using web services to enhance services for library users

Regarding Terry Reese's conference keynote, “Making our library services ‘next generation’ and how we get there”:

Given the WSU Libs current, peripheral work with the Summit software development that’s been so successful in its early stages, I’m enthused about this presentation for this. Terry noted that he works with a growing group of library developers who are interested in reducing the number of the search and delivery silos that library users have to deal with. Reese asserted that Innovative Interfaces’ products are very good at the staff functionality, and traditionally very good with public interfaces. For a third community cited by Terry, the library developer community, Innovative has been less friendly in terms of support.

He described the current situation in ILS support and development. Pre-Encore, functionality is gained thru hacks (in the most negative sense), individual programs built on top of an integrated system. As one example, Terry cited Library Thing (used in the Danbury Library Catalog); this functionality is obtained thru JavaScript, by inserting info into the OPAC. The negatives of this approach: hacks are fragile, relying on processes such as screen scraping, and they tend to be stop gap in nature.

Terry noted that he doesn’t like the “next generation” term, because it’s essentially undefined. He then went thru a couple of candidate systems, such as Amazon, the NCSU catalog, and an Encore system (Michigan State’s instance was shown). As I wrote in my LITA notes for the University of Rochester session, Terry also noted that these “next generation” examples meet, not exceed, the current expectations of our users. Instead of functionality, as exemplified in these three systems, Terry urged attendees to focus on the underlying data service models, models that enable developers to use data for different purposes.

This approach is disruptive, in that the focus shifts from technical development to “what you can do with the data.” The problem that Terry notes: when users come to the library, this service model (which appears in tools such as Amazon and Flickr) disappears. To resolve this problem, he calls for the disintegration of services; here are some attributes:

-Library services must use data that can be accessed thru multiple interfaces
-Library services much support multiple access methods
-Library services must allow the repurposing of data and services

Terry described he big picture goal for the OSU Libraries (re: where the Libraries wants to be in five years): “simple aggregation of resources will not be enough. Databases and websites will be replaced by mechanisms that fetch information into our users’ workflow.”

Terry then provided an assessment of what users want? His major points:

-Zero configuration services that are always on
-Services that auto advertise themselves and become live as users come into their domain (note the Starbucks and Apple partnership that Jeremy described at the LITA National Forum).

To meet these challenges, Terry noted that:

-How we deliver information is becoming as or more important as what information is being developed
-Patrons (and libraries) expect to have control over the information being sent

Terry also emphasized the importance of serving information to intermediate services instead of end users. He then described the Google/Amazon models, which are popular with developers because of their respective support for community development. This is an area where vendors (including Microsoft) are better at in comparison to library organizations. For example, Amazon gives out tagging, book cover images and reviews for free, as long as you’re linking back to the Amazon site (which is free advertising for the Amazon service). Other products/services that follow this model is Flickr and its API and development associated with Google Maps (example).

Terry then touched another “mash-up product” LibraryFind, which I heard described and saw briefly demo’d by Jeremy Frumkin ten days ago in Denver. He also noted that the Summit-related development work done in the consortium, which I’ve demo’d in a couple of meetings at WSU, is another example mash-up example, taking results from several data sources to create an improved UI.

In summary, Terry noted that the Web 2.0 data model underscores a recognition by the information provider that the use will find different and better uses for their resources. I’d like to think that we’ve (the NWDA development team, past and present, at WSU) supported a service-based model through the Northwest Digital Archives development. Finding aid information is accessible through a specialized user interface; is used to create search support in this UI (for example, the creation of a filing title drop-down list); is used to create an OAI-PMH data provider instance; and, is used to support Google Sitemap creation and search engine harvesting based upon an archival identifier standard, the Archival Resource Key. This data service emphasis in the NWDA project exists in large part because of the chronological time of this development (2003 forward). In a sense, we may have to reinvent our systems to accomplish this services or “middle layer” middle envisioned by Terry and articulated during his session.

Continuing on for the future: Terry noted the need to build services that are part of a larger organized library platform; he cited DSpace as an example, which is employed for the WSU Research Exchange service.

Terry noted that Innovative Interfaces needs to embrace a web services design for all its products, and to make these interfaces part of the base packages for these services (and not an add-on that requires payment for use). Additionally, he notes that III needs to support emerging protocols outside of the library community like OpenSearch, or library protocols such as SRU and NCIP. Finally, he noted the need for III to create more agnostic tools that can be used outside the library catalog (he cited the expensive add-on, Metadata Harvester [which I reviewed and rejected for NWDA use], as an example Innovative Interfaces application that supports data use outside of the OPAC).

Concluding, Terry noted that organizations need to cultivate a culture that encourages risk taking. This is a point that’s been discussed in my recent WSU Libs leadership cohort sessions. As one example, embracing the concept of the WSU Libs committing more resources to local software development to complement the licensing of commercial resources is a risk.

I got a lot out of his remarks, and am looking forward to Kyle’s conference-wide speech tomorrow on Summit search system development.