Bulletin, August/September 2006

Recent Developments in Electronic Resource Management in Libraries

by Rafal Kasprowski

Rafal Kasprowski is electronic resources coordinator at the University of Houston Libraries, Houston, TX 77204-2000 and can be reached by email at Rkasprowski<at>uh.edu and by phone at 713-743-9346.

At the 2005 Annual Meeting of ASIS&T, a panel composed of Ivy Anderson, representing Harvard University and the Digital Library Federation’s Electronic Resources Management Initiative (DLF ERMI) steering group; Barbara Weir, representing the Tri-College Consortium (Bryn Mawr, Haverford and Swarthmore); and Ted Fons, representing Innovative Interfaces Inc., discussed the recent advent of electronic resource management (ERM) systems for libraries.

ERM Processes
The steps required for managing e-resources are more complex than those for print resources. A process that consisted of selecting, ordering, cataloging and binding now includes selecting, evaluating, approving, licensing, billing and registering access and is accompanied by a series of technical aspects, such as usability, performance, access technologies, public interfaces, troubleshooting and usage statistics. E-resource management is a time-intensive and iterative process, often requiring more staff with a greater skill set at each stage. As demand for e-resources grew, storing and managing administrative information in spreadsheets, paper files or email folders became increasingly cumbersome, and a more integrated solution was called for.

Home-grown ERM Systems
At first, several libraries began to develop their own e-resource management systems (e.g., Colorado Alliance, Johns Hopkins, Massachusetts Institute of Technology, Penn State, Tri-College Consortium) to integrate such elements as the relationships between packages and their constituent parts, use permissions and constraints, authentication and contact information, as well as workflows for trialing, licensing, ordering, implementing access and notifying relevant staff.

Enter DLF ERMI, an initiative composed of librarians and vendors, spearheaded by Tim Jewell from the University of Washington. In 2001, DLF began a project to “develop common specifications and tools for managing the license agreements, related administrative information and internal processes associated with collections of licensed electronic resources” (retrieved June 9, 2006, from www.library.cornell.edu/cts/elicensestudy/dlf2002fall/home.htm). Ultimately, ERMI’s goal was to foster the development of ERM systems by commercial vendors and promote best practices and standards. A final report was produced in August 2004. The key technical aspects addressed in the document were a set of functional requirements, an entity-relationship diagram, a data element dictionary, the data structure and XML (eXtensible Mark-up Language) schemas.

Functional Requirements
The functional requirements were developed from an initial collaboration between Harvard University, Massachusetts Institute of Technology (MIT) and the vendor Ex Libris to define the necessary functions of an ERM system. The resulting document covered the information and workflows that an ERM system should integrate in the key areas of selection and acquisition, access provision, resource administration, user support and troubleshooting, renewal and retention. The document contained data elements new to ERM systems at that time, relevant to both users (e.g., license permissions and restrictions, resource availability) and staff (e.g., detailed license information, administrative IDs and passwords, usage statistics, training information). Also addressed were problem-solving activities (e.g., troubleshooting, vendor communications) and business functions (e.g., pricing models, renewal and termination information).

ERD, Data Element Dictionary, Data Structure
The entity-relationship diagram (ERD) is a visual representation of the major entities needed for managing electronic resources and the relationships between them. It is the basis for the database structure of an ERM system. At the heart of the ERD is the relationship of e-resources to licenses.

In the proposed diagram, the electronic product, which is an e-journal package or an article database, is a set of hierarchical relationships between the interface, the resources that use the interface and the child resources that are part of the product. License and access data relate to any level of that hierarchy. For example, the access information may relate to the entire interface or to an individual resource, while the license information may also relate to the entire interface as well as to each constituent product. The terms, governed by the license or acquisitions entity, define the number of simultaneous users, the various participating sites and the license permissions; they may have to be combined into prevailing terms to correspond to specific situations.

While the ERD presents the relationships between the entities, the data element dictionary defines the system’s data elements. The data structure document in turn links each data element with the entities and relationships represented in the ERD. ERMI plans to combine the data dictionary and data structure in a future version.

XML Investigation
In its initial attempt to develop an ERM XML schema, ERMI decided to limit its proof of concept to licensing. Digital rights management (DRM) and rights expression language (REL) schemas are related to licensing schemas, but ultimately do not fit library needs. ERMI recommended that a native XML schema be used instead to express license terms.

The three member colleges of the Tri-College Consortium share an Innovative Interfaces ILS system, have a common approval plan and jointly negotiate purchases of e-resources. In 2001, the Tri-College Consortium first developed their electronic resources tracking system (ERTS) to serve as a central repository for administrative metadata such as license restrictions, technical contacts, statistics availability and payment information. A separate database was later developed for managing trial information and included resource descriptions, access information, pricing information, license information and a note field. The ultimate goal was to merge ERTS and the trials database into a single system that would hold data from the trial phase through purchasing and renewal.

VERIFY, Phase 1
Following a process redesign project in 2004 to reorganize the various existing e-resource workflows into a coherent whole, the Tri-College Consortium decided to partner with VTLS to design a new ERM. This integrated library system (ILS) vendor was chosen over others as its product targeted the particular ERM needs of library consortia. VTLS based its system, called VERIFY, on the newly released ERMI specifications and used the FRBR (IFLA’s Functional Requirements for Bibliographic Records model) to link all versions of print and electronic titles.

The team tried to map the fields from ERTS and the trials database to the fields defined by ERMI, but most of the ERMI fields had no match as they had not been needed or considered or were recorded in a different field type. Moreover, VERIFY entered packages and titles without defining parent-child relationships. Its relationship structure was often inadequate; the trial entity, for example, was invariably attached to the acquisition entity whether or not a resource was actually purchased by one of the colleges. By April 2005, VTLS realized that the FRBR-based structure was too rigid to support the complex relationships of consortially purchased e-resources and pursued a more flexible XML-based system.

DLF ERMI and Consortia
The DLF ERMI guide is intended mostly for individual libraries; the intricacies of consortial arrangements still need to be resolved by libraries and vendors. Entities that do not require multiple fields for single libraries often do so for consortia. For example, all three colleges may run a trial separately, but only one may actually buy the resource. The Electronic Resource Status field, defined by the DLF as non-repeatable, could actually be Active, Under Review and Rejected by any of the three colleges at one time. Any field should also be customizable locally and not limited to a prescribed set of values.

VERIFY, Phase 2
By May 2005 VTLS had organized each resource into several entities – electronic resource, package contents, acquisitions, licenses, access information, workflow and trial, separated by tabs – and was ready to test VERIFY. The test system still needed improvement: data added would be lost with upgrades, many fields did not match the ERMI specifications, and it was still impossible to attribute acquisitions data to a single college. At that time, the colleges also began investigating a method for importing data from the Harrasowitz OttoSerials system or from spreadsheets into VERIFY.

ERM Industry Response
Around the same time that libraries began building their own systems, ILS and non-ILS vendors alike studied library-driven functional requirements and ERMI reference documents to build commercial ERM systems. Innovative Interfaces benefited from its close collaboration with several libraries and ERMI’s early deliverables for the 2003 release of its ERM system, which was also designed as a stand-alone component with other ILS systems. Endeavour and Ex Libris followed with their ERM systems in 2005; VTLS and Sirsi/Dynix ERM products were still in development in late 2005. Similar efforts by non-ILS vendors are represented by the Colorado Alliance’s Gold Rush product, released in partnership with TLC (The Library Company) in 2003, and Serials Solutions’ system launched in 2005.

ERM System Technology
Vendors like Innovative Interfaces and VTLS built their systems as an integral part of their ILS platform. In the traditional integrated systems model, the ERM component is integrated directly with the ILS system. In other cases, application protocol interfaces mediate the integration with relevant ILS modules. User interfaces can be built for the Windows or browser environment, and the Web services technology allows integration across computer platforms. Non-integrated systems built by Ex Libris and Endeavour to work in conjunction with their ILS systems or stand-alone products developed by the Colorado Alliance and Serials Solutions provide competing solutions. At the center of any ERM system is a knowledge base, containing bibliographic, URL and holdings information.

Interpretability, Ambiguity, Adaptability
Currently licensing values are often left open to varying interpretations. ERMI developed a value schema to take into account the entire range of permissions, explicit or not: Permitted (explicit), Prohibited (explicit), Permitted (interpreted), Prohibited (interpreted), Silent (uninterpreted), Not Applicable. Rights expression languages (RELs) only support two values: Permitted (explicit) and Prohibited (interpreted). It is believed that any new licensing language would have to accommodate a certain level of ambiguity, which would require a novel encoding system.
ERMI plans to develop an ERM language that is open enough to accept new trends and address the needs of all relevant communities, such as cultural memory organizations and creative commons. In participation with the joint National Information Standards Organization (NISO)/DLF/EDItEUR License Expression Working Group (LEWG), it seeks to develop a single standard to support internal library management, as well as exchanges between publishers, libraries and various cultural communities. EDItEUR is the international group coordinating the development of electronic commerce in the books and serials sector. EDItEUR has become involved in the license expression work initiated by ERMI and is working on a license transmission standard that will be part of the ONIX (Online Information eXchange) for Serials format.

Future Releases of ERM Systems
In future releases, vendors plan to incorporate overlap analysis and cost-per-use analysis into their ERM systems to help with collection development. Integration with content modules, such as e resource registration processes, incident report mechanisms and license descriptions, is also on the agenda. Standardizing usage data collection is a major objective of the ERM industry, as represented by the Standardized Usage Statistics Harvesting Initiative (SUSHI), whose objective is to develop a standard for downloading COUNTER-compliant (Counting Online Usage of Networked Electronic Resources) usage statistics from vendor sites into ERM systems.

ERM System Interoperability
A long-term goal for e-resource management is ensuring interoperability between systems. Imported data could originate from several sources, e.g., acquisitions data from the library’s serial vendor, XML-formatted license terms from publishers, aggregator packages from e-journal management services, serial subscription and holdings data in the ONIX for Serials format from publishers. ERM system data describing downtimes or use restrictions could be exported to the library’s online public access catalog (OPAC), link-resolver menu or Web-based news blog.

Questions and Answers
Following the presentations, panelists addressed questions from the audience. Participants inquired about public interfaces, MARC (the MAchine Readable Cataloging encoding format) compatibility, content types, URL verification and staffing. ERM system developers are looking to integrate public interfaces providing information relevant to users and an open structure enabling library staff to build their own interfaces, with the existing backend interfaces designed for staff use. The MARC format is not necessarily incompatible with ER data, but XML-based standards represent the current trend in streamlining data exchange between ERM systems and OPACs. The library community is becoming familiar with the new format, and XML-based standards, like ONIX for Serials for presenting licensing rights and package level data, may prevail. Besides online resources, like e-journals, databases and e-books, ERM systems can also accommodate print journals. Multiple instances of the same journal may not be managed efficiently by an OPAC especially in a consortial environment. An ERM system can provide a single entry for all instances, as the knowledge base can integrate both the title level and package level data. Although library systems can provide reports of broken record links, companies like TDNet or SerialsSolutions also update links as part of their journal management service and are, therefore, best suited for providing URL verification. Finally, any staffing issues in the online age should cause a shift in responsibilities, not loss of work, as e-resource management, in general, and ERM system implementation, in particular, are work-intensive processes. 

This first ASIS&T session on e-resource management was well attended and generated much interest. A follow-up session on ERM standards is planned for the 2006 Annual Meeting.