Library Technology Guides

Document Repository

OCLC announces the demise of SiteSearch

Information Today [July/August 2001] The Systems Librarian


In a surprise announcement, OCLC announced on June 8 that it plans to discontinue development of its SiteSearch suite of products. The next release, scheduled for delivery in Fall of 2001 will be the last. OCLC will continue to provide technical support to licensed users through at least the end of 2002. OCLC has been actively marketing SiteSearch up till the time of the announcement. Installations of SiteSearch currently number about 100, including several state-wide consortial implementations.

Though the basic announcement has been made, OCLC is still in the process of working through many of the details that relate to this product and its support. Doug Loynes, OCLC SiteSearch Product Manager states that “OCLC is interested in having an active dialog with current SiteSearch users on how best to meet their needs. We have a primary interest in addressing the concerns of current customers.” A meeting of the OCLC SiteSearch User's Group was already planned for the ALA Annual Conference in San Francisco. Interests of customers expressed at that meeting and in ongoing conversations will help to inform OCLC on the support options it will offer past 2002.

The primary reason stated for the discontinuation of SiteSearch have to do with the allocation of development resources. OCLC has recently articulated a very ambitions strategic plan, Extending the Cooperative: A three year strategy (see: Asked about the relationship between this initiative and the discontinuation of SiteSearch, OCLC Vice President Frank Hermes stated that OCLC is considering a project that would seek to integrate and extend its suite of portal services – SiteSearch, WebExpress, and FirstSearch – to provide libraries with a single “portal service” platform. OCLC will evaluate the market for this new service and announce its intentions it this area at approximately the same time that SiteSearch version 4.2 is released.

According to the official announcement issued by Jay Jordan, President of OCLC, “…while we launched SiteSearch with high hopes, it has never achieved our expectations in revenues and number of users.” Hermes further stated that new sales for SiteSearch had declined in the last two to three years. In order to work toward the objectives described in this strategy, OCLC finds that it would not be able to develop both SiteSearch and its new products and services. Like any organization, OCLC must make decisions about where to put development resources, and SiteSearch was apparently judged to be expendable relative to others in the context of a very ambitions agenda it has set for itself in the next three years.

Though OCLC will discontinue the development of SiteSearch, it will complete its planned next release. This new 4.2.0 release of SiteSearch will include a number of new features, including a completely new search engine. SiteSearch users all know that OCLC often uses names of apple varieties to describe its technical components. The now-dated “Newton” search engine will be replaced by a new Java-based search engine called “Pears.” OCLC indicates that the Pears search engine will be released as Open Source. The plan to offer Pairs in this way was made even before the decision to discontinue SiteSearch.

Loynes indicated that OCLC is investigating and considering a number of options for ongoing support for SiteSearch. One of the options might be a “community sourcing” model, where existing licensees of the product gain access to the source code of the product and provide mutual support. OCLC recognizes that there is a very strong SiteSearch community. Many SiteSearch users have deep technical skills and have a solid understanding of the SiteSearch underpinnings. A purely Open Source model of support where SiteSearch source code is provide to the whole world for free seems to be unlikely. Loynes emphasized that OCLC will “provide as much ongoing support for SiteSearch as possible given the circumstances.” Again, final decisions have not been made pending additional conversations with SiteSearch customers.

What is SiteSearch?

For those not familiar with< the product, SiteSearch is a suite of tools designed to help libraries deliver integrated access to diverse information resources through a Web-based environment. The initial version of SiteSearch was developed in 1992. The product has grown steadily since, with about one major version released every two years.

Some of the major components of SiteSearch include:

  • A customizable Web-based interface for search and retrieval
  • WebZ, a gateway for searching Z39.50-based resources
  • Database Builder, a facility for creating local databases based on OCLC's Newton search engine.
  • Record Builder, a facility for creating database records individually, rather than importing them from external sources.

SiteSearch is not a turnkey application, but more of a set of tools that libraries can use to build a customized search and retrieval environment. Though SiteSearch comes with an “Out-of-the-Box” interface, libraries have great flexibility in creating a tailor-made search environment with a look and feel that suits their needs. SiteSearch-based applications can be created to search many different kinds of resources, primarily through Z 39.50. Non-Z39.50 systems can be integrated as long as they support SQL or some other programmable query method. Local content can be incorporated into SiteSearch through its Database Builder component. Record Builder is based on the Newton search engine that OCLC developed for its own databases, such as FirstSearch and WorldCat. In earlier version of SiteSearch, local data had to be prepared externally and loaded in batch, but beginning with version 4.0 the Record Builder module allows data to be created directly.

Earlier versions of SiteSearch included a module called the Image Support Package (ISP) that allowed libraries to create collections of digital images. This software never lived up to expectations and was eventually dropped from the SiteSearch Suite. Record Builder in conjunction with external scanning and digitizing applications provided a more flexible approach for building digital image collections.

Creating applications with the SiteSearch Suite requires a high level of technical proficiency. Beginning with version 4.0, SiteSearch was based on the Java programming language. The process of configuring the system and building applications with SiteSearch requires knowledge of Java. Programmers with experience in Java are a scarce commodity in libraries, and command high salaries. One of the primary challenges libraries face in operating a SiteSearch installation involves recruiting and retaining the technical personnel required to operate the system.

The character of SiteSearch as a set of tools, or a development environment, increases the impact that its demise will have on libraries that have implemented this product. Given its personnel requirements, any change in the status of SiteSearch as a product is broader than might be the case for other categories of software. Libraries must also invest in server hardware in order to use SiteSearch.

OCLC has recently released a related product called WebExpress. Though similar in features to SiteSearch, WebExpress is much more of a turnkey product that libraries can manage without local programming support. OCLC officials state that no plans are in place to discontinue marketing, sales, or support of WebExpress.

The Holy Grail: A Single User Interface

SiteSearch was in large part created to fulfill the vision of providing access to multiple and diverse information resources through a single user interface. One of the great inconveniences of information access before Web involved all the different user interfaces that one had to master to gain access to different types of information. Each flavor of online catalog, database, or full-text e-journal had its own look-and-feel and used different commands for searching. The idea of using a tool such as SiteSearch to channel many different resources through a single customized interface using Z39.50 was enormously appealing.

But as the Web evolved, the basic need for systems like SiteSearch declined. Three factors played against this approach: basically consistent native Web interfaces, inconsistencies in Z39.50 implementations, and the burden of local database creation.

All information resources find convergence through the Web. In earlier times the differences among user interfaces were significant and end-users complained of not being able to effectively navigate all the many different systems they needed to use in the course of their research. Today as all the major information resources, including Web-based online catalogs, have Web interfaces, most users have little trouble using them in their native forms. In the Web environment, similarities far exceed the differences in searching, presentation, and overall operation. Even inexperienced users seem to be able to navigate through most Web-based information resources.

The second factor has been the difficulty in practice of using Z39.50 to search across diverse systems. The differences in the way that Z39.50 servers have been implemented greatly complicates the process of getting consistent search results from different systems. The recent emergence of Bath and other Z39.50 profiles offer theoretical solutions, but will not bring practical relief for a few years when all the developers of Z39.50 servers deliver profile-compliant servers.

Finally, the relative advantages of loading information locally no longer seems to be as strong as in previous times. Some libraries, for example, purchased tapes of ProQuest data from UMI to load into their SiteSearch system so that it could be accessed through the same interface as other resources. It takes significantly more time and effort to load and index information locally than it does to point users to a vendors own implementation of that data. As the vendor's interface becomes as easy--or easier--to use than the one locally created with additional features, the motivations for taking on the extra burden of locally loading the information fall away. In the current environment, the loading of local data generally is advantageous only for unique collections that are not offered by commercial vendors.

To the extent that these factors have eroded the basic approach of channeling resources through a single user interface, the demand for products like SiteSearch has also waned. Hence, the low sales in recent years for this product.

SiteSearch implementations

Many libraries licensed SiteSearch and used it to created mission-critical applications. At least 100 library sites, mostly larger institutions, have licensed SiteSearch. While the number of sites that have licensed SiteSearch isn't enormous, the scope of library users affected is. The state-wide and large consortial resources that have been crafted using SiteSearch serve wide audiences. A small sampling of the major resources that have been created with this software include:

GALILEO (Georgia Library Learning Online), an ambitious state-wide environment that provides a integrated access to licensed content, library catalogs, databases of documents and manuscripts, and other resources. ( This project was the early flagship implementation of SiteSearch. The University System of Georgia uses SiteSearch to provide access through a single user interface to both locally mounted databases and remote databases through Z39.50. In its more recent form, many of the locally mounted databases previously offered through the SiteSearch-based interface have been replaced by links to the native interfaces. The URL for GALILEO is

The Committee for Institutional Cooperation (CIC) uses SiteSearch for their Virtual Electronic Catalog (CIC VEL). Participating institutions include the University of Chicago , University of Illinois at Chicago , University of Illinois at Urbana-Champaign , Indiana University , University of Iowa, University of Michigan, Michigan State University , University of Minnesota , Northwestern University , Ohio State, Penn State University, Purdue University, University of Wisconsin-Madison. ( The CIC entered into a two-phase joint development project with OCLC in 1997 to create the Virtual Electronic Library. In July 2000, the OCLC announced that Phase I of the project was complete and that Phase II has been suspended. The second phase was to provide significant enhancements to the VEL for interlibrary loan and document delivery.

The Kentucky Virtual Library uses SiteSearch provides a common interface and simultaneous searching across libraries throughout the state. ( ; Athena, a similar effort for 11 participating libraries in the Nashville area , and a locally created bibliographic databases of materials related to a Vanderbilt's unique collections of Charles Baudelaire and Gilbert Sigaux (

The University of Saskatchewan Libraries used SiteSearch to create its “Resources for Aboriginal Studies,” providing access to a set of nine related databases. ( and several other local resources.

The University of Tennessee uses the WebZ component of SiteSearch to provide a Web-based catalog for their epixtech Horizon system ( At the time that UTK implemented Horizon, they were able to develop a more functional Web-based online catalog with SiteSearch than was possible through the Web OPAC delivered with Horizon.

VALE New Jersey (Virtual Academic Library Environment) provides a common interface to databases, library, catalogs, and other resources to institutions of higher education throughout New Jersey. (

Geac licensed the pre-Java WebZ component of SiteSearch to form the basis of their GeoWeb catalog that they offer for both their PLUS and Advance systems. Geac continues to use an older version of WebZ for GeoCat and is self-sufficient in their support and is unlikely to be seriously affected by this announcement.

Other libraries that licensed SiteSearch include Carnegie Mellon University; CALNET Capital Area Library Network as part of the Library of Michigan; Central New York Library Resources Council; Cleveland Public Library; College Center for Library Automation in Florida; Colorado Virtual Library; Harvard University; Indiana: Inspire; Long Island Library Resources Council ; MnLINK; North Dakota; North Suburban Library System; Pennsylvania State University; Seton Hall University; South Central Regional Library Council; Southeastern New York Library Resources Council; University of Alabama; University or Arizona; University of California - Santa Barbara; University of California – Berkeley; University of Kentucky; University of Arizona; University System of Maryland; University of Michigan (MIRLYN) Missouri University of Saskatchewan; University of San Francisco; University of Wisconsin; University of Toronto; Washington Research Library Consortium (co-developer for ISP); Washtenaw Community College; University Virginia Tech University. Other international users include the SABINET in South Africa and the National Chiao-Tung University in Taiwan.

Implication for Existing SiteSearch Customers

OCLC's discontinuation of SiteSearch will prompt many of the libraries and consortia that currently rely this product to consider their options. Given the expectation that OCLC will continue support for at least another year and a half, no libraries are forced to discontinue current services, and have some time to develop long-term strategies.

Even without direct support from OCLC, most sites running SiteSearch will have the capability to maintain their existing environments indefinitely. Since SiteSearch requires one or more Java programmers on hand, and most are thoroughly familiar with the internal workings of the system, it is likely that each library will be able to support its own SiteSearch-based applications for some time.

Libraries will need to consider when or if they should re-deploy their existing SiteSearch-based applications in some other way and whether to use SiteSearch for new projects. It does not seem any urgency in abandoning any current efforts. Depending on the what final decisions are made regarding support for the product, many libraries will be able to sustain SiteSearch-based applications indefinitely. For libraries with the proficient development staff, the “community source” model of support may offer them benefits over the current support system from OCLC. Libraries considering SiteSearch for new projects would likely consider other options given these recent circumstances.

Libraries that license SiteSearch currently pay about 15 percent of their original license fee to OCLC for support. If the final outcome results in OCLC not providing support, then the libraries should be able to continue to operate their existing systems without ongoing support payments. For the whatever period that OCLC offers support—that is at least through the end of 2002—libraries will continue to pay their SiteSearch support fees. OCLC prices and licenses SiteSearch according to the size of the user base that will access resources created with the system (measured in FTE), by the number of simultaneous user sessions allowed, or a combination of these two factors. Libraries will be expected to continue to abide by their license restrictions for SiteSearch even after the product is discontinued.

Public Relations Issues

The discontinuation of SiteSearch will be a public relations issue for OCLC. This type of action is not one that instills confidence in OCLC as a software provider. Especially in a business sector as conservative as the library market, long product life and vendor reliability are key values.

This is not the first time that OCLC has retreated from a product line. In the mid 1980's, for example, OCLC became involved in the library automation business, marketing a system called LS/2000. In 1990 OCLC sold off its Local Systems Division to Ameritech. While Ameritech did provide support for the OCLC's former library automation systems, they did not further develop the systems. Ameritech, rater, went on to purchase Dynix and NOTIS, and developed the product now known as Horizon. In December of 1999 Ameritech Library Services became epixtech. Inc. Many of the libraries that had purchased OCLC's library automation systems expressed concern at the time that they had been negatively affected by OCLC shift out of the library automation business.

Any negative impact that the withdrawal of SiteSearch as a product will have on OCLC's public image will be mitigated by a amiable arrangement of support with the current users of this product. Also to the extent that OCLC's users in general resonate with the organization's new strategic plan, they will be sympathetic with the trade-offs associated with focusing on a new set of products and services.

It does seem that OCLC is working hard to back away from SiteSearch more gracefully than it did a decade ago with its Local Systems Division. While it is firm in its decision that it can not develop SiteSearch as a product in the long term, OCLC emphasizes its determination to present palatable options for the libraries that have made investments in this product. Only time will tell whether these efforts have been successful.

An organization as large as OCLC that manages a broad slate of products over a long history is bound to have at lease some products that fall by the way. The landscape is changing very rapidly, and OCLC would be doing libraries no favors by taking a slow-going approach.

SiteSearch at Vanderbilt

To measure the reaction of this announcement by SiteSearch customers, I have to look no further than my own workplace. The libraries at Vanderbilt University have been long-time users of SiteSearch.

We have taken advantage of the database builder module to create two different resources related to French Studies—one related to the 19th century writer Charles Baudelaire and the other related to a large collection of materials related to theatre collected by Gilbert Sigaux.

The WebZ component of SiteSearch is the basis for two virtual catalogs we have developed. Athena, developed by Vanderbilt for the Nashville Area Library Alliance includes eleven libraries in Nashville and supports expedited interlibrary loan among the participating organizations. Kudzu, named for the infamous vine that “covers the south” connects the major research libraries in the southeast. The Kudzu catalog though hosted by Vanderbilt is an initiative of the Association of Southeastern Research Libraries (ASERL). Both virtual catalogs have worked well and facilitate resource sharing initiatives among each group of libraries. The Kudzu catalog is relatively new. We have just completed a successful pilot and are planning a production implementation of the virtual catalog with expedited interlibrary loan delivery of materials among the participating institutions by the 2001 Fall semester. We have just ordered a new Sun server for our SiteSearch applications.

Although we surprised by the announcement, we haven't seen any reason to panic. Even though we have just made commitments to use SiteSearch for a new project, we still plan to go forward with these plans. Even as we made the commitment to begin the Kudzu project with SiteSearch, we already intended to evaluate options in the near future to investigate systems that might offer more advanced interlibrary loan automation and resource sharing features.

We are confident in our ability to support the services we have created with SiteSearch even without significant support from OCLC. Some of the support alternatives currently being discussed seem to be ones that will allow us to continue to manage our SiteSearch environment into the future.

The fact that the next release of SiteSearch will be the last has also not raised major alarm. There are not any major features that the system lacks for which we have expressed an urgent need. As a gateway to information and as a digital library toolkit, SiteSearch has grown into a reasonable level of maturity.

Like other libraries, we are eager to learn all the details and final arrangements that will be worked out between OCLC and the SiteSearch users. The upcoming SiteSearch User's Group at ALA should be interesting.

View Citation
Publication Year:2001
Type of Material:Article
Language English
Published in: Information Today
Publication Info:Volume 18 Number 07
Issue:July/August 2001
Publisher:Information Today, Inc.
Series: Systems Librarian
Place of Publication:Medford, NJ
Notes:Systems Librarian Column
Company: OCLC
Products: SiteSearch
Record Number:9273
Last Update:2024-02-23 02:26:00
Date Created:0000-00-00 00:00:00