She begins with a summary of key findings from the 2005 Bibliographic Services Task Force report:
- Focus on what users want--Users care about delivery as much as discovery
- Reduced click streams key, even at expense of good description
Still need to understand user behavior--do people click on something because it's at the top of the list or because it's what they want?
Need to define collections in new ways; print-digital distinction blurring. Also, unique special collections are more widely available.
Embed services where users are, such as learning environments, Web sites, desktops, and other applications.
Meet user needs and solve user problems. Try something, assess it, and change quickly if needed.
We need more Clickstream analysis and eyeball tracking to better understand user behavior.
The Ubiquitous Librarian is a strong proponent of innovative uses of Twitter in libraries. He recently joined staff at UCSB.
Librarians are a distinct class of users; will always have tensions between needs of end users and those of librarians.
Many reasons to partner with OCLC:
- Amazing record base--140 million records, 60 million in queue
- Similar visions/goals/interests
- Chance to leverage investment across institutions
10% of searches return no results at present. Could suggest alternate spellings in future.
Pilot: 10 campus specific views and 1 UC-wide view. Only one ILS represented per campus, which is sometimes a problem.
Interoperability with UC eLinks and Request is crucial; so far better success with UC eLinks than Request, in pilot.
Future task: Mining authority records for recommender services, a la Amazon.
For large record sets, retrieval speed will need to improve. Changes in the works.
Pilot to be extended until all outstanding requirements are met by OCLC--Request, multiple ILS's at one campus, retrieval speed. No firm time for when pilot concludes; "Melvyl not going away just yet."