Forum Paper: Selection and Presentation of Commercially Available Electronic Resources:Issues and Practices

Selection and Presentation of Commercially Available Electronic Resources: Issues and Practices

Unedited draft for exclusive use of participants in the DLF Spring 2001 Forum

Timothy D. Jewell
University of Washington Libraries

Table of Contents

  1. Introduction
  2. Selection and presentation issues and practices
  3. Conclusion and future considerations
  4. Appendix A. Selection and presentation documents and web pages
  5. Appendix B. Functions and data elements for managing electronic resources

1. Introduction

Over the last decade libraries of all kinds have been spending larger and larger shares of their resource budgets to acquire or gain access to electronic resources from publishers and vendors. Ten years ago the user of a typical ARL library would have found little more than a handful of the more prominent periodical indexes and abstracts -- possibly in CD-ROM format. Today the same person would find a daunting array of resources that might include hundreds of databases and thousands of electronic journals.

One of the primary reasons for this shift in spending and emphasis is that electronic resources have enabled libraries to dramatically improve services in a variety of ways. First, most electronic resources come equipped with powerful search and retrieval tools that enable users to perform literature searches more efficiently and effectively than was previously possible. Since the majority of relevant electronic resources are now available via the web, users can now have the convenience of desktop access 24 hours a day. The added convenience of being able to navigate directly from indexing databases to the full text of the articles referenced, and to follow further links from there, is now becoming available. Needless to say, most users have welcomed these developments and have come to expect them as a matter of course. Nevertheless, libraries have faced and continue to face a number of technical, funding, and organizational challenges to bring these changes about. The goal of this report is to review some of these key issues and some of the more promising practices that libraries have devised - individually and collectively -- to deal with them.

1.1. Scope and Perspectives

The focus of the present report is on practices related to the selection and presentation of "commercially available electronic resources." However, as part of the DLF Collection Practices Initiative, it shares the goal of identifying and propagating practices that support the growth of "sustainable and scaleable" collections. For what follows, it is important to briefly discuss how these terms are being interpreted.

In an attempt to be as inclusive as possible, commercially available resources are understood to extend to virtually any electronic product or service for which libraries spend resource funds. This is a very broad definition, since it encompasses even more types of resources than those listed by Demas a few years ago (Kovacs 2000, p. 11): directories; dictionaries; abstracts, indexes and table of contents services; encyclopedias and almanacs; e-serials; bibliographies and bibliographic databases; and "key primary documents." To this list must now be added a few new "genres," such as "aggregator" databases that combine features of indexes and e-serials, historical full-text collections, and electronic books.

It is necessary to turn for a moment to the conditions that might enable commercially available collections or resources to be "sustainable." In current discourse, librarians tend to characterize pricing models, or the information marketplace in general, as sustainable or not, and this is the sense in which the term is used in a couple of recent, influential documents. For example, a few years ago the International Coalition of Library Consortia issued a document [reference] that included this statement: "current pricing models for e-information, which are developing during a period of experimentation, are not sustainable." The more recent "Tempe Principles for Emerging Systems of Scholarly Publishing" [reference] employs the term similarly but goes further, declaring that the current system of scholarly publishing and communication " . . . has become too costly for the academic community to sustain." Vendors and providers of electronic resources are also be concerned about the economic sustainability or viability of their products, and their views may be in conflict or tension with what libraries view as the realities of their funding situations.

Since pricing issues are so fundamental to sustainability, some attention will be paid in this report to some emerging strategies for exerting economic pressure within the marketplace for electronic resources. However, since substantial staff time and effort may be necessary to acquire and effectively provide access to databases and e-journals, sustainability is also an operational question. With that in mind, much of the report focuses on operational and organizational issues and practices. Nevertheless, it is important to recognize the influence that vendor design and presentation choices have on the amount and kind of time and effort required of libraries. It is assumed that both libraries and vendors have important roles to play in fostering an environment with resources and services that are sustainable for both communities.

These broad definitions of commercially available resources and of sustainability affected the scope of the study. Most importantly, it limited the number of libraries whose internal policies and practices could be surveyed, and the way in which that could be done. Since problems of scale, or size of collection, are among the key concerns of this report series, it made sense to focus on libraries having relatively large collections of commercially available electronic resources. The member libraries of the Digital Library Federation generally fit that description and number fewer than 30. Since they could also be communicated with relatively easily and asked to respond sample of libraries to focus on. Although primary attention was paid to this group, an attempt was made to look beyond them when trying to identify useful practices, and it is hoped that the results of the research will be of interest beyond the DLF membership.

1.2. Methodology, Aims, and Organization

The foregoing considerations formed the starting point for an extensive review of internal documents and local practices that started with two recent and very useful ARL SPEC Kits: numbers 248 (on Managing the Licensing of Electronic Products) and 253 (on Networked Information Resources). The documents reproduced in these publications often pointed the way to additional documents and discoveries, or led to telephone conversations and e-mails with librarians involved in the practices discussed or depicted. The public Web sites of DLF member libraries and several other ARL libraries with large collections of electronic resources were also reviewed with some care. This led to further contacts, conversations, internal documents, and ideas.

The main body of the report presents ten topical discussions of issues and practices related to selection and presentation. The first of these is titled "The Economic Context of Electronic Resource Selection," and it briefly discusses the two most visible means libraries have so far developed for dealing with the pricing of electronic resources: consortial purchasing and alternative scholarly communication initiatives. The next two sections treat selection policies, strategic plans, and organizational matters. The remaining sections deal more directly with operational questions, and are organized roughly in a "resource life cycle" sequence. That is, it starts with initial selection issues and proceeds through ordering and purchasing, establishing and organizing access, providing support, and ongoing evaluation or assessment. It concludes with a discussion of some local databases and systems devised to support or help rationalize the treatment of electronic resources. It is hoped that the information provided will help libraries define standard functions and data elements, and otherwise collaborate in developing improved systems for supporting the acquisition and maintenance of licensed electronic resources.

The most promising general "effective practice" strategies are assembled and briefly discussed in the concluding section. Given the dynamic nature of electronic resources, many of these suggested practices may soon be outdated and irrelevant. Accordingly, the goal is to assist in local decision-making, rather than to establish a set of formal standards. Because of their potential value to readers of the report, a table of links to local practice documents and related Web sites is provided in the Appendix.

2. Selection issues and practices

2.1. The Economic Context of Electronic Resource Selection

Over the last decade, ARL and DLF libraries have gradually increased both the amount and proportion of resource budgets spent for electronic resources [reference]. In each of the most recent 5 or 6 years an additional 1% of resource budgets has been devoted to electronic resources, and in 1998-1999 both ARL and DLF member libraries spent about 10% of resource budgets. More than 85% of such expenditures are for serials and represent ongoing commitments. If prices for electronic resources rise as rapidly as have serial prices in general, the shift in spending could lead to additional financial pressures for libraries. The apparent need to maintain print copies of electronic resources while electronic archiving agreements and technology are developed will add further pressures.

There seem to be good reasons to predict that e-resource expenditures will actually accelerate rapidly over the next few years. For example, more and more journals from established publishers are quickly becoming available on the web, and tens of thousands of electronic books are now available through a single new company: netLibrary. Questia, another new company, has begun to offer an extensive collection of electronic resources directly to users. Although it is too early to tell whether such new services will take hold, there is substantial concern in some libraries that they must respond to this competition to maintain credibility as important sources of information on their campuses. It seems likely that one such response to this new "competitive space" (Hughes 2000) will be for libraries to invest more heavily in electronic resources. Accordingly, libraries may be very different in another ten years.

As libraries spend more money acquiring electronic resources or access to them, there are growing concerns over the inter-related problems of vendor pricing and institutional finances. In particular, there is a common perception that individual libraries are at a decided disadvantage when acting alone in this environment -- and that collaborative effort is necessary. This is perceived to be especially true when they negotiate with large corporate entities. Two important and fairly distinct types of collaboration have begun to shape the electronic resource market hold down prices. The first of these is cooperative purchasing through library consortia, and the second is the development of alternative outlets for scholarly communication. Their respective impact over time on prices and the marketplace remain to be seen.

2.1.1 Consortial Purchasing and Pricing

To librarians involved in acquiring electronic resources the importance and impact of consortia and consortial buying have become obvious - both within the United States and around the world (footnote - reference to ICOLC, international representation, etc). Consortia have been formed for a variety of reasons and exhibit a number of similarities and differences. A number of them -- such as OhioLINK, VIVA, and the California Digital Library -- are state-based, and limited to academic libraries. Other important academic library consortia, such as the Committee on Institutional Cooperation, include members in multiple states within a region. The NorthEast Research Library (NERL) consortium also includes members from several states, but its members are primarily private universities. There are also sizable, state-based multi-type consortia - such as NC Live in North Carolina and INCOLSA in Indiana - as well as national consortia based in the UK, Canada, Australia, and elsewhere. Some of these originated with broad mandates to foster resource sharing through online catalogs, but they are relevant to the extent that they attempt to negotiate better pricing and other terms for electronic resources.

Consortial arrangements for databases can benefit libraries financially in various ways. One theme that is common to most arrangements is that as the buying group expands, prices are lowered. One of the more common examples of this is when a price "per Full Time Equivalent (FTE) student" is based on a sliding scale and is lowered as more FTE students are added to the consortial contract. Another important current approach taken by some journal publishers is to determine the scope and content of a consortial e-journal collection on the basis of library holdings within the consortium. Examples of this include the arrangements that OhioLINK and NERL have with Elsevier - where members have access to all available ScienceDirect titles. Since no individual library is likely to be able to subscribe to all of this publisher's titles on their own, even the best funded of them stand to gain access to considerable content in such scenarios. Smaller libraries, in particular, may be able to dramatically improve access to resources for relatively small amounts of money. Sanville (2000) has shown that this content is likely to be used when made available, and has calling into question the traditional assumptions on which librarians have made local journal selection decisions.

Consortial purchasing must be profitable for publishers and vendors to view them as "sustainable," and many publishers see in consortial arrangements opportunities to reduce marketing and invoicing costs. It has also been argued that consortial contracts for e-journal packages enable publishers to gain revenue from smaller institutions that could not otherwise afford to subscribe to any of their journals. Although consortial arrangements can provide libraries with much better pricing or other terms, those benefits must be weighed against the other costs of doing business consortially, which can include substantial staff time and some losses in local control and flexibility. An added complication is that libraries can be formal or informal members of a number of different and sometimes competing or overlapping consortia and buying groups. Deciding which arrangements make the most sense in a given situation may be difficult.

An examination of public consortia member lists and other sources indicates that most DLF libraries are members of multiple consortia. As indicated by the consortial memberships listed in Table 1, they are most likely to be members of one of a handful of regional consortia, including the CIC, NERL, ASERL, or the Big 12 Plus. In most cases, they also belong to at least one major state-based consortium - and often more than one. For example, Penn State is simultaneously a member of the regional CIC and two state-based groups: PALCI and Palinet. As noted, though, many group-buying arrangements seem to be ad hoc and difficult to identify. As a consequence, it seems likely that the typical DLF library (and by extension, ARL library) will be involved in many different collaborative buying arrangements with varying degrees of formalization.

Table 1. DLF Member Libraries' Consortial Memberships
DLF Member Institution Regional Consortia State Consortia
University of California, Berkeley CDL
Carnegie Mellon University NERL (affiliate) PALCI; PALINET
University of Chicago CIC
Columbia University NERL
Emory University ASERL GALILEO
Harvard University NERL
University of Illinois at Urbana-Champagne CIC ILLNET; IDAL; ILSCSO; IICMP
University of Indiana CIC INCOLSA
Library of Congress
University of Michigan CIC Michigan Library Consortia
University of Minnesota CIC MINITEX
New York Public Library
North Carolina State University ASERL NCLive; TRLN
Pennsylvania State University CIC PALCI; PALINET
University of Pennsylvania NERL PALCI
Princeton University NERL
University of Southern California Big 12 Plus SCELC
University of Tennessee ASERL
University of Texas at Austin Big 12 Plus TexShare
University of Virginia ASERL VIVA
University of Washington Big 12 Plus Washington Cooperative Library Project
Yale University NERL, NELINET

ASERL = Association of Southeastern Research Libraries
CDL = California Digital Library
CIC = Center for Institutional Cooperation
NERL = NorthEastern Research Libraries
SCELC = Statewide California Electronic Library Consortium

An attempt was made to gain a general impression of the extent to which these libraries are currently purchasing selected databases and e-journal access via consortia by checking the Web sites of member libraries for a limited number of expensive, "big ticket" databases and selected e-journal packages. Appropriate consortia web sites were then checked to determine whether there seemed to be a consortial role, and further contact was made with a few key individuals to double check the information. It is important to note that no claim to the complete accuracy of this process is being made.

Several of the databases selected for this review were prominent general-interest and business-related full text aggregator services from several companies, including Academic Universe from Lexis-Nexis; EBSCO's, Gale's, and Bell & Howell's primary Academic and Business databases, and H.W. Wilson Company's "Wilson Select" and Omni products. To help reflect buying patterns in Scientific, Technical, and Medical (STM) fields, CAS' Scifinder Scholar and ISI's Web of Science products were chosen. Holdings of significant full text databases in the Humanities from Chadwyck-Healey and participation in the Early English Books Online project were also reviewed.

Of the full-text aggregator databases selected for this review, the most widely held among DLF libraries is Academic Universe, to which all of the academic members currently subscribe. This may not be too surprising, since over 6,000,000 FTE students nationally have access through a "national consortial" arrangement brokered by Solinet. The popularity of this service among academic libraries of all sizes is probably due in part to the vendor's adoption of a "sliding scale" fee structure described earlier -- through which the FTE price becomes increasingly attractive as more libraries subscribe. The situation with the other competing full-text aggregator databases offered by EBSCO, Gale, Proquest/Bell&Howell and H.W. Wilson Company is much less clear. The Proquest ABI/INFORM and Research Library databases are the most commonly held of these services (about half of the libraries subscribe to one or both of them), but consortial arrangements seemed to be involved only a quarter (or less) of those cases. Gale's Expanded Academic Index and Business databases are also relatively popular, but consortial buying appeared to be involved in only 4 cases. EBSCO's general academic and business databases were somewhat less popular, but statewide contracts appear to be the major factor where they were available. And, lastly, H. W. Wilson's databases appear rarely to be purchased by DLF libraries through consortia.

ISI's Web of Science is almost as popular among DLF libraries as Academic Universe, with 21 of the 22 academics offering it. Roughly 80% of the subscribing libraries were purchasing it through consortia - most through either the CIC or NERL. All eight of the nine DLF libraries having access to Scifinder Scholar subscribe through consortia -- but of those, six were buying through NERL. Fourteen DLF libraries provide significant full text databases from Chadwyck-Healey. Six of these libraries were buying through NERL. Seven libraries were found to subscribe to Early English Books Online (EEBO), of which five are NERL members. Consortial arrangements appear to be important factors in buying most of these databases - with the apparent exception of full-text aggregator databases.

To gain a sense the current role of consortia in the selection of electronic journals by DLF member libraries, a list of more than 20 prominent e-journal publishers and vendors was assembled. The for-profit publishers selected were Academic Press, Annual Reviews, Blackwell Science/Munksgaard, Elsevier, Kluwer, MCB Universities Press, Springer, and Wiley. Several university presses or university-based publishers or providers were also included: Cambridge University Press, Oxford University Press, Project Muse, and HighWire. In addition, a number of association publishers were included, such as the Association for Computing Machinery (ACM), the American Chemical Society (ACS) the American Mathematical Society, IEEE, the Royal Society of Chemistry, and the Society for Industrial and Applied Mathematics (SIAM). Lastly, the two first JSTOR collections (Arts and Sciences I and General Sciences) were also included, although they do not provide the current journal access that the others publishers do.

To determine which DLF libraries had online access to journals from these publishers and providers, their gateway e-journal lists and online catalogs were searched. Publisher and consortia web pages were also examined to see which DLF-relevant consortia might have signed agreements with which publishers. Because of the prominence of its Electronic Journal Center, OhioLINK was also included in the study and was contacted for a current list of e-journal packages available to its members. It was decided to consider only the access that was available at the end of December 2000.

A few tentative conclusions were drawn from this fairly informal investigation. First, many of the listed publishers were offering individual libraries "free electronic access with print subscriptions" at the time of study. Not surprisingly, most individual DLF libraries had taken advantage of those offers and were providing their users with such access. Second, consortial arrangements appeared to be somewhat less common and important for the not-for-profit publishers than the for-profit group.

To illustrate, almost all DLF libraries provided access to the two JSTOR collections, but largely because JSTOR does not currently provide special consortial pricing, only 3 libraries subscribed via consortia. Other publishers with relatively high rates of subscription (60% or more) but low rates of consortial adoption (i.e. 4 libraries or fewer) were ACM, the American Institute of Physics, SIAM, and Oxford University Press. The publishers showing similar rates of subscription, but higher rates of consortial adoption (i.e. 6 or more libraries) included two not-for-profits (Project Muse and American Chemical Society), and a number of for-profit publishers: Academic Press Ideal, Annual Reviews, Elsevier, Springer, and Wiley. Less than half of these libraries appeared to have online access to journals from two other for-profit publishers (Kluwer and Blackwell Science). However consortial arrangements were generally found to exist for the Kluwer subscribers, and almost always for Blackwell Science.

2.1.2. Scholarly Communication Reform Initiatives

Although the consortial movement appears to have had a significant effect on prices, the recent development of Web-based alternative outlets for scholarly communication may also play a significant role in shaping the market for scholarly information. One of the earliest and most successful of these alternatives is the physics preprint server at the Los Alamos National Laboratories, now known as arXiv [reference], which is provided free of charge and serves researchers in a wide variety of physical science disciplines. Within a relatively short time, it has quickly become an important part of the research culture in several of these disciplines. For example, a recent study of the 20 most cited sources in Chemical Abstracts [reference] revealed that the 4th, 8th, and 16th most-cited sources are (respectively) the arXiv preprint archives for High Energy Physics, Condensed Matter, and Astrophysics. The success of this initiative has inspired other efforts to make research articles freely via the web, such as Pubmed Central and the Public Library of Science.

The Association of Research Libraries has organized several initiatives aimed at addressing the problem. Most recently, these efforts have focused on three complementary programs. The first of these is SPARC (the Scholarly Publishing and Resources Coalition; references) - which has helped fund the development of a number of electronically based scientific journals and resources. Since one of the main goals of SPARC is to foster competition in the STM journal market, some SPARC journals have been developed to compete directly with specific journals deemed to have especially high prices. SPARC's related Create Change and Declaring Independence initiatives are aimed at raising journal editors' awareness of pricing issues, and at encouraging them to take remedial actions. Suggested strategies include negotiating with publishers about pricing policies, or starting competing journals if pricing negotiations prove unsuccessful.

There have been indications that some of these efforts have resulted in success. For example, the American Association of Physical Anthropologists reported that the publisher of its journal agreed to significantly reduce its subscription price after lengthy discussions and negotiations [reference]. The SPARC initiative itself appears to be well received among the ARL membership, and most DLF member libraries have supported it by becoming founding members or initiating SPARC-affiliated projects, such as Columbia's Earthscape, Cornell's Project Euclid, and the California Digital Library's e-scholarship initiative. Nevertheless, the program is not free of controversy. For instance, librarians sometimes complain that since faculty may not accept the cancellation of established but expensive titles or the packages that include them, funds must now be found both for SPARC titles and those they compete against.

Although there seem to be good reasons for libraries to support both the consortial and scholarly communication reform movements, a number of recent articles and responses to them have pointed to a tension between these approaches. For example, Landesman and Van Reenen (2000) have posited that there is a natural affinity or "congruence" between consortia and large publishers or vendors. They both are looking for the economies of scale that are represented by large packages, and although these packages typically do provide subscribers with more content for their money than their individual print subscriptions used to, the large dollar commitments required can quickly claim larger and larger shares of scarce resource dollars. These commitments will appear more justified as usage climbs, due to visibility and easy availability. When that happens, there will be less money to pay for the individual titles and smaller packages that consortia are poorly equipped or unprepared to handle - which will tend to be the more modestly priced offerings from the scholarly associations and university presses. In a similar vein, Frazier (2001) has warned of the dangers of the "Big Deal" - high cost packages of e-journals from for-profit publishers. In response, supporters of such package deals (Sanville et al. 2001, Mulliner 2001) have emphasized the added content that may be made available, the importance of user choice, and stated that libraries do not surrender their ability to negotiate terms when they enter such agreements.

There seems to be much value in both the consortial and "reform" approaches to influencing the market, and thoughtful people might reasonably lean in one or the other direction. However, as Landesman and Van Reenen (2000) suggest, it does seem that consortia could help assure the success of SPARC-like initiatives if those initiatives could be made more "consortia friendly." Bundling significant amounts of content, as the SPARC BioOne initiative is currently doing, seems to be one key way to addressing the need for economies of scale. As noted earlier, some for-profit publishers have also found that consortial contracts can provide them with new income from smaller institutions if the contracts provide those institutions with substantial content for a small amount of money. In addition to supporting them financially, consortia could also facilitate faculty education efforts and help create alternatives to established publishing outlets.

2.2. Selection Policies and Strategic Plans

In the midst of high user expectations, rapid change, and competitive pressures, it is not surprising that many librarians characterize local selection decisions as ad hoc, opportunity-driven." Such decisions are also sometimes described as having strong political elements - which may be especially apt when the purchase of one expensive service that serves a particular group rules out the acquisition of another that a different constituency may want or feel it needs. Many libraries have tried to overcome such tendencies and perceptions and give shape to their licensed digital collections through the writing and adoption of formalized local policies and plans. It seems clear that such documents should reflect and support the differing missions of these libraries and their parent institutions, and several do so - if sometimes in a rather general way. For instance, MIT's brief document entitled "Toward a Networked Resource Policy" [reference] begins by noting that the library "seeks to acquire access to the primary digital information resources which support the educational and research missions of the Institute." A similar statement appears in USC's "Collection Policy Statement for Information in Electronic Formats." [reference]

A general sense of institutional mission may or may not prove helpful in making specific "real life" decisions, though. Those libraries that are funded to serve broader state, regional, or national constituencies have special dilemmas to address in the digital environment. For example, the Library of Congress has traditionally served an archival function for printed materials, and that has been and will be more difficult when these materials are both digital and licensed [reference]. Similarly, the libraries of large, state-funded "flagship" universities are often expected to share their resources with other libraries or individuals in their states. When subscriptions to scholarly e-journals come with significant license-based restrictions or prohibitions on use for Interlibrary Loan, the institution can find itself caught in a new way between the priorities of local users and other groups.

2.2.1. Selection Policies

Perhaps the single most common thread running through the selection policy documents sampled for this project is that although electronic resources raise some new and different questions, the value system brought to bear on selecting more traditional resources is still valid. That idea was promulgated several years ago in Demas' (1994, 1995) descriptions of efforts to "mainstream" selection of electronic resources at Cornell, and is also expressed in policy documents in use at Penn State [reference] The California Digital Library [reference], and the Library of Congress [reference]. The University of Texas' "General Libraries Digital Collection Development Framework" elaborates on this notion, noting that

As with all formats, digital material should meet the same subject, chronological, geographical, language and other guidelines as outlined in the library's subject collection policies; and possess the same standards of excellence, comprehensiveness, and authority that the library expects from all of its acquisitions. [reference]

As was suggested regarding a sense of institutional mission, it may be difficult to translate such general considerations -- well stated as they are -- directly into decision-making. Several selection policy documents also refer to the need to support programs, or for a constituency to be served by a given resource. Other interesting considerations mentioned in some of the documents include the perceived need to maintain a balance among disciplines or subject areas, or with traditional formats, when choosing electronic resources. And, of course, there need to be identifiable and strong reasons for selecting electronic access over print or other formats, and the Collection Framework document adopted by the California Digital Library points to six pertinent factors. Among them are the potential added values of greater timeliness, more extensive content, and greater functionality or access.

Selection guidelines also typically list and discuss factors unique to electronic resources. In some cases these are presented in checklist form, with questions grouped together. Yale's "Examining Networked Resources" checklists [reference] are unusually thorough -- extending to 13 pages - but have much in common with those in use elsewhere. The Database Selection Criteria in use by the California Digital Library [reference] introduces an interesting refinement to the checklist approach by differentiating between non-content factors deemed "critical" and those viewed only as "important." Several of the following topical categories are drawn from that list:

2.2.2. Strategic Plans

An important trend at several institutions is to more clearly and consciously project selection decisions into the future, and to begin looking at and strategizing about the acquisition of different types of resources. For example, the Texas "Framework" document mentioned earlier incorporates strategic considerations in its section on "Observations and Qualifications" concerning different categories of resources, such as electronic journals, indexing and abstracting databases, etc. These brief sections discuss the context for each type of resource and attempt to delineate goals.

Several other DLF member libraries have gone substantially further in strategic planning for electronic resources, and there are interesting differences in their respective planning timeframes. For example, the time span covered by Cornell's "Digital Futures Plan" [reference] is the current two years, and seems especially valuable because it targets specific planned actions or outcomes. Virginia's "Library of Tomorrow" plan [reference] is somewhat more visionary, as appropriate for its five-year scope. Carnegie-Mellon's "Digital Library Plan" is for an even longer period of time (7 years), and presents three progressively more ambitious levels of project development for different levels of funding: "steady state," "higher profile," and "leading." The more open-ended University of Illinois Library Electronic Collections Plan [reference] is notable because it mentions the challenge licensing poses to its traditional role as a resource for other libraries in the state - a concern that is probably shared by many state-funded DLF and ARL libraries.

The projected futures articulated by these plans and the issues identified in them are somewhat different, but like selection policy documents, they highlight a "core set" of concerns. These might serve as useful starting points for other libraries interested in forming their own strategic plans for electronic resources - possibly including those to be digitized locally or selected from among freely available Web sites

2.3. Institutional Finance and Organization

The discussion of costs and expenditures thus far has focused on subscription and purchase prices, and on how collaborative action through consortia and publishing initiatives can help control those prices or mitigate their rates of increase. As is widely recognized, the prices of electronic resources represent only a part of the larger cost picture.

The University of Michigan's 1997 Electronic Resources Task Force Report [reference] offers a useful list of some "non-content" costs, and shows how those costs might be distributed within the local context. As with non-electronic resources, there are acquisition and processing costs, and the costs of providing "intellectual access." The latter costs used to be attributable strictly to cataloging, but since electronic resources are now also commonly presented through gateway lists, there are now also costs of adding and maintaining list entries. Montgomery and Sparks' interesting analysis of Drexel University's moves toward a completely electronic journal collection [reference] found that the costs of acquiring and providing intellectual access to electronic journals were actually higher than for other formats. The Michigan report also identified Access Systems (including interface design, application development, server capacity, etc.) and Library Infrastructure (workstations and connectivity) as requiring added funding. And, of course, there is the need for what the Michigan report identified as User Support. Drexel referred to this as Information Services, and determined that increased staff time was required for reference support and instruction, preparation of documentation, and selection.

The decision to acquire a given resource must somehow take these costs and ramifications into account. As noted in the preceding section on consortia and publishing initiatives, important general questions of policy, strategy, and mission may also be involved. It is not surprising that libraries have recently been experimenting with different organizational approaches to these questions. One problem that some of these efforts attempt to address is managing the apparent tension between centralized and decentralized decision-making models. On one hand, a number of institutions have made significant amounts of money available to centralized budget lines. This seems particularly appropriate when resources are heavily interdisciplinary or multi-disciplinary nature, when large amounts of resource funds are involved, or when there is a contractual requirement to maintain a certain level of expenditure with a given publisher. On the other hand, many institutions also believe it important for subject selectors or specialists to have the leeway to spend non-central or subject funds for electronic resources without having to go through multiple levels of review and approval. These considerations have prompted some institutions to write guidelines on how different categories of funds should be used, who should be involved in decision-making, and so on - and to the development of new committee structures.

A number of interesting but fairly similar organizational models rely heavily for decision-making on a specially constituted, broadly representative group. For example, Yale's Collection Development Council has reporting to it a Committee for Digital General Resources (CoDGeR) [reference], which consists of nearly a dozen members who represent various subject or functional areas. Harvard's Committee on Electronic Reference Services [reference] includes well over a dozen representatives from a wide range of libraries but, in contrast to the Yale model, reports to the Libraries' Public Services Committee. The California Digital Library has a Joint Steering Committee for Shared Collections, which, again, has representatives from all of the UC campuses.

A couple of institutions also involve subject-oriented groups in discussion and decision-making. For example, since the 1980's Stanford has had an Access to Information Committee reporting to its Collection Development Officer. It now has added three "resource groups": for Humanities; Social Sciences/Government Publications; and Sciences and Engineering. In addition to evaluating resources within their subject areas, each of these groups is expected to maintain contact with the libraries' Systems department. Michigan also has an overall steering committee (called the "E-team"), plus a "core resources" group and several other groups that focus on broad subject areas, such as Humanities or Science. How these arrangements actually work in practice and how tied they may be to a particular setting and context is an interesting question that could not be pursued during the course of this project.

Aside from the absorption of a number of related tasks by many staff throughout these libraries, there appear to be two fairly distinct and interesting trends with respect to the way in which job tasks are distributed. The first of these is the trend toward having single individuals responsible for overall coordination or orchestration of electronic resource purchases. Examples of these positions among the DLF member libraries and elsewhere are abundant. For example, Stanford has recently defined and filled the position of Digital Program Officer - who reports to the Collection Development Officer. Harvard has the position of Coordinator for Digital Acquisitions, who reports to the Systems Department, and MIT an Assistant Acquisitions Librarian for Digital Resources. The California Digital Library has a Director of Shared Content.

One of the dilemmas for electronic resource coordinator positions is the apparently limitless scope of possible responsibility. Several of these institutions have also been developing ways of distributing some of the tasks. The CDL, for example, has defined what it calls Resource Liaisons who are responsible for monitoring the "technical and content performance" of major products or product groupings, for reviewing and compiling usage data, and for identifying and communicating "enhancement requests and performance failure reports" to vendors. Harvard has a somewhat similar Resource Stewardship Program, with a "stewardship" role comparable to the CDL's Resource Liaisons. Similarly, MIT has defined the role of Product Sponsor, and Yale maintains a "list of contacts for electronic resources."

2.4. Internal Procedures for Initial Evaluation and Purchase

Many larger libraries have recently invested substantial time and effort in trying to understand, document, streamline or rationalize, and communicate their local procedures for dealing with e-resource acquisition. In addition to clarifying local processes, the resulting documentation may be of significant value to other libraries trying to grapple with the same issues. Loghry and Shannon (2000) provide a window onto many operational complexities in their discussion of workflows and forms devised for use at the University of Nevada - Reno. Although the California Digital Library is larger and more complex than most library systems at single institutions, its Acquisitions Procedures outline [reference] could serve as a helpful template for some of them, since it identifies a number of common steps. A couple of MIT's internal documents provide especially clear, helpful glimpses into their local process. The first of these is a unique, graphical "Process Map" [reference] that illustrates the process of acquiring electronic resources. The second is a detailed Workflow Proposal [reference] that deals with ordering, cataloging, management and maintenance.

Many different pieces of information must be gathered and organized as a particular acquisition makes its way through most local processes. As with the acquisition of more traditional materials, much of this information will be fairly standard - such as basic descriptive information, vendor, selector or selectors, fund or funds involved, etc. For resources accessed remotely via the web, the appropriate URL (or URLs), relevant usernames and passwords, the IP ranges from which they are available and system requirements will also be needed. The ability to determine the "status" of a resource within the overall selection process is also important, that can be a complex question. For example, libraries quite typically establish trials of resources that they are considering buying, and it is important to know when and to whom they are available for review, how and by whom input about them is to be received, and so on. Establishing funding for large expenditures can also take time, as can the process of working through licensing details and issues.

Owing to the complexity and range of these details, standardized forms for selectors and other staff have been developed and successfully put in place by a number of libraries. As is the case at Nevada-Reno, these forms may be used simply as templates for gathering appropriate information - which is then used in paper form or transferred more or less by hand to other systems or web pages. In large-scale operations it may be a challenge simply to track and effectively identify what resources may be under consideration or "in process" at any given time. A number of the larger libraries, including Harvard, Yale, the CIC, and the California Digital Library maintain "trial" or "status" web pages for staff and sometimes users. Although some of these pages continue to be edited "by hand," some other libraries have developed more automated approaches - which will be explored in section 3.7.

2.5. Licensing Issues and Practices

One of the biggest and most often-discussed changes facing libraries as they move toward increasing reliance on electronic resources is that they are typically governed by contract, rather than solely by copyright. The full significance and impact of this development on libraries has yet to be fully realized, but there are reasons for concern.

For example, while existing copyright law provides for the time-honored practice of interlibrary loan, licenses may forbid or so constrain it as to make it much more costly or otherwise impractical. As noted in the section on selection policies and strategic planning, such license provisions can also seriously undermine a library's ability to continue serving as a regional or state resource. Licenses may also seriously interfere with a library's archival roles and responsibilities if ongoing access rights are excluded or if there are severe restrictions on copying. There may also be prohibitions against user copying that are unenforceable from a library's perspective. Noncompliance by library staff or end-users may permit the vendor to discontinue the service without a refund for time remaining on the contract. In addition, the parent institution may be put at financial risk if its libraries are required to accept responsibility for user actions or if they agree to indemnify the provider against third party damage claims. And if, as frequently happens, a distant state or country's laws and courts are named as the contractual authority, substantial costs could be incurred by the institution.

Many of these issues are now widely recognized within the library community - partly as a result of a number of important initiatives that have been organized over the last few years to help educate librarians. For example, for several years now ARL has offered excellent training materials and classes for librarians wishing to upgrade their skills and institute organized local processes for dealing with licensing issues. The popular Liblicense web site and Liblicense-l listserv have also had considerable impact, since they facilitate ongoing discussions among librarians and publishers regarding licensing issues. Efforts toward the development of more favorable license terms and standardized language have also been important and influential. For example, the International Coalition of Library Consortia's 1998 Statement of Current Perspective and Preferred Practices for the Selection and Purchase of Electronic Information [reference] articulated a library-oriented viewpoint on a number of important licensing issues, including fair use and perpetual access rights, and liability for user actions. The concentrated purchasing power within that group has undoubtedly prompted some vendors to be more responsive to the concerns of potential buyers about licensing terms. Some consortia [reference] and individual libraries [reference] have clarified and strengthened their own bargaining positions by defining their own sets of standardized license terms or requirements. Additional impetus toward standardization has been provided by the development of the CLIR-Liblicense model license and software -- which libraries and publishers can use to efficiently arrive at mutually agreeable licensing terms.

Despite these noteworthy efforts libraries may still agree to terms that they regard as less than ideal. For example, if libraries are not required to accept responsibility for user behavior, they may still agree to make "reasonable efforts" to inform users of license terms. For the most part, library compliance efforts have tended to focus on making standardized disclaimers available and visible to users, such as the following statement found at the top of MIT's database and e-journal lists:

Use of many of these resources is governed by license agreements which restrict use to the MIT community and to individuals who use the MIT Libraries' facilities. It is the responsibility of each user to ensure that he or she uses these products only for individual, noncommercial use without systematically downloading, distributing, or retaining substantial portions of information. [reference]

Adoption and use of such a general disclaimer should be relatively easy for libraries, but since license details vary from one product or service to another, reliance on a general statement may be insufficient to make users aware of relevant terms. Some licenses require that subscribing libraries route their users through a "click-through" page containing some standardized language about use restrictions, and may also require users to certify that they are eligible to use the product. It seems impractical for libraries to consider investing the staff time required to write and maintain such pages for a large number of products.

If vendors do not provide rules of use or brief versions of license terms through their own web sites and services, libraries must devise and implement their own ways of tracking license terms and making them available both to users and staff. As will be seen later, several libraries have already done so. As interesting and promising as their efforts appear to be, there are still significant barriers to developing truly efficient systems for addressing this problem. One alternative that has recently been proposed is for libraries to share the effort of analyzing and reporting license details - much as cataloging effort is shared through bibliographic utilities. Libraries could then incorporate a license "profile" for a given product into appropriate gateway or catalog descriptions of the resource. This seems like an idea with potential, but since licenses for a product can vary from institution to institution, it could be difficult to establish the "canonical" version. The development of standardized definitions of key provisions could facilitate the growth of such a system.

2.6. Web Presentation Strategies

There are three general strategies for providing access to commercially available electronic resources: "local loading" by an individual library, loading by a consortium, and reliance on vendor sites. Individual libraries have generally been moving away from the local load strategy for cost considerations, as have consortia - though there are some notable exceptions, such as OhioLINK. Whether reliance on publisher and vendor services is truly the most cost effective strategy is an interesting question, since an argument can be made that it entails a range of hidden costs. Local loading continues to provide some important advantages - such as control over the number and functioning of user interfaces and normalization of usage measurement.

Effective presentation of electronic resources to users poses challenges for collections of locally digitized, commercially available, and "free" web resources, although each type of resource comes with its own set of distinctive issues and complications. The role of online catalogs is controversial because users accustomed to the Web expect to be able to locate resources through just a few mouse clicks. If they cannot, students will tend to rely even more extensively on web search engines than they currently appear to do. A number of institutions have documented that usage of specific resources jumps or declines quickly based on their gateway placement and visibility. [Document]

Larger libraries appear to follow a strategy in which resources are presented both through catalogs and e-resource gateways. These gateways typically provide alphabetical or subject lists of databases or electronic journals, along with an effective gateway search tool. Providing such multiple paths to resources in a cost-effective manner requires establishing synergy between cataloging and the methods used to generate lists, which is increasingly accomplished through standardized metadata.

A number of interesting alternatives to this general approach have recently been devised. For example, a system currently in use at the University of Pennsylvania [reference], organizes resources according to flexibly defined communities of interest. Bibliographers can identify the resources they believe to be the most critical to a particular user community, and new resources are brought to the attention of users through highlighting. The practice making new resources more visible has been integrated into web pages elsewhere - including the University of Texas [reference]. Other institutions (including N.C. State, the University of Washington, and the California Digital Library) have also introduced fairly simple but effective ways for users to "personalize" their views of available resources. Through them, once users have identified resources of particular interest, the selected resources are visible as a "default" when that user logs in to the local system.

Despite their value, the established and newer presentation strategies tend to treat e-resources as distinct entities that are located and then searched and used one at a time. However effective the presentation of distinct databases may be, users may find the effective integration of disparate resources to be just as important. One important step in that direction was the development of "broadcast search" functions - which were introduced first for traditional indexing and abstracting databases. Useful as those functions are, it is also important to integrate full text content with the indexing and abstracting services.

As suggested in the section on buying strategies, many libraries subscribe to one or more "aggregator" full text databases that may cost tens of thousands of dollars and contain full text coverage for several hundred to several thousand periodicals. Until recently, few libraries have been able to effectively and systematically make users aware of what periodicals are included in which aggregator databases. An important step toward making the content of these collections more generally visible and usable is the development of the open source community's jake (for Jointly Administered Knowledge Environment) initiative [reference] and the related jake2marc program [reference]. jake currently incorporates periodical holdings or coverage information for nearly 200 databases, enabling users to determine which database or databases include a given title. Institutions can customize jake to reflect their own holdings. The jake2marc service uses the jake database to enable libraries to generate catalog records for local use. Of considerable importance is the fact that these services are open to interested libraries, which has encouraged development and minimized usage.

Still greater user convenience is provided by systems that provide article-level links between abstracting and indexing databases on the one hand, and aggregator databases and e-journal collections on the other. A number of vendors have recently developed services aimed doing so. These offerings include PubMed's LinkOut, Silverplatter's Silverlinker, ISI's Web of Science linking feature, and OCLC's links between their databases and Electronic Collections Online (ECO). Similar services are also available from Cambridge Scientific Abstracts, OVID, and others. Though popular and beneficial, they each present drawbacks, such as the ability to link only to specific "islands" of content. For example, Silverlinker and ISI links feature offer links only to content from publishers with which these companies have agreements. Similarly, the OCLC linking utility only works for those e-journals that a library is accessing through the ECO service. This means that many links to available and locally licensed content cannot be presented, and differences in vendors' linking coverage can be confusing to users and staff alike. Most also require setup and ongoing maintenance, and since larger libraries often have subscriptions to abstracting and indexing databases from multiple vendor vendors, much of this effort must be duplicated.

Libraries clearly have reasons to want more universal, standardized solutions to the problem of providing such links to full text. One approach that has been recently introduced is Crossref; an industry-based initiative aimed at enabling article linkages across participating publishers [reference]. Although the initiative now includes a significant number of for profit and nonprofit publishers, it has significant drawbacks, perhaps chief among them being libraries' inability to control which links are enabled. This can be an important issue, since publishers can and sometimes do decide to provide Crossref links only for their "premiere" e-journal services - meaning that links to less-costly alternatives from a given publisher may not work. From the perspectives of many libraries, what is needed is a single utility that can be used to establish and maintain full text links for indexing databases from various publishers, and which provides them with control over which specific links are to be enabled. Of the alternative solutions proposed for this problem, one that appears to be quite promising is an approach based on "open URLs" - most visibly offered via the SFX service [references] now marketed through Ex Libris. In addition to local choice and control of links, SFX offers libraries the ability to administer and maintain them for multiple vendor offerings without duplication of effort.

Somewhat more visionary is the idea of establishing a broader "scholars' portal" (Campbell 2000) or commons that could be searched more or less like such established web search engines as Yahoo, Alta Vista, Google, and others. The advantage of such an approach lies in combining convenient access with appropriately "vetted" academic content. Practical means toward achieving such a vision may be available through metadata harvesting - a method that is now being investigated and developed through the Open Archives Initiative [reference]. However, the viability of the scholars' portal idea may depend on the availability of consistent and comparable metadata - which may be difficult to achieve.

2.7. User Support

As Libraries have made increasing investments in electronic resources, the problem of providing ongoing support has become larger and more complex. For example, as libraries rely more heavily on access through multiple remote vendors, interfaces and potential points of failure have proliferated. As indexing services and e-journal collections become more closely linked, such "pieced together" or eclectic systems present additional possible points of failure. Any number of problems may arise for users accessing licensed resources remotely. For example, a particular service may suddenly become slow to respond or unavailable due to technical problems, web browser configuration or unrelated connectivity issues, or to invoicing and payment problems. Users coming to a library's gateway through a commercial Internet Service Provider may also find services to be unavailable to them and not realize that this may be because they have not authenticated themselves through an available proxy service.

Most users of web-based services want and expect these services to be understandable and usable with a minimum of help or intervention from others, but when support is needed, expectations are high. Among the challenges is that -- as other web-based businesses and services have come to realize -- users may wish to make use of a consortium's or library's gateway at any time of day, and may expect "live" support at those times. Just what support may be needed and how it is to be provided are continuing questions, and libraries have developed a range of responses to them based on local perceptions, available resources and priorities. Of course, users need to know what services are available, who is eligible to use them, and how to connect to them. These basic needs have typically been addressed through the design of gateway web pages and e-resource lists mentioned in the previous sections. Slightly more elaborate approaches incorporate a basic "how to use" instructional element with information about how to make a connection [reference].

Connectivity and other technical problems are more difficult and seem likely to require ongoing efforts of various kinds. Among the simpler tactics have been to post lists of known problems and solutions for users [reference], make available browser "configuration" pages that inform users if they need to adjust settings [reference], and to provide information on how to contact staff for help. Some libraries have begun to experiment with "online chat" functions for reference and other services [reference], and it seems reasonable to suppose that such an approach would work for reporting and receiving advice on access problems.

Several libraries and consortia have determined that they can respond more effectively to support problems if the work can be distributed among a number of units or individuals - such as those identified as "resource coordinators" or liaisons. This strategy necessitates some coordination and orchestration - which some libraries have achieved by establishing and communicating "triage" paths for different situations. Doing so seems to be made easier if vendor contact information can be assembled and maintained centrally.

2.8. Ongoing Evaluation and Usage Information

Libraries have long had a strong interest in knowing how, how much, and by whom their collections are being used. Of course, a prevailing motivation has been to focus spending on those parts of the collection that appear to be used most heavily - though it has also been recognized that some specialized materials are likely never to receive much use. Measuring the use of print collections has been approached in various ways, with early efforts to develop and apply unobtrusive techniques giving way to reliance on online systems for tracking external circulation and systematic recording of in-library use.

For many years now it has been possible to track the usage of electronic resources, and many libraries now have considerable experience gathering, analyzing and presenting such data for internal and external consumption. In addition, libraries have come to expect vendors to provide usable and useful data on an ongoing basis. Although it may be possible to state such an expectation fairly simply, defining it has proven to be difficult. Just what should be measured, how the information should be presented, what data might mean, and how they might be used confidently in decision-making are all questions without obvious answers.

ARL has recently launched an intensive investigation of some of these questions. The first extensive report on this project (Shim et al. 2000) provides an interesting and useful taxonomy of three types of situations in which such usage information may be used for decision-making. The first category is termed "external resource contracts" - which includes the use of data for journal or database renewal or cancellation, or changes in numbers of concurrent users. The second is "reporting and communication," and includes budget justification (the most common response in this category), strategic planning, and comparisons to other institutions. Last is "service assessment and improvement," which includes the use of data for redesign of web pages, marketing and instruction efforts, staffing changes, and evaluation or assessment of resources available on a trial basis.

There are some obvious impediments to obtaining data useful for these purposes. First, most libraries currently depend on vendors for usage information, and many supply no data at all. As noted by Luther in her recent white paper on e-journal usage statistics (2000) "less than half of the publishers who offer journals in electronic form today are able to provide statistics on the usage of these journals." Second, data provided by different vendors may not be comparable. Again, quoting Luther, "librarians currently receive reports with different data elements that are not clearly defined and that cover different time periods, making it impossible to analyze them in a consistent way." The extent of this problem is indicated by the frequency with which libraries complained about it to the ARL team (Shim et al. 2000), and by the wide diversity of measures and other practices reported in the detailed inventory of "electronic vendor statistical reporting capabilities" reproduced in the team's report.

A number of responses to these dilemmas seem especially notable. First, it is important for libraries to institutionalize the ongoing evaluation of their electronic resources based on what is currently available from vendors. One part of such a strategy may be to treat the available data -- despite its inconsistencies and other problems - as good enough for "best guess" decision-making. Such an outlook has been adopted by Virginia Tech, which routinely assembles the available data into fairly simple but useful spreadsheets [reference]. Yale, Harvard, and several other libraries have also established clear and understandable Web pages for reporting usage information to interested staff [references]. Another interesting approach is to adopt a regular plan for evaluating each resource in advance of renewal and to distribute responsibility for the evaluations - as done by Harvard's resource stewards and coordinators [reference], and by the California Digital Library's resource liaisons [reference].

Second, it is important for libraries and vendors or publishers to reach a consensus on standard definitions and reporting practices. The International Coalition of Library Consortia took a very significant step in this direction when it published its "Guidelines for Statistical Measures" in November 1998 [reference]. This document has provided a reasonable standard that many vendors have already attempted to reach. The group is currently reviewing the guidelines and vendor responses, and plans to review the extent of vendor compliance - which may have further positive impact. The ARL E-metrics initiative that led to the study by Shim et al. may also hold potential for establishing additional consensus, as well as encouraging reform and standardization of vendor practice. And, of course, it is important for individual libraries to encourage vendors that do not provide what is needed to do so.

Finally, standardization need not be the only goal of measurement efforts. Those libraries and consortia that rely on their own equipment and staff for access to electronic resources - such as OhioLINK -- are well positioned to develop and initiate measurement and presentation methods that could become standards in the future. The usage measurement and presentation methods developed at the University of Pennsylvania demonstrate that it is possible to be innovative without an extensive local investment in infrastructure. Some libraries are also attempting to show how electronic resources affect student learning or facilitate research productivity [reference]. If successful, these efforts could become widely emulated.

2.9. Preservation and Archiving

The need to preserve electronic resources and continue to make them available has been recognized for a number of years, and there have recently been a number of serious and broad-based efforts. There are many possible explanations for this, but one of the more important is that many libraries now have experience with significant numbers of electronic journals and have started to consider whether they can truly afford the cost of both online and print subscriptions. The prospect of relying completely on electronic access to substantial numbers of journals or other resources raises a number of questions. For example, if libraries rely completely on electronic access and have to cancel their electronic subscriptions in the future, or the publisher goes out of business or discontinues access to older content, there is a real risk of losing access to back issues. If the publisher provides libraries with electronic files in those eventualities, how are libraries to utilize them? Lastly, there is increasing awareness that electronic versions of journals can and do differ significantly from printed versions. Electronic versions may include video and sound clips, special purpose data requiring specific software, and links to ephemeral web sites that may be difficult or impossible to archive.

A number of recently published articles have explored the meaning of electronic archiving, and who should take responsibility for it [references]. There seems to be a growing consensus that libraries are the appropriate agencies to serve this function - rather than publishers [references]. Nevertheless, possible costs could prove to be quite substantial - if, for example, data must be frequently refreshed and migrated, or if emulation software must be developed and maintained [references]. Most libraries will have difficulty devoting substantial local funds to these activities.

A few key and related initiatives take on special significance in this context. The first is the effort by CLIR and DLF to establish shared understandings within and across the publishing and library communities about the different elements that will be required to successfully establish working archival repositories. The second is the funding by the Mellon Foundation of planning grants at a number of larger libraries. The roles of these libraries is identified by CLIR as follows:

Yale, Harvard, and the University of Pennsylvania will work with individual publishers on archiving the range of their electronic journals. Cornell and the New York Public Library will work on archiving journals in specific disciplines. MIT's project involves archiving "dynamic" e-journals that change frequently, and Stanford's involves the development of specific archiving software tools. [reference]

The importance of these efforts to publishers and libraries is indicated by the joint announcement by Yale and Elsevier of their intent to work together on this problem. Both parties hope to establish a model archive in two years or so.

2.10. Toward Integrated Systems for Managing Electronic Resources

As noted earlier, a number of locally developed computer-based systems for acquiring, managing and supporting electronic resources were identified during the research for this project. Among these electronic resource management systems were MIT's VERA (Virtual Electronic Resource Access; reference) system, Penn State's ERLIC (Electronic Resources Licensing and Information Center; reference) and the License Tracker system developed at the University of Texas at Austin [reference]. These and other systems were found at larger institutions with significant levels of investment in electronic resources and substantial organizational complexity. In some cases, considerable amounts of staff time have clearly been devoted to designing, implementing, and maintaining them.

Although these electronic resource management systems vary in purpose and functionality, they have all been implemented to remedy perceived deficiencies in these institutions' established online cataloging, acquisitions, or other systems. For example, some are used to generate alphabetical lists of databases and electronic journals, or to keep track of important license terms. Some integrate the listing and license tracking functions, and others focus on tracking acquisition status or organizing technical information. As these systems' similarities and common functions began to emerge, the value of a more thorough and systematic inventory became apparent. That is, if many libraries are trying to solve much the same problem, it might be possible to devise common functional and data definitions and standards that could be used as the basis for future design and implementation work - by libraries working individually or collectively, or by vendors.

With this in mind, the author and another librarian at the University of Washington carried out a thorough review of known electronic resource management systems. Based on initial reviews of several systems, a coding form containing appropriate functions and data elements was developed and used to profile them. Staff at the institutions having relevant systems generally performed these reviews, but University of Washington staff analyzed a paper-based system in place at the University of Nevada - Reno on the basis of a published article about it (Loghry and Shannon 2000). In a couple of other cases an initial analysis of a system's functions and data elements was done by UW staff and then sent for review to staff at the library responsible for that system. Thirteen electronic resource management systems were analyzed. These included the four already mentioned (MIT, Penn State, Texas, and Nevada - Reno), and five other systems that are in place and operating (Michigan, Notre Dame, Simon Fraser, Virginia, and Yale). Three of the remaining systems (Cornell, Stanford, and UCLA) -- are in various phases of planning or development. A similar analysis of functions and data elements was also done for the University of Washington, although its current system consists of its Innovative Interfaces system, Digital Registry, and paper-based license file and inventory tool.

The results of the analysis are presented in spreadsheet format in Appendix B (Functions and Data Elements for Managing Electronic Resources). The systems summarized are presented in three categories. Because Nevada - Reno's is essentially paper-based, it is presented by itself. Since eight of the remaining systems are "in production," they are presented together. Lastly, the three systems in development are grouped along with the University of Washington. Between 140 and 150 separate functions or data elements are listed - although careful analysis will probably reveal some overlap and generate a somewhat smaller list. Seven fairly distinct functional areas were identified, including listing/descriptive; license-related; financial/purchasing; process/status; systems/technical; contact and support; and usage. For the first category, it seemed useful to distinguish between "listing or reporting" functions and the data elements themselves. That distinction seemed less clear or relevant for the other categories.

For each of the systems analyzed on the spreadsheet, an "X" generally appears in the relevant spreadsheet cell where a function or data element was identified. However, several of the systems interact with and support online acquisitions or cataloging systems, or systems for gateway page generation. Because of that, other codes - especially those for purchasing or financial functions and data elements -- appear. For example, those libraries using their Innovative Interfaces acquisitions systems to manage online subscriptions will show "III" where appropriate. Since the entries for Cornell are based on a planning document that identified functions and data elements as "required," "recommended," or "optional" for a future system, those values were entered as indicated. Since each of the profiled systems is complex, this highly schematic summary may tend to oversimplify or misrepresent them in some way.

Similarities and differences - including some unique and interesting features and functions - emerge from a review of the spreadsheet. First, they are based on a number of different database software packages or other platforms, although Filemaker and Microsoft Access are in use at 5 of the institutions. As seen in category 1.A. of the spreadsheet (listing and descriptive functions), several of the systems are used to generate "production" alphabetical and subject lists of resources for users and staff. Separate lists of databases and electronic journals are frequently generated, and several libraries generate "composite" full text lists that include both e-journals and coverage by "aggregator" full text database providers. A couple of libraries also generate lists by "package" or publisher. Many of the data elements used to describe print resources, such as title and publishers, are shown in category 1.B (listing and descriptive data elements). A few fields that are particularly appropriate or uniquely relevant for electronic resources are also shown, such as description, genre, and inclusion of full-text. Two institutions also facilitate their reporting of e-resource expenditures to ARL by recording appropriate expenditure categories from the ARL Supplemental Survey.

License term recording and display practice is somewhat more varied. Several institutions track and display common license terms for users and staff. For example, Yale's system presents informs users of whether a resource can be used for Interlibrary Loan, e-reserves, or course packs -- in a clear, standardized format. One of the more interesting display strategies was MIT's. In addition to a standardized notice about appropriate and prohibited use, MIT's resource lists display a red "L" icon appears when a license for a resources contains specific terms of which users and staff should be aware. When the icon is clicked on, the terms are displayed. Several institutions record and display links to electronic versions of their licenses, and Penn State has initiated a special project to digitize and make its licenses available.

As noted, libraries rely fairly extensively on established online acquisitions systems for tracking and reporting financial details. Whether the complexities involved in e-journal pricing and their many permutations can be successfully captured in this way is open to some question, however. Interestingly, several libraries try to save key e-mail correspondence with vendors because they record important transaction details.

Where traditionally designed online acquisition systems appear to be least successful in supporting the purchase of electronic resources is in how they allow for coverage of "process" or status information. The importance of doing so is indicated by the fact that most of the electronic resource management systems incorporate some sort of "order status" information. However, the range of relevant system characteristics suggests that devising a standardized approach to process tracking may be difficult. One reason for this is that successfully keeping track of a resource as it passes through the local acquisition process is likely to require the inclusion of steps that are highly specific to a given institution and its particular organization and workflow. Several of the systems track and provide reports of upcoming renewals. This is important for two reasons. First, an unpaid invoice may result in the sudden interruption of an institution's access to an electronic resource. Secondly, as mentioned in the section on evaluation and usage information, several libraries and consortia conduct cyclical reviews of electronic resources prior to renewing them. It would be difficult to plan for and coordinate such reviews without being able to see which resources will be up for renewal at what time. A couple of institutions have also included a "follow-up needed" reporting function in their local e-resource management systems, which should help prevent staff from losing track of problems.

Section 5 of the spreadsheet summarizes the wide variety of technical and access details that are routinely recorded in order to manage collections of e-resources. Six of the systems allow for tracking availability or "problem status," which is an issue when collections consist of resources from dozens and dozens of vendors. A related and attractive feature of the VERA system is that a special "resource broken" icon can be displayed on resource lists when a resource is unavailable. To support multiple vendor offerings, the recording of vendor billing and technical support information is also required, and most of the systems evaluated allow for that. In addition, most of them allow for the identification of an internal contact person for a specific resource - who might be responsible for resolving problems and reporting back on them to staff and users. A few of these systems also allow for recording where usage information can be obtained, or include a usage reporting function.

3. Conclusions and Future Considerations

One of the goals of the DLF Collection Practices Initiative was the identification of "best practices," but this goal began to seem overly ambitious as research for the present report progressed -- and more and more complexities and specialized local circumstances surfaced. It seems more appropriate in closing to identify a few practices that appear to be effective - or that just seem to be promising.

Before doing so, it is important to recall that the introduction to this report noted that commercially available resources must be affordable for libraries to be able to build and sustain significant collections of them. That comment served as the rationale for exploring two strategies for controlling costs -- consortial purchasing and scholarly communication initiatives. As noted, recent literature concerning both general strategies indicates that there is much to be said for both. Because of that, and since it is difficult to assess their current impact or to predict how either might help control prices for electronic resources over the long term, neither is being put forward here as an "effective practice." Affordability remains a necessary precondition for sustainability, however.

Leaving such strategies and related controversies aside, it was still difficult to choose objectively from among the practices discussed in the other sections of the report. Nevertheless, a selection was made and is presented as Table B, entitled "An Idealized Model of Effective Selection and Presentation Practices" Another researcher could easily have chosen differently, and that element of subjectivity should be kept in mind when reviewing it. Table B follows the organization of the text and the lengthier list of practices and links in Appendix A, so rather than provide a discussion of why each item was selected, a few brief comments will be given. Interested readers can look back at the appropriate text and Appendix sections for more information and discussion.

Table B: Idealized Model of Effective Selection and Presentation Practices
Topical Area Suggested Practices
2.2. Selection policies and strategic plans
  • Well-developed selection guidelines and policies
  • Strategic approach/articulated goals for e-resource development
2.3 Institutional Finance and Organization
  • Broad-based oversight/coordination committee structures
  • E-resource coordinators
  • Distributed resource "stewardship" responsibilities
2.4 Internal Procedures for Initial Evaluation and Purchase
  • Work flows systematized and clear; appropriate forms utilized to expedite handling
  • Order status of a given e-resource can be easily determined
  • Standardized information about library (FTE's, IP ranges, site definition, licensing policies) readily available to vendors.
  • Clear system of conducting trials that includes clear communication of availability and process to staff (and users, if appropriate)
2.5 Licensing Practices
  • Smooth handling of licenses with clear policies and responsibilities
  • General and specific licensing terms systematically made known to staff and users
2.6 Web Presentation Strategies
  • Aggregator database periodical holdings visible to users
  • A & I database citations linked to e-journal holdings
  • Personalization ("My Gateway")
2.7. User Support
  • General support information readily available to users
  • Clear problem escalation/triage paths for staff
2.8 Ongoing Evaluation and Usage Information
  • Planned/cyclic reviews prior to renewal
  • Systematic reporting of usage to staff
2.9. Preservation and Archiving
  • Supports joint efforts to establish preservation techniques and standards
  • assesses local capacities vs. risk factors for resources
2.10 Toward Integrated Systems for Managing Electronic Resources
  • E-Resource support systems in place or in planning

The discussion of selection policies and strategic plans pointed to the importance for libraries to try to articulate what they wish to achieve through their acquisition of commercial electronic resources, and to identify the values that should inform their decision-making. Creating strategic plans for e-resource collection development - especially those that incorporate concrete steps and benchmarks - seems to be another very worthwhile activity. For large libraries, the purchase and presentation of electronic resources can be a complex process involving large numbers of staff, and coordination of decision-making and other activities is a significant problem. Quite often coordination is achieved partly through one or more e-resource committees, for which there are many interesting organizational models. An important, emerging job role that appears to be common to many institutions is that of "E-resource coordinator." Lastly, as resources grow and diversify, some distribution of responsibility for resources seems inevitable, and a number of institutions have complemented "e-resource coordinator" positions with some form of "stewardship" role for specific resources or platforms.

The acquisition of electronic resources can also be operationally complex, and efforts to make the local process more coherent seem particularly important. Steps such as systematizing and documenting work flows, organizing and providing standardized information for vendors, and the adoption of appropriate forms all hold potential for expedited order handling. Managing the licensing of electronic resources also counts as one of the bigger challenges that libraries now face. Operationally, effective local licensing practice requires the adoption of clear policies and responsibilities, as well as the creation and communication of smooth processes for dealing with them.

Ideas about how best to present licensed resources to users continue to emerge and develop rapidly. Recent initiatives aimed at making the contents of aggregator databases more visible to staff and users, and at linking indexing databases to licensed e-journal collections enable libraries to derive more benefit from those resources. Establishing open standards for both seems important - since to do otherwise may mean supporting multiple vendor-specific systems. Efforts to better identify the needs and interests of particular user groups, and to enable users to personalize their own views of relevant resources based on their own interests and usage patterns, also stand out..

Extensive reliance on vendors for access to e-resources - and the attendant proliferation of interfaces and possible sources of problems - appears likely to necessitate a creative look at how the support function can best be organized. Establishing clear "problem triage" paths is an important step. Reliance on vendor performance has also been problematic for libraries that rely on them exclusively for usage information. Ongoing efforts to establish standards for such data and to urge vendor compliance are quite important, and individual libraries can and should support those efforts in their contacts with vendors. The systematic, ongoing presentation of such data to staff, and its review under organized, ongoing programs when resources are up for renewal also seem to qualify as good practice.

Though clear solutions do not yet seem to be on the horizon, it is extremely helpful that a number of organizations and libraries are taking a keen interest in the crucial issue of archiving and preserving electronic resources. And lastly, it is clear from the time and effort invested by many libraries creating local systems for managing electronic resources that existing library management systems and software lack important features and functionality. Although developing such local systems probably contributes to effective local practice, coordinated efforts to define needs and establish standards may prove to be of broad benefit.

References

Anderson, Rick. "Is the Digital Archive a New Beast Entirely?" Serials Review. Vol. 26, No. 3 (2000), pp. 50-52.

Barber, David. "Internet-accessible Full-text Electronic Journal & Periodical Collections for Libraries." Library Technology Reports, Vol. 36, no. 5 (September-October 2000). Chicago: ALA, 2000.

Bleiler, Richard and Terry Plum. Networked Information Resources. SPEC Kit 253 (December 1999) Association of Research Libraries.

Boyce, Peter B. "Who Will Keep the Archives? Wrong Question!" Serials Review. Vol. 26, No. 3 (2000), pp. 52-55.

Brown, John Seely and Paul Duguid. The Social Life of Information. Boston, Mass.: Harvard Business School Press, 2000.

Campbell, Jerry. "The Case for Creating a Scholars' Portal to the Web: A White Paper." (2000)

Chudnov, Daniel, Cynthia Crooker and Kimberly Parker. "jake: Overview and Status Report." Serials Review. Vol. 26, No. 4 (Oct. - Dec. 2000), pp. 12-17.

Cox, John. "Model Generic Licenses: Cooperation and Competition," Serials Review. Vol. 26, Issue 1 (April 2000), pp. 3-9. < http://www.sciencedirect.com/web-editions?_ob=MImg&_imagekey=B6W63-40JFYV1-2->

Crane, Gregory. "Commercial Digital Libraries and the Academic Community: How New Firms Might Develop New Relationships Between "Publisher" and Higher Education," D-Lib Magazine, Vol. 7 No. 1 (January 2001). http://www.dlib.org/dlib/january01/01crane.html.

Curtis, Donnelyn, Virginia M. Scheschy, and Adolfo R. Tarango. Developing and Managing Electronic Journal Collections: a How-To-Do-It Manual for Librarians. N.Y., Neal-Schuman: 2000.

Demas, Sam. "Collection Development for the Electronic Library: a Conceptual and Organizational Model," Library Hi Tech, Vol. 12 No. 3 (1994): pp. 71-80.

Demas, S. G., P. McDonald and G. Lawrence. " The Internet and Collection Development: Mainstreaming Selection of Internet Resources," Library Resources and Technical Services, Vol. 39 No. 3 (1995): pp. 275-90.

Douglas, Kimberly. "Digital Archiving in the Context of Cultural Change." Serials Review. Vol. 26, No. 3 (2000), pp. 55-59. < http://www.sciencedirect.com/web-editions?_ob=MImg&_imagekey=B6W63-41S4V97-9->

Duranceau, Ellen Finnie. "License Compliance," Serials Review, Vol. 26, No. 1 (2000): pp. 53-58. < http://www.sciencedirect.com/web-editions?_ob=MImg&_imagekey=B6W63-40JFYV1-6->

Duranceau, Ellen Finnie. "License Tracking," Serials Review, Vol. 26, No. 3 (2000): pp. 69-73. < http://www.sciencedirect.com/web-editions?_ob=MImg&_imagekey=B6W63-41S4V97-D->

Frazier, Kenneth. "The Librarians' Dilemma: Contemplating the Costs of the "Big Deal," D-Lib Magazine Vol. 7, Number 3 (March 2001)

Graham, Rebecca A. "Evolution of Archiving in the Digital Age." Serials Review. Vol. 26, No. 3 (2000), pp. 59-62. < http://www.sciencedirect.com/web-editions?_ob=MImg&_imagekey=B6W63-41S4V97-9->

Holleman, Curt. "Electronic Resources: Are Basic Criteria for the Selection of Materials Changing?" Library Trends, Vol. 48, No. 4 (Spring 2000): pp. 694-710.

Hughes, Carol Ann. "Information Services for Higher Education: A New Competitive Space." D-Lib Magazine, Vol. 6, no. 12 (December 2000)

Hunter, Karen. "Digital Archiving." Serials Review. Vol. 26, No. 3 (2000), pp. 62-64. < http://www.sciencedirect.com/web-editions?_ob=MImg&_imagekey=B6W63-41S4V97-B->

Kovacs, Diane. Building Electronic Library Collections: The Essential Guide to Selection Criteria and Core Subject Collections. N.Y., Neal-Schuman: 2000.

Kovacs, Diane K. and Angela Elkhordy. "Collection Development in Cyberspace: Building an Electronic Library Collection," Library Hi Tech, Vol. 18, No. 4 (2000): pp. 335-359.

Landesman, Margaret and Johann Van Reenen. "Creating Congruence." JEP: the Journal of Electronic Publishing, Vol. 6, Issue 2 (December 2000)

Leggate, Peter and Mike Hannant. "The Archiving of Online Journals," Learned Publishing, Vol. 13, No. 4 (October 2000): pp. 246-250.

Loghry, Patricia A. and Amy W. Shannon. "Managing Selection and Implementation of Electronic Products: One Tiny Step in Organization, One Giant Step for the University of Nevada, Reno." Serials Review, Vol. 26, No. 3 (2000): pp. 32-44. < http://www.sciencedirect.com/web-editions?_ob=MImg&_imagekey=B6W63-41S4V97-4-

Marcum, Deanna. "Digital Archiving," College & Research Library News. Vol. 61, No. 9 (October 2000): pp. 794-797, 807.

Metz, Paul. "Principles of Selection for Electronic Resources." Library Trends, Vol. 48, No. 4 (Spring 2000): pp. 711-728.

Michalak, Sara. "The Evolution of SPARC." Serials Review, Vol. 26, Issue 1 (April 2000): pp. 10-21.

Montgomery, Carol Hansen and JoAnne L. Sparks. "The Transition to an Electronic Journal Collection: Managing the Organizational Changes." Serials Review, Vol. 26, No. 3 (October 2000): pp. 4-18. < http://www.sciencedirect.com/web-editions?_ob=MImg&_imagekey=B6W63-41S4V97-2->

Morris, Sally. "Archiving Electronic Publications: What are the Problems and Who Should Solve Them?" Serials Review. Vol. 26, No. 3 (2000), pp. 64-68. < http://www.sciencedirect.com/web-editions?_ob=MImg&_imagekey=B6W63-41S4V97-C->

Okerson, Ann. "Are We There Yet? Online E-Resources Ten Years After." Library Trends, Vol. 48, No. 4 (Spring 2000): pp. 671-693.

Rothenberg, Jeff. "Avoiding Technological Quicksand: Finding a Viable Technical Foundation for Digital Preservation: A Report to the Council on Library and Information Resources." http://www.clir.org/pubs/abstract/pub77.html

Sanders, Thomas R., Helen Goldman and Jack Fitzpatrick. "Title-Level Analytics for Journal Aggregators." Serials Review. Vol. 26, Issue 4 (Oct. - Dec. 2000), pp. 18-29.

Sanville, Thomas J. "A Method Out of the Madness: OhioLINK's Collaborative Response to the Serials Crisis. Three Years Later - Progress Report." Paper presented at the North American Serials Interest Group Conference, University of California, San Diego, June 22-25 2000.

Shim, Wonsik, Charles R. McClure and John Carlo Bertot, "Data Gathering Practices in the Networked Environment." ARL: A Bimonthly Report on Research Library Issues and Actions from ARL, CNI, and SPARC. December 2000: pp. 6-7.

Soete, George J. and Trisha Davis. Managing the Licensing of Electronic Products. SPEC Kit 248 (August 1999) Association of Research Libraries.

Troll, Denise. "How and Why Are Libraries Changing?" http://www.diglib.org/use/whitepaper

Van de Sompel, Herbert and Patrick Hochstenbach, "Reference Linking in a Hybrid Library Environment, Part 1: Frameworks for Linking," D-Lib Magazine Vol. 5, Issue 4 (April 1999) http://dlib.ejournal.ascc.net/dlib/april99/van_de_sompel/04van_de_sompel-pt1.html.

Van de Sompel, Herbert and Patrick Hochstenbach, "Reference Linking in a Hybrid Library Environment, Part 2: SFX, a Generic Linking Solution," D-Lib Magazine Vol. 5, Issue 4 (April 1999) http://dlib.ejournal.ascc.net/dlib/april99/van_de_sompel/04van_de_sompel-pt2.html.

Van de Sompel, Herbert and Patrick Hochstenbach, "Reference Linking in a Hybrid Library Environment, Part 3: Generalizing the SFX solution in the "SFX@Ghent & SFX@LANL" experiment," D-Lib Magazine Vol. 5, Number 10 (October 1999) http://dlib.ejournal.ascc.net/dlib/october99/van_de_sompel/10van_de_sompel.html. TimFinal2; 04/29/01; 8:08 PM Page 6