CIL2008: The Open Source Landscape

This is the presentation I hoped to have in yesterday’s keynote

Marshall maintains a list of who has what catalogs on his Library Technology Guides site.
Federated search systems: LibraryFind; dbWiz (Simon Fraser); Masterkey (developed by Index Data). masterkey.indexdata.com for a demo.
OCLC offers some open source software — but not cutting edge stuff. Fedora is a major digital repository engine. VTLS Vital is based on Fedora. Fedora Commons is a support service around it. Keystone — also by Index Data.

Open Source Discovery Products (i.e, Next Generation Catalogs)

– VUFind. Apache Solr/Lucene.
– eXtensible Catalog (Mellon funded). Not a product now, but will be one day. XC are currently seeking institutional participation. This will “probably become a player” in the coming years.
– Others, such as Fac-Bac-OPAC, Scriblio (formerly WPopac).

Open Source in the ILS Arena

Shifting from open source being risky to open source being mainstream. Medium-sized public libraries are going with open source solutions for catalog; it no long requires massive technological effort or as much risk as it did.
In 2002, the open source ILS was a distant possibility — 3 of 4 tools Marshall reviewed then (Avanti, Pytheas, OpenBook, and Koha) are now defunct. In 2002, open source ILS wasn’t a trend.
In 2007, world starting to change. Slowly. A few hundred libraries had purchased an open source ILS; 40,000 had purchased a commercial product. In March 2008 — early adopters are now catalysts for others. There’s a small installed base, which makes others see the possibilities as being real. It seems now that we have a bona fide trend.
The ILS industry is “in turmoil”. Companies are merging; libraries are faced with fewer choices from commercial vendors; this gives more credence to ILS arena from standpoint of competition.
Decision to go open source is still primarily a business decision — as a library, need to demonstrated that the open source ILS best supports the mission of the library.

Current Product Options

Koha first open source ILS. Based on Perl, Apache, MySQL, Zebra search engine (from Info Data). Has 300+ libraries using it. Including Santa Cruz Public Library, 10 sites and 2 million volumes. Has relevance-ranked search, book jackets, facets, all that jazz.
Evergreen. Developed by Georgia Public Library consortium. Two year development cycle (6/2004 – 9/2006). A single shared environment shared by all libraries. One library card. Switched from SIRSI Unicorn. Succeeded in part because of standardization of policies across libraries (lending policies, etc.). Used in Georgia, British Columbia, Kent County (Maryland), and under consideration by a group of academic libraries in Canada. So far, only publics have adapted).
OPALS Open Source Automated Library System. Developed by Media Flex. Both installed ($250) and hosted ($170) services. Used by a consortium of K-12 schools in NY.
NextGenLib ILS designed for the developing world. 122 installations (India, Syria, Sudan, Cambodia). Originally closed, converted to open source in early 2008. More information from Library Technology.
Learning Access ILS. Designed for underserved rural public and tribal libraries — a turnkey solution. But may be defunct, according to Marshall. Built on an early version of Koha, but customized.

Open Source Business Front

Lots of companies offer a business plan to help support ILS software. Index Data, LibLime (Koha), Equinox (Pines), Care Affiliates; MediaFlex.
Duke is working on an open source ILS for higher education (looking for funding from Mellon; Marshall is involved).

Open Source Issues

Rise in interest led by disillusionment with traditional vendors. But total cost of ownership is probably about the same between open source and traditional tools. Libraries hope that they are less vulnerable to mergers and acquisitions. There’s no lump sum payment (though still need hardware, support — internal or external — and development costs. Not always clear who is funding the next generation of the current system.
Risk factors: dependency on community organizations and commercial companies. Decisions are often based on philosophical reasons, but they shouldn’t be — you need to consider the merits of the system itself. Make sure features and functionality are what you need.
Open Source vendors/providers need to develop and present their total cost of ownership — with documentation.
“Urgent need for a new generation of library automation designed for current and future-looking library missions and workflows.” That is, systems built for our digital and print collections. Open source tools do OK for systems of yesterday; will they meet the needs of the new library?
Q: How close are we to a system that does not utilize MARC records?
A: Not very. We need systems that do MARC, and Dublin Core, and ONYX, and RDF, etc., etc. The value in existing MARC records is too large to ditch. (Of course, it needs to be MARC XML.)