18th
Peter Binkley created Firefox extension: November 30th *
Art Rhyno creates bookmarklet rewriting URLs to proxied URLs
WAG (Windsor-Alberta-Georgia) the Dog formed to address localization: December 3rd
Art moves his project server-side, proxying GS through Tomcat/Cocoon
PHP Web Localizer is born
Openly Informatics takes over development/maintenance of Peter's extension

Peter Binkley's original Firefox extension: http://gslocal.sourceforge.net/
PHP Web Localizer: http://rsinger.library.gatech.edu/localizer/localizer.html
Openly Informatics' OpenURL Referrer: http://www.openly.com/openurlref/

Firefox extension:

Google strikes back

Scholar Preferences

Original pilot with 30 schools (mostly SFX) early '05
For items with DOIs or PubMed IDs, link resolver links are placed in citation
Available either by special link from institution's website or by user choosing affiliation
Requires the institution to send holdings list to Google
"Title link" in citation still may not go to appropriate copy

Scholar Preferences: http://scholar.google.com/scholar_preferences?prev=/
Scholar Support for Libraries: http://scholar.google.com/scholar/libraries.html

A simple alternative

(Latent/Embedded/Persistent) OpenURL (Autodiscovery)

Dan Chudnov & Jeremy Frumkin: Service Autodiscovery for Rapid Information Movement
OpenURL links that point to no resolver and have no anchor text
CiteULike, Connotea (NPG), Hubmed, Scholar's Box
Opening up OpenURLs with Autodiscovery. Chudnov, Cameron, Frumkin, Singer, Yee: Ariadne #43
Eric Hellman coins term "Latent OpenURL", pushes spec based on OpenURL v1.0 (Z39.88)

Service Autodiscovery for Rapid Information Movement: http://curtis.med.yale.edu/dchud/writings/sa4rim.html
CiteULike: http://citeulike.org
Connotea: http://connotea.org
HubMed: http://www.hubmed.org
Opening up OpenURLs with Autodiscovery: http://www.ariadne.ac.uk/issue43/chudnov/
Latent OpenURLs in HTML: http://www.openly.com/openurlref/latent.html

WAGging

Installing the WAGger

Google Scholar: Wing Drag | Hegel's Dialectic | "ingenta.com"

Scirus
PVAMU DB List
MIT Databases for Physics
Caltech: Physics
Wikipedia: Chaos Theory
OpenURL Autodiscovery
- Canary DB | CiteULike | HubMed

Things still to do

Direct linking with incomplete citations *ahem*
Database/Journal Advisement - the WAGNet
Metasearch
Virtual Reference
Advocacy/Acceptance of OpenURL autodiscovery tags
Folksonomic Linking - Connotea, CiteULike, del.icio.us, technorati, OPAC
OpenURL Autodiscovery
Toolbar, Greasemonkey/Firefox extension, "WAGger persistence"

WAGger vs. Scholar Preferences

	Scholar Preferences	WAG the Dog
Advantages	Instantaneous response; Little requirement for users; OpenURL links definitively go to fulltext every time; Links to fulltext in citation are highlighted	Not limited to Google Scholar; Tries to content/services for every citation; Ability to push more of our collections/services to user based on context
Disadvantages	Limited coverage for OpenURL; Strange requirements for libraries to join; No possibility of ILL, print, etc. for items not held in fulltext	Only as accurate as the citation data; requires user to install something; requires user to activate on every target; the more targets = slower response

Conclusion

A lot of work yet to do
Worry less about "perfect" results; focus more on helping find "more relevant" results
Thanks to rigid library standards, "sloppy" interfaces can flourish in our environments... if we let them
We ~~can~~ have to all be friends

http://rsinger.library.gatech.edu/GUGM/ • May 19th, 2005

WAGging Google Scholar