KEYWORDS: Digital libraries, online information resources, usability, online databases, interorganizational networks, institutionalism, postmodernism
OI resources, as we will refer to them in this review, are curated collections of indexed electronic databases with supporting distribution services. Online service vendors have traditionally provided fee-for-service modem access to mainframes containing these databases of strategic business, scientific, legal and financial information. Initially, the services were entirely text-based. Now, most OI service providers supplement their mainframe offerings with GUI-enhanced CD-ROM products. They have also begun to provide additional access points via consumer utilities like CompuServe and America Online, and the World Wide Web.
Some of the most successful online services trace their origins to early 1970's outgrowths of government defense contractors, like Lockheed's DIALOG, while others were spawned by wire services and business news publishers in the US and Europe, like Dow Jones and Reuters. Some not-so-successful attempts to commercialize online information were initiated in the early 1980's by joint ventures between cable TV, telecommunications and publishing companies which offered videotext services directly to consumer households. At the time, videotext was considered to be an important vehicle for ushering in the Information Age, but only France's Minitel lived up to the aspirations of its initiators. Jerome Aumente points out that the causes of failure for most videotext services have not been well-researched [3], but he suggests that technical infeasibilities were to blame and that the services were 'technologically ahead of their time.' He theorizes that the videotext failures opened a window of opportunity for the slow, steady rise of the plain text online commercial database services, such as DIALOG and NEXIS, and for ERIC--one of the precursors of modern digital libraries.
Aumente blames the technology, and few would quarrel that technological feasibility is an essential ingredient in the fabrication of OI resources. But the Minitel success has been largely attributed to an important combination of political and technical factors [5]. These stories, and the social, economic and historical analyses of Bell, Cortada, Porat and Rubin, Yates and others, provide a grounding for the examination of OI resource usability from a variety of perspectives and at several different levels [4, 7, 34, 39]. In the balance of this paper we will examine the research literature on OI usability. We will follow the discussion from technological usability, to content usability, to organizational usability and to interorganizational usability. We will be looking to see where empirical evidence supports existing theories which guide the construction of digital libraries, and where it suggests a need for modified conceptualizations of usability or the need for further research.
We have classified the aspects of OI usability that we find in the literature into four major categories: HCI Usability, Content Usability, Organizational Usability and Interorganizational Usability. HCI Usability involves issues of individual access and individual resource usage on a personal level or within working groups. It highlights human-computer interaction with the OI resource interface, but also takes into consideration how individuals might need to coordinate resource usage in a work group setting. Content Usability involves characteristics of the information itself. It addresses issues of information relevance, information awareness, information retrieval methods, and expertise. Organizational Usability involves the situated use of OI and OI resources. A library or organization may provide training and support staff to help people use the available resources, or it may not. The information itself may be used collaboratively among working groups or symbolically as post-decision support. Interorganizational Usability involves use of OI and OI resources within interorganizational networks. This may entail uses of OI to achieve competitive advantage, or to coordinate professionals of the virtual organization. Table 1 outlines the components of these usability categories, and it identifies the research literature which discusses them. The table also provides a classification of the discussions by the perspectives which inform them. We will examine these perspectives in depth as our discussion proceeds.
Rational Human Institutional Postmodern
Relations
HCI Usability
individual access Eisenberg 83 Culnan 85 Aumente 87 Amiran et
group work/benefit Hesse et al. 93 Grudin 90 Orlikowski 93 al. 92
Content Usability
awareness and Larson 89 Kaye Nicholas Oddy et al. 92 Luke 91
retrieval strategic 90 and Erbach Wilson 95
information 89
Pfaffenberge
r 90
Organizational
Usability
training and support Koenig 90 King Ackerman 94 Kling & Poster 90
control or & Grover 91 Mintzberg 75 Elliott 94
cooperative use O'Day &
symbolic use Jeffries 93
Feldman &
March 81
Interorganizational
Usability
competitive use Marchand & UNCTC 82 Libmann 90 Clegg 90
interorganizational Horton 86
networks
Martha E. Williams has been accumulating and analyzing the statistics of the online database industry for a number of years. Her classifications of "users" are at the institutional level: academics, information brokers, government, industry, legal institutions, medical institutions, not-for-profits and libraries [38]. Levy and Marshall also confirm that digital libraries are "users" of OI resources [22].
The people searching these databases are also called "users", and they are often segmented on the basis of their information usage into two groups: end-users and intermediaries. Intermediaries are usually associated with libraries or corporate information centers. They search the OI resources at their disposal on behalf of others who will use the information retrieved in decision-making or other tasks. Aumente quotes 1987 DIALOG usage statistics as 80% by search intermediaries [3]. End-users are commonly considered to be individuals who use the information they retrieve directly in their own work tasks. There has been a significant amount of research attention paid to the end-user. [Ref. 10, 28, 9, 33] However, some researchers have found it difficult to segregate the intermediary from the end-user in situated use. They have concluded that in their studies all individual OI resource users were essentially intermediaries because "the work practices of their settings led them to share their search results with others, either by request or on their own initiative" [30].
Who the user is and what the user does has been the quintessentially interesting research question of the OI resource provider and the digital library architect. However, as was demonstrated by the videotext fiasco, the providers and architects are not always willing or able to perform the research and evaluate the results before constructing or marketing the resources.
When empirical usability research does not exist or cannot be done because the technologies under construction have no current equivalent, OI resource providers in industry and academia rely on theoretical perspectives that are widely espoused by respected members of their communities to guide them. One perspective on OI usability is widely and tenaciously held. It can be characterized as optimistically rational, and it is espoused as a complex array of social, political and organizational strategies which emphasize the importance of information, information resources and information technology for economic survival and social progress. We call these strategies "informational imperatives." They guide organizations and individuals to adopt more information-intensive practices, and they set expectations for particular socio-technical results.
Why these informational imperatives exist and persist is the topic of another paper, but for the purposes of this review, it is important to note that much of the "problem" with implementing OI resources and other digital library resources may actually be a discrepancy between how researchers, providers and users themselves expect that the resources will be used, and how they actually are used.
Rationalists expect the aggregate use of OI resources by working groups to yield productivity benefits and stimulate additional use. Hesse et al. set out to measure this productivity among oceanographers connected to SCIENCEnet, suggesting that "[i]ncreasing ease of access to these resources should lead to a concomitant increase in scientific productivity" [15:90]. They made some positive correlation between access and productivity, but because SCIENCEnet only provides access to oceanographic data and does not contain the data itself, they did not address issues of content usability.
Content Usability. We characterized Content Usability as follows: Content Usability involves characteristics of the information itself. It addresses issues of information relevance, information awareness, information retrieval methods, and expertise. Rationalists believe that when individuals learn about OI resources, they will recognize the potential value of OI contents and use them accordingly. If OI resources are not being used, then it must be because people just don't know enough about them. They believe, as Jon Kaye does, that "to take full advantage of on-line database services, the threshold of awareness of the value of information must be raised. Users are still not understanding the value of information as a strategic tool" [16].
However, rationalists working with content usability understand that establishing awareness is only one facet of the problem. Once access and awareness have been established, there are still issues of information retrieval and information overload. Some researchers, like Ray R. Larson, try to resolve both of these problems simultaneously by developing online systems which use techniques derived from information retrieval research. Recognizing that users sometimes just "give up" when their queries don't return relevant results immediately, Larson attempts to cluster the data into meaningful classifications that, when coupled with statistical ranking mechanisms, can "reduce the cognitive load on the user in selecting relevant documents" [21:129].
Organizational Usability. As we stated earlier, Organizational Usability involves the situated use of OI and OI resources. An organization may provide training and support staff to help people use the available resources, or it may not. The information itself may be used collaboratively among working groups or symbolically as post-decision support. One of the reasons why rationalist managers of an organization may choose not to provide technical support or training on OI resources is because it costs too much. Researchers like Michael Koenig seek to convince them otherwise by extrapolating from economic theory to affirm that organizations systematically underinvest in OI support services and libraries [20]. He then goes on to assert that the existence of these organizational infrastructures enhances the free-flow of information throughout the organization, and leads directly to productivity gains and improved decision-making, but his studies do not identify the mechanism by which they do so.
Along with providing OI resource usage support, rational organizations in the Information Age are expected to develop formal organizational strategies to guide the optimized use of both internal and external information available through these resources. But a survey of IS managers by King and Grover found that "[n]early 55% reported either 'irregular' or 'no formal process', suggesting that analyses dealing with information technology and information as a strategic resource have not yet become a routine and regular part of the planning process" [17:300]. King and Grover suggest that, in order to successfully 'mine' online databases for potentially powerful information, organizations should adopt a model of competitive advantage like those developed by rationalists concerned with interorganizational OI usability.
Interorganizational Usability. Marchand and Horton adopt the view that efficiently functioning markets are the highest good [26]. And, for rationalists, healthy competition is the hallmark of efficiently functioning markets. We said that Interorganizational Usability involves use of OI and OI resources within interorganizational networks. This may entail uses of OI to achieve competitive advantage, or to coordinate professionals of the virtual organization. According to Marchand and Horton, managers must learn to use information strategically, in order to correctly assess environmental changes in their own industries and to spot potential opportunities in heretofore "other" industries. To do this, managers will have to realign the organizational structure with TQM-like methodologies.
The profitability and efficiency concerns of corporate information consumers may not be motivating concepts for digital library planners, but an interorganizational usability awareness may allow them to consider that "outsourcing" organizations may take advantage of publicly available digital libraries. Rational perspectives which encourage efficiency and the attainment of competitive advantage rarely consider the consequences of shifting the productivity burden to public institutions.
HCI Usability. While rationalists assume that OI resource access naturally leads to use, human relations (HR) theorists seek to broaden the definition of accessibility. Mary Culnan has personalized the access question by breaking it down into three behaviorally oriented components: physical access (i.e. finding a terminal), interface access (i.e. using a query command language) and informational access (i.e. displaying data.) We have classified the last two components as aspects of Content Usability, largely on Culnan's findings that "physical access to the terminal used to access the system and access to the actual information system are independent dimensions" [8:303]. Furthermore, on the physical access dimension, she suggests that physical accessibility does not necessarily lead to use. Contrary to rationalist expectations, she finds that experience with a resource improves perceptions of accessibility and potential usefulness of the information attainable through that resource.
HR theorists researching at the group level are also very different from rationalists in one important respect--their expectations incorporate a notion of equity. Unlike Flores et al. [12], who suggest that efficient control is the answer to group work coordination problems, Grudin suggests that a balance in terms of who does the work and who benefits from that work would go further to alleviate coordination issues [14]. Although Grudin's work focuses on internal OI resources, like scheduling systems; his concepts can be seen to apply to external OI resources and digital libraries as well--especially when research functions are organizationally segmented.
Content Usability. HR theorists do not approach OI resource content usability from a perspective which assumes that mere awareness of the potential usefulness of a resource is enough to ensure use. In fact, Nicholas and Erbach conclude their examination of end-user searching in four commercially available OI resources with an important observation [29]. They find that the services are dissimilar, the pricing structures are divergent, downloading speeds are highly variable independent of baud rates, and in-depth knowledge about even a small set of OI services is difficult for any one person to obtain and maintain. Their findings are based on end-user observations and are motivated by a concern for lowering the frustration level of the end-user by identifying user behaviors which are not well-accommodated by the OI resources in the study. Nicholas and Erbach make numerous suggestions for improving these existing resources, which if followed, would lead to a standard approach to OI resource design. However, they do not explicitly make this recommendation.
Pfaffenberger agrees that there are a great many obstacles to end-user searching in OI resources [33]. He points out that equal opportunity of information access was rhetoricized as the inevitable by-product of technologically enhanced information networks. But he identifies inequity of expertise as the most serious threat to equity of access. Pfaffenberger suggests that online databases will only attract end-users who are highly educated and are able to judge the significance of the material they encounter online. Much of his discussion takes on a critical theory perspective, especially when he refers to the decline in public education quality and the proletarianization of the white collar work force, but we have included it here because his suggestion to restructure the online information formats so they are more accessible to end-users is a human-factors solution to the problem of expertise.
Organizational Usability. HR guided researchers are nearly unanimous in their advocacy of training and support infrastructures for OI resource usage. They may suggest that an organization provide basic skills training and technical support for its OI workforce, or they may suggest a kind of self-help system which blends humans and technology into a single package. Mark S. Ackerman describes how one such system has been developed to combine information sharing with expertise sharing [1]. His system is not merely a factual repository, it is a device for growing organizational memory, and so it encourages managers to reorganize work--especially how expertise is captured and how it is made available to all levels of the workforce.
Mintzberg takes a different tack [27]. Rather than suggesting a technological solution to information access, he concentrates his attention on intermediaries and collaboration. While trying to answer the question: "Why do managers not use information as they apparently should?", he focuses on identifying how managers work. He suggests that organizational impediments to OI resource usage arise from dysfunctional goal setting and the nature of managerial work and advancement. Mintzberg also encourages the reorganization of work to alleviate the problem, but he states that "[m]anagers need broad-based formal information systems, in large part independent of the computer...There is reason to believe that the best filters and pattern recognizers are intelligent human beings--bright specialists (and alter egos) who can stand between the manager and mass of data out there" [27:17-20].
Interorganizational Usability. The Human Relations perspective usually focuses on issues of job satisfaction and worker productivity, so there is not much HR literature which addresses OI resource usability at an interorganizational level. One notable exception is a report issued in 1982 by the United Nations Centre on Transnational Corporations [36]. The UN was worried that the "information gap" between developed and developing countries had been widening and lessening the potential for real competition by developing countries in world markets. They commissioned this study which suggested two organizational solutions to level the playing field: build human resource infrastructures in the developing countries, like the Mexican information center, or establish search centers in the information-rich countries. Whereas the rationalist is interested in allowing markets to function and information to flow, researchers with an HR view are more interested in removing barriers to entry and leveling the playing field.
The digital library community has taken many of the guiding principles of the public service library community to heart, and it strongly embraces the concepts of HR theorists. But concerns for equity of access, fairness in the distribution of work and advocacy for adequate training in distributed, multi-organizational environments should be guided by an awareness of these non-academic contexts of use in addition to traditional academic contexts.
HCI Usability. Institutional discussions of OI resource usage do not assign an active, rational decision-making role to the individual. They are more likely to present OI resource user choices as a complex set of constraints set up by the various organizations and institutions which provide and administer OI services. As we discussed earlier, Aumente presents an historical overview of the videotext and online database industry, with an accent on HCI usability aspects [3]. He finds fault with the usability of videotext formats and the communication exchange capabilities of the medium, but notes that similarly constrained services, like online databases, have been successful. He looks to see where the technological constraints have been overcome within the videotext arena, for example France's Minitel, and identifies the external political factors as key enablers. Aumente's set of recommendations for the establishment of usable and viable OI resources involves technological enhancements which are tightly coupled with political support for developing a national OI infrastructure.
Wanda Orlikowski has examined the introduction of groupware databases which served as digital libraries throughout a large, geographically dispersed organization [32]. Her study found that the groupware had been developed expressly for collaborative use. For those groups within her organization who shared information in a cooperative mode, the system was usable. But for the consultants, who worked in a competitive, non-information sharing mode, the system was largely unusable. They did not have any established norms for sharing information through digital libraries. Furthermore, they had real economic incentives not to share, but to retain and obtain as much expertise for themselves as possible. Orlikowski does not assume that everyone will want to use the resource, and she does not assume that equal access brings with it any incentives to use. She locates the implementation problem of her study outside the technological infrastructure, within the competitively-focused organizational incentive structures for the consultants, and suggests that modifications will have to be made here before the OI resource usage practices of the consultants can be expected to change.
Content Usability. Institutionalist information retrieval approaches try to acknowledge the taken-for-granted ways of doing things that can be discerned in particular organizations or environments. Oddy et al. have tried to incorporate situational information into OI retrieval systems. They attempt to generate generic research scripts and to match search processes with markers in the script so that the system will "know" where the searcher is in his or her research project. Oddy et al. believe that "[p]eople engaged in empirical work can be viewed as members of a community: they complete similar tasks, espouse shared social and technical norms, and operate within a unique reward structure" [31]. So knowing where a researcher is in a research project can help a search intermediary determine what is the most relevant information. Oddy et al. were able to identify some important patterns along these lines, but their study results suggest that "both users and intermediaries are constrained by their expectations about current systems and by their views of their own roles in the search process" [31].
Patrick Wilson also finds that users are significantly constrained in their use of OI resources [37]. In fact, his studies of R&D information gathering practices show that in some organizations the nonuse of strategic information may be dictated by policy. "A research group of any size--discipline, subdiscipline, field, specialty--may operate with a tacit explicit convention of ignoring categories of relevant information" [37:47]. Wilson lists a number of reasons for this including information overload and time constraints. But he also identifies disciplinary territoriality, or the cognitive division of labor that has come to define the boundaries of the scholarly research disciplines, as formulating a policy of information nonuse.
Organizational Usability. Kling and Elliott have investigated how the organizational environment can influence the use of OI resources and have made some suggestions for organizational support of the use of OI resources in digital libraries [19]. They focus attention on how these systems can be effectively integrated into the work practices of specific organizations. They suggest that the systems could be more organizationally usable if they were designed to include the infrastructure necessary to support and accommodate people as they learn to maintain and use these resources. However, Kling and Elliott recognize that their advice may be ignored by OI systems designers and digital library architects who make assumptions about how their systems will be used which are based on the taken-for-granted concepts of their own cultural and organizational ways of doing things.
Organizational use of OI resources is often associated with a library or an information center. Research services may be provided to an entire organization by a small group of information specialists who make use of OI resources, and print information resources, on behalf of others. As noted earlier, this confounds the issue of who is actually the "user"--someone who performs the OI search, or someone who uses the information retrieved by the searcher. In cooperative work settings, identifying the end-user of the information as someone separate from the information intermediary becomes even more difficult. O'Day and Jeffries have concluded that, in the organizations they studied, virtually everyone functioned as an intermediary in some fashion; either by updating team members, consulting, broadcasting or putting the information into a shared archive [30]. Although rationalists might suggest that the intermediary represents an inefficiency in the system of getting relevant information directly and quickly to the end-user, O'Day and Jeffries show that modes of work which involve information sharing, as opposed to intensive, individualized use, are prevalent.
While informational imperatives may encourage intensive information gathering practices, not all the information gathered is actually used in decision making. In fact, Feldman and March suggest that information gathering is largely a symbolic act of the decision making process [11]. They cite studies which show that little systematic relation can be established between the time of receiving the results of an information search and the time of making a decision. However, these studies do find that when the intrinsic quality of a decision is difficult to assess, the signals generated by information gathering during the decision making process may impart a legitimacy that affects the perceived quality of that decision. So, when managers cannot show that their decision is "correct", they can at least build confidence by showing that they have followed accepted practice in formulating it. With respect to OI resources, the conclusions of Feldman and March would suggest that organizational decision makers might be more likely to use them in a "scanning" mode when decision criteria are ambiguous, performance measures are vague or when decision quality requires a long period to become established.
Interorganizational Usability. As institutionalists have shown, there are a number of environmental constraints on OI resource use at individual, group and organizational levels. When a resource is constructed expressly for use at the interorganizational level, the reasons for its underutilization may be attributed to these constraints and to additional environmental influences. Francois Libmann's study of European online technology transfer databases shows that reasons for OI resource implementation failure can be linked to cultural concepts [23]. He finds that "[n]o one has sought to create synergies between databases and other elements in the world of technology transfer (interpersonal networks for example)...databases are on the whole little or inefficiently used to optimise innovatory process. This is a cultural problem and should be dealt with as such, and is one which is not exclusive to technology transfer databases" [23:202]. Libmann suggests that national policy must be modified to address this need.
US policy makers have encouraged the construction of digital libraries, and have stressed the importance of their educational support functions. But organizations on the periphery of academia, such as Libmann's who are promoting university/industry technology transfer efforts, are also interested in constructing or using digital library services. Digital libraries may serve these organizations, or may be embedded within them, depending on how central or supplemental the research function is to their core organizational processes.
HCI Usability. Postmodern views of OI usability do not fall neatly into the admittedly rationalist framework we have constructed for this review, as does much of the other literature we have discussed. While Amiran, Unsworth and Chaski discuss issues of individual accessibility in their article on electronic publishing, what they actually concentrate on is the rhetoric about the accessibility of electronic publishing [2]. They dispute unfounded claims that reading 'the text' is somehow fundamentally different when it is conveyed in electronic form: "There is no psycholinguistic evidence that the process of reading suddenly changes when it is conducted in a different medium" [2:51]. But they do find significant differences in the malleability of electronic texts, and suggest that a shift toward electronic scholarly publications may encourage a reinterpretation of the concepts of authorship, ownership and scholarship at individual and institutional levels. These are just the kinds of debates that we now see in the digital library literature.
Unlike rationalists, Amiran et al. are not preoccupied by providing quick, efficient access to OI resources, and they are less expectant of coordinated consensus as working groups share these resources. Their expectations for OI resource usability are very like institutionalist views in that they readily identify the legitimating functions of existing ways of using OI resources, but, as postmodernists, they are more likely to attend to signals which indicate that role boundaries, such as those of librarians and other intermediaries, are blurring as a result of OI resource usage.
Content Usability. Usability of 'the text', according to Timothy W. Luke, lies in its ability to serve as a site of shared cultural consciousness [24]. Luke examines Baudrillard's concept of 'hyperreality' and presents it, not as a future capability, but as an extension of existing social and political simulations which use current implementations of electronic media. He suggests that simulation and the shared cultural consciousness it can achieve are important contributions to the social infrastructure. Luke and postmodernists see fiction (e.g. TV and movie dramas like "Earthquake") and simulation (e.g. earthquake-preparedness drills) as preceding and defining reality (i.e. the actual earthquake) for those residents involved--that's hyperreality.
Although this perspective takes environmental and institutional factors into account, the postmodern view of content usability differs greatly from the institutional view, which is focused on how OI resources are manipulated by users. Postmodernists examine how users' perceptions are modified by the contents of OI resources.
Organizational Usability. With respect to organizational use, the postmodern emphasis is less concerned with training organizational members to use OI resources effectively, and is more focused on understanding how OI resources, as critical components of the technical infrastructure, sustain organizational bureaucracy. Mark Poster presents an insightful overview of the impacts of Bell's forecasts for the Information Society [35]. He identifies information as the sustaining materiel of bureaucracy, and he acknowledges Foucault's interpretation of online databases as the major structural component of domination in postindustrial society. But he suggests that databases also have a democratic side. In personal communication, he has associated democratic empowerment with database exchanges on the Internet to exemplify the dual potential of OI resources. In either case, the postmodern view of organizational OI resource usability can be seen to resemble the institutional view in some important ways. The end-user of the information is often difficult to identify. She may be the intermediary, the bureaucrat, or some collective identity that is amorphously referred to as 'the organization.'
Interorganizational Usability. In postmodern discussions, the terms 'organizational' and 'interorganizational' may be used interchangeably. Clegg's analysis of modern organizations defines one of the more recent examples, the Benneton firm, "less as an organization per se and rather more as an organized network of market relations premised on complex forms of contracting made possible by advances in microelectronics technology" [6:121]. Clegg does not refer exclusively to OI resources, but to IT generally speaking, when he suggests that the classic Weberian bureaucracy may be replaced by technologically enabled organizational forms such as Benneton's.
Unlike Marchand and Horton's rational analysis of interorganizational use of OI resources, Clegg's analysis does not stop at the managerial methodology level. And unlike institutional analyses which call for governmental intervention to enact cultural change so that existing organizational forms can make more effective use of OI resources and traditional library bureaucracies, Clegg identifies environmental and cultural changes which are redefining the boundaries of the organizations themselves.
The postmodern perspective encourages us to recognize that the boundaries which separate digital libraries, academic enclaves and commercial organizations are permeable and changeable. The roles of librarians and information intermediaries may be equally transmutable.
Our review of the research literature on OI resource usability, and its applicability to digital library contexts, is not exhaustive, but it is representative. We find that, when contrasted with the rational and human relations perspectives, the studies informed by institutional perspectives add a great deal of explanational value to discussions of HCI, content and organizational usability. Institutional theories provide conceptual tools which allow the use of OI resources to be analyzed as something other than a simple aggregation of individual activities molded by local contingencies. The institutionalist studies seem to offer the most robust explanations of observed OI resource use. These studies show how individuals and groups are constrained from using OI resources by existing incentive structures and taken-for-granted ways of doing things. But they also show how organizations are able to take advantage of cooperative work modes to share the results of OI searches. This body of research suggests a need for modified conceptualizations of OI usability within the driving imperatives of the Information Age. But it also points to the need for further research. Institutional perspectives on interorganizational digital library contexts of OI usability are not well-developed, and empirical studies, like Libmann's, are not often found in the literature. Postmodern studies at the interorganizational level, such as the organizational networks analyzed by Clegg, add important insights, and could be used to augment institutionalist explanation. This area appears to be the least well-explained, but possibly the most interesting one for extending empirical research. We suggest, as a result of our review, that empirical studies into interorganizational use modes of OI resources may shed light on key aspects of digital library use; and we expect that these studies would provide grounding for extending or supporting existing institutional perspectives.