Whirlwind in Windsor surrounding integrated library systems: My symposium notes
On November 15 Rob Fox and I attended a symposium at the University of Windsor on the topic of integrated library systems. This text documents my experiences, and in a sentence, the symposium re-enforced much of what I had already thought regarding "next generation" library catalogs and at the same time it brought much more depth to the issue than I had previously given it.
Art Rhyno (University of Windsor) set the stage by providing a background of the current environment in "The Trip so far: A Journey with the ILS". He compared initial "integrated library systems" to the venerable Ford Mustang. Both were things a person could tinker with and customize to one's own taste. As things developed they have become increasingly less malleable. Integrated library systems vendors are also stuck in a difficult position. On one hand users' expectations regarding search and browse are at an all time high, and the systems just don't make the grade in this regard. On the other hand librarians still desire to keep their traditional workflows. These are competing desires resulting in a disconnect. Rhyno advocated a couple of things in order to move forward. First the information technology profession needs to develop sets of metrics used to measure success. Secondly, he advocated using existing building blocks to create any sort of "next generation" integrated library system. Specifically, the use of something like Lucene and its indexing technology for search, and something like PeopleSoft or SAP software for things like acquisitions. Finally, he compared the process of writing software to barn raising. Both processes are collaborative. Both processes are empowering. Both processes are educational opportunities. He subtly advocated open source software.
The keynote presentation, "Applying the Service Oriented Architecture (SOA) Model to Libraries", was given by Peter Murray (OhioLINK). He began by defining SOA through a number of negatives; SOA is not:
- a programming language or a network protocol
- about layering applications, but an architecture of discrete business processes
- about starting over, but about reusing thing in use now
- only about Web Services, but the use of Web Services as a means to an end
- about proprietary systems but standards
- about the how something gets done, but the what
A quote from the current version of Wikipedia echos much of Murray's definition:
A service-oriented architecture is not tied to a specific technology. It may be implemented using a wide range of technologies, including REST, RPC, DCOM, ORB or Web Services. SOA can be implemented without any of these protocols, and might, for example, use a file system mechanism to communicate data conforming to a defined interface specification between processes conforming to the SOA concept. The key is independent services with defined interfaces that can be called to perform their tasks in a standard way, without the service having pre-knowledge of the calling application, and without the application having or needing knowledge of how the service actually performs its tasks.
Murray then characterized the integrated library system market with words and phrases such as imploding, monolithic, and an environment of dueling press releases. Increasingly he has noticed smaller libraries using smaller software solutions and in the process shying away from the larger library systems. Things like Koha are workable solutions for the smaller library.
So, how might SOA be applied to the integrated library system? Easy. Outline sets of services to be implemented and then create applications that perform just those functions. Some of those services might include:
- known item search
- browse related items
- list recommendations
- read reviews
- provide relevance-ranked filters
- create automated descriptions of items
- display visual representations of lists
- annotate an item
- track-back an item from a blog
Once the these functions have been implemented using SOA techniques it will easy to integrate them into information retrieval systems ("catalogs"), metasearch systems, electronic portfolios, and/or course management systems. Additionally Murray advocated using the same techniques to make sure library content is discoverable by outside agents such as Google or via "mash-ups" with WorldCat.
I found the presentation by the folk of PINES (Brad LaJeunesse, Mike Rylander, David Singleton, & Julie Walker) to be the most interesting. Their presentation was called, "Evergreen: The ILs is open and everyone is invited!" They began by sharing an overview of the public library environment including 44 public library systems, 252 libraries located in 123 counties, 8.8 million records, and 1.6 million state-wide library card holders. This environment coupled with technological challenges (Y2K compliance, computer performance ceilings, and the continual need to work around vendor-supported software) incubated a desire to build their own integrated library system. After facilitating numerous focus group interviews and building consensus surrounding library policies such as the lending and returning of books, they wrote a cataloging, user-interface, and circulation modules using commodity hardware and open source software. While the impetuous for Evergreen began prior to the year 2000, I believe the actual development time was less than two years. When it was all said and done they were able to outline a bit of a cost-comparison regarding their implementation:
|support||free for 3 years||$200,000/year|
|staff||4 people||2 people|
In other words, the folks of PINES reallocated their dollars shifting it away from vendor support and invested it into staff, saving money all along the way.
Some of their articulated advantages, besides money, of writing and maintaining their own integrated library system include:
- more control over their own destiny
- more ability to set priorities
- more direct access to computer/technical staff
Their demonstration of Evergreen was quite impressive. Simple. Elegant. Employed modern technologies. They were particularly proud of their "book bag", "shelf browser", and links to Galileo (a state-wide set of licensed bibliographic indexes) features. Technically speaking, much of their implementation employed SOA, as described by Murray. Future developments include:
- acquisitions and serials modules
- a "children's portal"
- online bill paying
- more "social Web" features
- a complete Spanish translation
- deep links to Galileo
- access via mobile devices
- NCIP (a circulation interchange protocol) implementation
In short, the Evergreen project can easily be described as a success. Development happened on time and under budget. Follow-up focus group interviews and surveys have been very positive. Use of the public library system in Georgia is increasing. The folks of PINES are well on their way to remaining relevant in our increasingly networked environment. Instead of outsourcing their bread & butter activities they have taken the bull by the horns and retained control over their own computing environment. Kudos!
The last formal presentation was given by Alan Darnell (Ontario Scholars Portal) called "Welcoming the prodigal child: Integrating e-resources and print resources in the next generation OPAC/ILS". In it he shared how he is combining the content of his traditional library catalog (metadata records describing print resources) with the full-text of 10 million articles from 7,500 journals and the citations from 130 abstracting and indexing databases. He has been able to do this exploration because his hosting institution does not just license access to these additional materials but licenses the content itself. They do this for a number of reasons: 1) archiving and preservation, 2) ease of access, 3) the desire to "capture the conversation" of scholarly communication. Darnell compared the journal literature to the prodigal child. A child that goes away, tries to become its own person, yet returns and desires to be a part of the family again while the stay-at-home child complains about the "fun" its sibling had while out and about. He elaborated by comparing and contrasting the content of the traditional OPAC with e-resource (A&I) systems:
|relies on authority control||exploits relevance ranking|
|describes the whole item||built from component parts|
|contains surrogate data||contains digital objects|
(It would be nice if the profession were able to take the useful characteristics of both environments and combine them into something whose whole is greater than the parts. Open content. Relevancy ranking. Digital objects. Authority control.)
Once he acquires his content he converts it into different flavors of XML (MARCXML, etc.), stores it, indexes it, provides access to the index, and "mashes-up" the results create thing going beyond lists of search results. Pictures from here. Reviews from there. Annotations. Citation lists of similar items. His ability to perform these functions is all premised on his direct access to the content. "I don't need your interface. Just give me the data."
The symposium was brought to a close through a panel discussion. In it quite a number of action items and/or next steps were articulated. Many of them are listed below but in no priority order:
- allow yourself to take risks
- convince vendors to add value to projects without breaking standards
- create forums for future discussions through mailing lists, IFLA, or other professional associations
- describe projects in a language the intended audience can understand
- foster a culture of innovation
- foster communities around PINES/Evergreen
- get more involved in the wider world to create standards, such as the interactivity of the OpenSearch and SRU communities
- integrate computer technology into the entire aspect of librarianship
- keep in mind the needs of users before the needs of libraries
- keep in the mind the purpose of libraries, and retain what worked
- know your contexts and its purpose
- look at the emotional as well as the logical aspects of projects
- permit library employees to play and explore issues
- prioritize the aspects of an ILS, break off the pieces, implement them, and repeat
- spend time coalescing computing skills across the library community
- step back, take a holistic view, and assess the problems that need to be solved
- understand that we have more things in common than differences
- use existing standards -- "Don't do what you don't know how to do."
- use measures like the Apache Foundation and apply them to library computing
- write a manifesto
- write business cases accounting for the costs of ILS's across institutions and then consider pooling resources
- symposium home page - http://infoservices.uwindsor.ca/ils/
- Art Ryhno's presentation - http://infoservices.uwindsor.ca/ils/sessions/ils/
- Peter Murray's presentation - http://dltj.org/2006/11/windsor-soa-presentation/
- SOA from Wikipedia - http://en.wikipedia.org/wiki/Service-oriented_architecture
- PINES/Evergreen - http://open-ils.org/
- Ontario Scholars Portal - http://www.scholarsportal.info/
- Peter Binkley's blog entry - http://www.wallandbinkley.com/quaedam/?p=86
- Rob Fox's travel log - http://www.library.nd.edu/daiad/fox/windsor_ils.shtml
Creator: Eric Lease Morgan <firstname.lastname@example.org>
Source: This text was published here first.
Date created: 2006-11-29
Date updated: 2006-12-03
Subject(s): next-generation library catalogs;