destefano.p65 58 College & Research Libraries January 2001 Selection for Digital Conversion in Academic Libraries Paula De Stefano Electronic technology has begun to change the way scholars conduct their research. Before this new approach to scholarly inquiry becomes a viable and productive method in institutions of higher learning, the exist­ ing resources that a scholar normally would use in the library must be converted to a digital format in order to be accessible electronically. How do academic libraries set about creating a body of knowledge and begin to convert traditional print collections to a digital format in order to sat­ isfy what today’s researchers want? This article examines previous meth­ ods of selection and collection building, and applies those supporting principles to today’s collection-building efforts for digital collections. election decisions in academic libraries have never been clear- cut or straightforward. One has only to examine the litera­ ture to confirm that. However, a rudimen­ tary principle that library professionals historically have agreed on is this: Like other processes in the library, selection should be aligned closely with the mis­ sion and goals of the parent institution. This simple, but important, tenet of aca­ demic librarianship is supremely mean­ ingful in light of the resources that digi­ tal conversion activities consume, such as staff and funding. More strongly stated, it is incumbent upon the academic library community to develop a carefully rea­ soned approach to the selection of library materials for digital conversion that is fis­ cally responsible to both itself and its par­ ent institution. To select and select well is critical to the success of the digital library. As Clifford Lynch has pointed out: Libraries face both opportunity and potentially unmanageable budget­ ary demands from all quarters. The questions now facing libraries arise less from the availability of technol­ ogy than out of the development of strategies for collection develop­ ment and management and sup­ porting resour ce allocation choices.1 Despite the urgency to develop a co­ herent and sustainable approach to the selection process, the academic library community has yet to produce one. Per­ haps the biggest reason for this hesita­ tion has been the newness of digital tech­ nology itself. Much experimentation has taken place in the formative years of this technology in libraries, and more re­ search and development is needed to explore the capabilities of digital technol­ ogy, specifically, for what it has to offer Paula De Stefano is the Barbara Goldsmith Curator for Preservation in the New York University Librar­ ies; e-mail: paula.destefano@nyu.edu. 58 mailto:paula.destefano@nyu.edu Selection for Digital Conversion in Academic Libraries 59 the research community. Those efforts should be deliberate, well planned, and designed for the greater good of this community on a global scale. At the same time, however, it now is clear that a criti­ cal mass of a newly reformatted body of knowledge must be created to assist in this new electronic approach to research. This article examines the process in­ volved in developing an implementable rationale for the selection of library ma­ terials for digital conversion by review­ ing analogous conditions in collection- building history. In the hope that the past may inform the future (if not in practice, at least in principle), what follows is a reexamination of past collection build­ ing, including the advent of the printing press and the ensuing drive to develop print collections; an overview of the se­ lection process as it evolved in modern research libraries; and the selection cri­ teria developed to cope with brittle books and preservation microfilming. Then, turning to contemporary efforts, a small sampling of project managers of recent digital projects provides information about some of the contemporary selec­ tion methods used in today’s projects. Developing Print Collections The act of selecting and acquiring books to build and develop printed book col­ lections has evolved over the centuries since the time of the manuscript book. During the secular period of the manu­ script book and the early rise of the uni­ versity in the twelfth century in Europe, the demand for books began to move be­ yond the monasteries. The drive to pro­ duce manuscript books was fueled by the need for books in education and research. Universities employed professional craftsmen to copy texts by hand “expedi­ tiously and cheaply” for their courses.2 Following Johannes Gutenberg’s inven­ tion of the printing press and William Caxton’s perfection of movable type and the inking process, both in the mid-fif­ teenth century, the race was on to dupli­ cate and convert scholarly works from manuscript books to printed books. In the words of Lucien Febvre, the book rendered vital service to research by immediately transmitting results from one researcher to another… . By doing so, it gave their ideas a new lease on life and endowed them with unparalleled strength and vigour [sic]. They came to have a new kind of coherence and, by the same token, an incomparable power for both transformation and propagation. Fresh concepts crossed whole regions of the globe in the very shortest time, wherever language did not deny them ac­ cess.3 This produced a remarkable transition in both library history and the book trade, and in many ways is comparable to the transformation facing libraries to­ day. Febvre goes on to say that the mag­ nitude of reproducing texts in print was quite staggering at the time: “[S]oon the potential of the new process became ob­ vious, as did its rôle as a force for change as it began to make texts accessible on such a scale as to give them an impact which the manuscript book had never achieved.”4 With respect to the selection of texts for printing, Febvre reminds us that the process of setting up a printing shop— acquiring equipment and supplies—re­ quired a significant investment. There­ fore, 15th-century publishers only fi­ nanced the kind of book they felt sure would sell enough copies to show a profit in a reasonable time. We should not therefore be sur­ prised to find that the immediate effect of printing was merely to fur­ ther increase the circulation of those works which had already enjoyed success in manuscript, and often to consign other less popular texts to oblivion. By multiplying books by the hundred and then thousand, 60 College & Research Libraries January 2001 the press achieved both an in­ creased volume and at the same time more rigorous selection.5 Most of the earliest books were reli­ gious because clerics generated most of the demand. But as the demand for books broadened, “selection soon became im­ perative as the decision had to be made as to which of the many thousands of medieval manuscripts were worth print­ ing.”6 Febvre aptly mines the depths of this issue, but for the purposes of this ar­ ticle, what his research distills very clearly is that the selection of manu­ scripts for printing was based on the profitable demands of an elite society for education and knowledge. Later, this would expand to include a broader au­ dience with a thirst for learning. What is most strikingly similar in this limited comparison between the early days of printing and the current status of electronic technology is the need to develop a body of knowledge. Today, this endeavor supports an effort to re­ alize the potential of electronic technol­ ogy similar to what Febvre recognized when he said the printing effort sought “to make texts accessible on such a scale as to give them an impact the manu­ script book had never achieved.” We now seek to realize the potential of elec­ tronic technology “to make texts acces­ sible on such a scale as to give them an impact” that, indeed, surpasses what the printed book has achieved. If today’s libraries are to take any direction from the past, they would not go wrong to follow the example of their medieval counterparts and focus resources and attention on the education of the com­ munities they serve. For academic li­ braries, this is the research community, with faculty and students of the univer­ sity being the primary beneficiaries. Strictly followed, this entails allocating internal funds for the digital conversion of collections exclusively to support the needs of the immediate user community. It is this mandate of the academic library that precludes a focus on K–12 educa­ tion, leaving those needs for digital materials to places such as the Library of Congress, the New York Public Li­ brary, and other public libraries.7 The exclusive application of resources for specific mission-related uses and needs may be construed as something too obvious to point out. However, much of what has been scanned by libraries and archives to date are low-use special col­ lection materials, simply because they are “signature” collections. Although these efforts produce educational information sites, rarely do they actually produce a digital collection deep enough to satisfy the broader research needs of the local constituency. In effect, many of these sites are more suitable to the needs of the K– 12 audience, rather than higher educa­ tion. Given the costs of conversion, se­ lection decisions must remain organic to the mission of the parent institution, or the library stands to lose its credibility within the university and its scholarly structure. Overview of Past Collection-Building Practices Though in no way new, the question of how libraries should proceed in build­ ing a body of knowledge is quite daunt­ ing in the electronic environment. In a very broad assessment of what digital technology portends, Carla Hesse saw the current environment as an opportu­ nity to achieve the “most cherished ideal of modern democratic polities and the libraries they have created: universal ac­ cess to all forms of human knowledge.”8 For academic libraries and their select­ ing policies, the idea of collecting “all forms of human knowledge” is an old one. In the nineteenth century, early book selection in academic libraries began to adhere to the ideal of comprehensiveness and completeness.9 However, in the twentieth century, the idea of complete­ ness was given up as its practical impossibil­ ity came to be realized. The cause was the enormous proliferation of Selection for Digital Conversion in Academic Libraries 61 knowledge and the resulting vast increase in publication.10 Now, with the technology and the de­ mand in place to convert library materi­ als into digital form, we find ourselves in a situation comparable to our earlier counterparts. As in the traditional print environment, the stultifying problem of overabundance amplifies the need for The choices involved in microfilm­ ing brittle books are recognizably similar to the choices libraries must make in deciding what to reformat digitally, but that intersection is brief and extends only backward, not forward. selectivity and increases the need for pro­ fessionals to do the selecting: “the more there is to select out there, the more sub­ ject expertise is needed to select quality and specifically in order to satisfy needs and demand.”11 In the early days of collection build­ ing in American academic libraries, most collections were built by faculty.12 J. Periam Danton emphasized the vagaries that resulted in this practice: The majority of titles in the book stock of the typical American uni­ versity library are there as the re­ sult of scores of thousands of indi­ vidual, uncoordinated, usually iso­ lated decisions, independently made by hundreds of faculty.13 Danton was not the earliest or only critic of the system that permitted fac­ ulty the exclusive responsibility for se­ lecting library materials. Corroborating this view, Raven Fonfa cited the period 1876–1939 as a period of widely shared discontent and criticism among librar­ ians and others, noting that data col­ lected for the Waples and Lasswell study, published in 1936 and entitled National Libraries and Foreign Scholarship, stated that collections developed by faculty in academic institutions were both unbal­ anced and lacking, whereas collections developed by librarians in public insti­ tutions “showed significantly more bal­ anced holdings.”14 Danton also cited the Waples and Lasswell study, which found that Harvard and the universities of Chi­ cago, California, and Michigan had sig­ nificantly lower percentages of 500 En­ glish, French, and German works in the areas of social sciences, “judged by spe­ cialists in those fields to be of primary scholarly importance,” whereas the New York Public Library, “where book selec­ tion is … entirely the responsibility of a corps of subject specialist librarians, held 92 percent.”15 In the years after 1939, as librarianship became more professional, librarians began to win support for col­ lection development in academic librar­ ies and the theories and practices that in­ form materials selection evolved simul­ taneously. Undoubtedly, at its most basic level, the “selection of materials has almost al­ ways been based on clientele.”16 In the academic environment, the university, as the parent institution, dictates the devel­ opment of collections for the research and teaching needs of the faculty and stu­ dents. Ross Atkinson, clarifies this con­ dition by saying, [w]hile the individual library can make the micro-decisions concern­ ing the particular items to which access should be provided, the broader policy decisions that define the parameters within which the library’s collection building effort must operate, are largely based upon stipulations made in advance by the supported (and in support of) [the] user community.17 As stated above, the huge expense of digital conversion of library and archive materials requires a fiscal responsibility to the academic library’s parent institu­ tion. In fact, this responsibility to the uni­ versity is heightened further by poten­ tial misjudgments and technical vagar­ ies that could grossly waste precious re­ http:community.17 http:faculty.13 http:faculty.12 http:publication.10 62 College & Research Libraries January 2001 sources and would, in the end, represent a glaring disservice to the library’s cli­ entele. With this in mind, there is an amplified need for an implementable plan that adheres to the basic tenet that collection development—what John Rutledge and Luke Swindler defined as the “macro-decision”—should coincide with the directives of the parent institu­ tion.18 In addition, libraries’ collecting decisions should be consistent and cog­ nizant of holistic responsibilities. Hendrik Edelman broadly defined col­ lection development as the first level in the development of a policy or plan that considers the goals of the library as far as the collections are concerned, taking them into account and correlating them with the environmental as­ pects such as audience demand, need, and expectation, the informa­ tion world, fiscal plans, and the his­ tory of the collections.”19 In the digital environment, conformity to this established principle is certainly possible and looms as a fiscal imperative. However, David Fielding and Carl Lagoze raised a good point in the con­ text of digital libraries when they asked, “is it really necessary to select materials in specific groupings or ‘collections’ to begin with, or just proceed on [a] use basis, or curriculum needs basis?”20 In other words, does the “distributed digi­ tal library” or “virtual library” obliter­ ate the need to follow traditional collec­ tion development principles as they have been applied broadly to the macro-col­ lection-building function? Indeed, digi­ tal technology affirms the capability to forego the need for one institution to build and develop isolated “collections,” especially when such an endeavor can otherwise be accomplished on a multi- institutional, collaborative basis with in­ dividual institutions contributing on what Rutledge and Swindler call a mi- cro-decision level. As long as an institution’s selection efforts coincide with the goals of the parent institution, cooperatively built virtual collections can satisfy the needs of the immediate user community as well as the pressing need to be fiscally responsible. Proceeding from the idea that tradi­ tional (macro-level) collection building could be jettisoned, the need to formu­ late a strategy to build digital collections still persists. It is possible that the clue to a coherent selection strategy is em­ bedded elsewhere in traditional collec­ tion development practices. Moving on to Edelman’s second level of collecting practice, he defined the title-by-title se­ lection process as implementation of the library’s overall collection development policy.21 Rutledge and Swindler called this the micro-decision and set out the six “most relevant” factors that a selec­ tor must consider in title-by-title selec­ tion: “(1) subject, (2) intellectual content, (3) potential use, (4) relation to collec­ tion, (5) bibliographic considerations, and (6) language.”22 These categories could be adopted easily in a micro-level decision-making process for electronic collection development; however, they are too broad and too inclusive to be selective enough. Even more than in the print world, the caution to “remember that selection implies selectivity” is pro­ foundly relevant.23 Here, at this micro- level of collection building for the digi­ tal library, is where the real crux of the problem exists. Following the estab­ lished traditions of the omnivorous li­ brary (“which sees nothing as out of scope”) is no longer appropriate, nor is the “just-in-case” model of selecting.24 Both are far too inclusive to be compat­ ible with the implied fiscal responsibil­ ity of digital technology.25 Selection Methodologies for Preservation A third avenue to explore as a path to­ ward a digital conversion strategy lies in the methodologies that evolved to sup­ port the decision making for the preser­ vation of brittle books using microfilm technology. This, too, is a reformatting http:technology.25 http:selecting.24 http:relevant.23 http:policy.21 Selection for Digital Conversion in Academic Libraries 63 decision and, perhaps, explains why there is a persistent practice in the library community of linking preservation and digitization particularly with regard to selection activities. The choices involved in microfilming brittle books are recog­ nizably similar to the choices libraries must make in deciding what to reformat digitally, but that intersection is brief and extends only backward, not forward. Whether driven by the demands of re­ searchers or the demands posed by poor condition, the similarity between micro­ film and digital technologies lies only in the initial desire to duplicate what al­ ready exists. Beyond that, the two tech­ nologies diverge significantly in pur­ pose. Nevertheless, a brief tour of the selection decisions established for pres­ ervation microfilming is a useful exercise and possibly may distill a vision of pur­ pose in developing digital collections, even if only in the obverse. What follows is a brief description of the decision pro­ cesses applied to microfilming. Tangentially, it is too tempting to resist the observation that digital technology may serve to rescue much of what exists on microfilm—a format that researchers tend to avoid. An early and popular selection meth­ odology for brittle books was developed in the 1980s and adopted by the Research Libraries Group (RLG) for use in coop­ erative preservation microfilming projects. The “clean-sweep” approach utilized date parameters applied to sub­ ject areas of a collection. The date param­ eters attempted to approximate those years in which paper manufacture pro­ duced highly acidic paper, 1870 and 1910.26 The distinctive feature of this ap­ proach was its comprehensiveness: It sought to include all materials within a subject area between the chosen date pa­ rameters on the grounds that what had been collected in the past would have potential use in the future. Initially, this method had great attraction primarily because “little time is expended on deci­ sion making… . And it is argued that there might someday be a use even for materials whose importance is not evi­ dent at present.”27 Thus, the merit of this approach lies squarely in its lack of se­ lectivity. The efficiency of such a prospec­ tive approach works well when applied to micr ofilm technology, due to microfilm’s low storage costs. But the in­ herent weakness of the clean-sweep ap­ proach “is that materials which may never be needed by scholars take up time and money and thus displace more im­ portant materials that aren’t in the cho­ sen group.”28 High labor costs associated with pre- and postmicrofilming activities hardly justified the benefits of its built- in decision-making efficiencies, and the clean-sweep method of selection fell out of favor. In the electronic world, the va­ garies of this approach are equally intol­ erable given the high cost of digital con­ version. Therefore, this paradigm is clearly one to avoid. In an effort to improve upon the clean- sweep approach, the RLG later fostered the concept of great collections based on the RLG Conspectus rating of a collection as a measure of its worth for preserva­ tion.29 This selection method also was subject driven and focused attention on those comprehensive research collections, built over time, that were in danger of disappearing due to embrittlement. The great collections approach retained many of the characteristics endemic to the clean- sweep approach but also introduced the element of physical condition. Date pa­ rameters were still observed, but if an item was not brittle, the great collections ap­ proach excluded it from microfilming. Comprehensive in its approach, this method also was not very selective. Again, microfilm technology lends itself well to the massive attempt to reformat brittle books, but evaluated from the perspective of digital technology, the problem of what to do with low-use materials resurfaces. In the digital environment, it makes little sense to enhance access to low-use mate­ rials because it is difficult to justify their costly conversion. 64 College & Research Libraries January 2001 A more practical approach to selec­ tion for preservation was provided by Ross Atkinson, in “Selection for Preser­ vation: A Materialistic Approach,” in which he discussed three classes of li­ brary materials and their appropriate corresponding preservation treatment.30 Class 1 includes special collection and unique materials, such as rare book and manuscripts; class 2 “consists of high- use items that are currently in demon­ strable demand for curriculum and re­ search purposes”; and class 3 are low- use, or less frequently used, research materials.31 The preservation treatment that Atkinson advises for these group­ ings of materials are conservation treat­ ment for class 1; replacement and/or preservation photocopying for class 2; and microfilm reformatting for class 3. This is an acceptable formula, generally, for organizing preservation treatments, and it works well for low-use materials when microfilm is your reformatting tool of choice. Unfortunately, like the previous two methods described above, this approach is not appropriate either. A strict translation of Atkinson’s para­ digm would substitute digital conver­ sion as the reformatting agent for class 3 materials instead of microfilm—a fruitless and costly undertaking for low- use materials. Before leaving Atkinson’s model, it cannot be overlooked that many, if not most, digital projects to date have fo­ cused largely on Atkinson’s class 1 spe­ cial collections and unique materials. These efforts have attracted funding and attention and have provided fer­ tile ground for testing this new tech­ nology. Whether they do now or ever will suit the research objectives of aca­ demic libraries is debatable. Research requires in-depth collection building. Converting that depth to electronic technology in a specific subject area for scholarly use would be extremely ex­ p e n s i v e a n d i n e ff i c i e n t b e c a u s e i t would lead to the conversion of mate­ rials that are used by only a small seg­ ment of researchers. In addition, build­ ing electronic collections on this basis requires the bibliographer, as in print collections, to speculate and project which materials will be needed by the researcher, an activity that has been deemed both subjective and difficult, often leading to overbuying. Still pursuing the analogy of selection methodologies for preservation micro­ filming, additional strategies include the condition-and-use model and the edito­ rial model. Selection based on condition and use for preservation microfilming was an approach first suggested by Christenger Tomer in 1979 (and again in 1985) and then later by Barclay W. Ogden in 1987.32 Proponents of this method ar­ gue that scarce resources should be allo­ cated for preservation based on poor con­ dition and the amount of use an item re­ ceives, the theory being that the combi­ nation of use and poor condition places an item in a higher-risk category than those items that remain untouched on the library shelf. Tomer logically explained that the documents at most serious risk are those whose interest to readers exceeds in longevity their physical capacity to support the consequent handling.33 As part of this process, identification of items in need of preservation occurs at the point of use. Condition plays the largest role in the decision and, thereby, makes access to materials currently in demand its initial priority. Of course, bibliographic review of these materials is essential because everything used in a library is not necessarily worth pre­ serving. The condition-and-use selec­ tion decision proceeds on a title-by-title basis and assumes that the heart of the preservation mission should first con­ sider the end user for whom the mate­ rial is saved. The condition-and-use selection method brings us closer to a translatable decision-making paradigm for digital conversion primarily because of its fo­ http:handling.33 http:materials.31 http:treatment.30 Selection for Digital Conversion in Academic Libraries 65 cus on use; however, the element of con­ dition is misplaced. In the digital envi­ ronment, it is important to distinguish between the factors that motivate a pres­ ervation decision and those that seek to enhance access. The two decisions may “overlap where materials are both en­ dangered and in demand,” but in the context of collection building (i.e., creat­ ing a body of knowledge in a new me­ dium for research purposes), selection decisions for digital conversion need not consider physical condition of the origi­ nal.34 Digital technology is ideal when used to enhance access but assists in pres­ ervation only when creating surrogate copies of materials that are likely to ben­ efit by reduced handling. It cannot be used to actually preserve an item.35 The editorial model of selecting ma­ terials for retrospective collection devel­ opment has been employed commonly by commercial microfilmers as a method of distributing collected works and genres. It was not widely used as a method for selecting materials for pres­ ervation microfilming. Only a few pres­ ervation-microfilming projects explored the use of editorial boards of scholars and national bibliographies to identify im­ portant titles and core materials. Most notable among them are the American Theological Society Serials Project and the American Philological Association’s effort to preserve classics materials.36 It is not entirely clear why the editorial method was not used more widely to make preservation decisions. Perhaps the general distaste among scholars for microfilm is responsible, compounded by a fractious relationship between fac­ ulty and the library in the wake of large microfilming grants. Digital technology seems to be bridging that chasm as librar­ ies look to their respective faculties to assist them in selection decisions for digi­ tal conversion. Currently, several conver­ sion projects are using this method to develop digital collections in specific subject areas: the Perseus Project and the National Agriculture Library’s CORE Project.37 For the most part, it seems that selec­ tion models for preservation microfilm­ ing are inapplicable to digital technology. The clean-sweep approach is wholly un­ suited to guide the selection decision for digital conversion because it is far too in­ clusive and would involve the reformat­ ting of items that may never be needed or used. Likewise, the great collections approach is equally inappropriate for the same reason. Although Tomer ’s condi­ tion-and-use approach is more appeal­ ing, strictly applied, it is only appropri­ ate in digital conversions where protec­ tion of the original is an issue. It appears that even though the choices involved in digital conversion are reminiscent of the choices required for preservation micro­ filming, they are not translatable. True, both approaches involve a reformatting decision, but their trajectories diverge from there because their intents conflict in purpose. At the most basic level, one thwarts or decreases access whereas the other enhances or increases it. Tangentially, it is too tempting to re­ sist the observation that digital technol­ ogy may serve to rescue much of what exists on microfilm—a format that re­ searchers tend to avoid. Just as Febvre recognized that printing in the fifteenth century “resurrected long-forgotten writ­ ings in which the fifteenth century seem[ed] to have new interest,” it is pos­ sible that when libraries begin to offer on-demand digital conversion of micro­ filmed materials, much of the scholarship ”hidden” on microfilm will be “resur­ rected” in the same way.38 Conversely, Febvre’s notion that printing was respon­ sible for “consigning … less popular books to oblivion” has a rather unattrac­ tive, yet valid, modern-day correlation to the reformatting of brittle books to mi­ crofilm. Conclusion Salient points taken from the collection- building and selection decision-making models offered above settle most harmo­ niously around the overriding directive of research libraries to align their collec­ http:Project.37 http:materials.36 66 College & Research Libraries January 2001 tion development practices with their parent institutions and the utility of do­ ing so. The idea of use, especially high use, is fundamental to collection devel­ opment and is the common thread in all selection decisions. Coupled with the ex­ traordinary access capabilities permitted by digital technology, use holds signifi­ cant promise as a guiding factor in se­ lecting materials for digital conversion. It represents an opportunity to extend more resources and offer highly im­ proved services to the academic library’s local user community—undergraduates, graduates, faculty, and scholars alike. Whether digitizing core collections, fol­ lowing a curriculum-based approach, or creating partnerships with faculty, col­ lection-building efforts based on use are more likely to garner the support of the parent institution. A use-based directive never suc­ ceeded in preservation reformatting, but it has been widely accepted and, indeed, very successful in replacement and physical treatment decisions in collec­ tions conservation activities—where funding tends to be most scarce and de­ cision making most practical. Here, Atkinson’s proposal for class 2 materi­ als is quite relevant. In its application to digital conversion, it offers three advan­ tages: first, it could dovetail nicely with the work of faculty and students and thus support the teaching process and curriculum. Second, high-use items suf­ fer from the wear and tear of heavy han­ dling, which digital reformatting would alleviate, or at least reduce. And third, because high-use materials are the most likely to be duplicated across academic collections, it might foster opportunities for interinstitutional collaboration. In fact, because use is often similar across institutions, especially at the core collec­ tion and curriculum levels, the latter could lead to the development of a digi­ tized core collection and, in so doing, force the issue of cooperation to exploit the efficiencies and economies of shar­ ing, carefully tailored to avoid duplicate efforts and expenditures. In Atkinson’s parsing of library mate­ rials into three classes, he says, the objective of class 2 [high-use] preservation … is to preserve ma­ terials currently being used, or very likely to be used as projected on the basis of what is currently being used. It is in class 2 preservation, moreover, that bibliographers have the most important role to play in the preservation process, for the knowledge amassed by bibliogra­ phers as the current needs and ac­ tivities of the users and the current trends in the subject are precisely the criteria that must be applied to class 2 preservation selection deci­ sions. Class 2 preservation is, in fact, really only an extension of or supplement to the core building and maintenance done by most se­ lectors in most libraries.39 This statement is just as pertinent when read in the context of today’s en­ vironment and applied to selection ef­ forts for digital conversion, using the same argument in favor of converting materials in support of an institution’s core curriculum. Demonstrable use mo­ tivated current efforts such as the Na­ tional Science Foundation’s Digital Li­ brary Initiative Phase II (1998–2002), which seeks to “explore the linking of digital library research efforts and testbeds for undergraduate education” in science, mathematics, engineering, and technology.40 Similarly, one of the find­ ings of the studies conducted as part of the Columbia On-line Book Project was that faculty had a high regard for the usefulness of offering students Internet access to reading assignments.41 Except for a few examples such as these, however, it seems that a use-based approach still is no more popular as a selection criterion for digital conversion than it was a selection criterion for pres­ ervation microfilming. 42 Having re­ viewed an array of possible selection cri­ teria and approaches deployed in the http:microfilming.42 http:assignments.41 http:technology.40 http:libraries.39 Selection for Digital Conversion in Academic Libraries 67 past to develop book collections and pre­ serve brittle books in academic libraries vis-à-vis digital conversion, it is surpris­ ing to discover what is actually driving the endeavor to create digital collections in academic libraries today. An informal survey of twenty-five current digital projects in academic libraries showed that the most popular approach to select­ ing collections for digital conversion is a subject-and-date-parameter approach applied, by and large, to special collec­ tions, with little regard for use, faculty recommendations, scholarly input, edi­ torial boards, or curriculum.43 When queried about their goals, project man­ agers most often responded that im­ proved and/or enhanced access was the primary goal of converting collections to an electronically accessible format. It is hard to imagine that a broad-based local user community benefits by the im­ proved access to special collections. It is only a matter of time until the question emerges as to how long the parent insti­ tution will be satisfied with supporting the costly conversion of their library’s materials to improve access for narrowly defined audiences that may not even be their primary local constituents. Hark­ ing back once more to the mid-fifteenth century, the building of printed book col­ lections in libraries was driven initially by the education and research needs of the academic community. Simplistic as it sounds, half a millennium later, the education and research needs of the academy are still the academic library’s primary responsibility; thus, it must be prepared to account for a digital conver­ sion selection methodology that sup­ ports and complements that relationship. Notes 1. Clifford Lynch, “The Role of Digitization in Building Electronic Collections: Economic and Programmatic Choices,” in Selecting Library and Archive Collections for Digital Reformat­ ting, proceedings from an RLG Symposium held Nov. 5–6, 1995, Washington, D.C. (Moun­ tain View, Calif.: Research Libraries Group, 1996), 2. 2. Lucien Febvre and Henri-Jean Martin, The Coming of the Book, trans. David Gerard, eds. Geoffrey Nowell-Smith and David Wooton (New York: Verso, 1997), 19; first published as L’Apparition du livre (Paris: Éditions Albin Michel, 1958). 3. Ibid., 11. 4. Ibid, 248–49. Febvre says, “Some 30,000–35,000 different editions printed between 1450 and 1500 have survived, representing 10,000–15,000 different texts, and if we were to take into account those which have not survived the figures would perhaps be much larger. As­ suming an average print run to be no greater than 500, then about 20 million books were printed before 1500, an impressive total even by 20th-century standards.” However, he fur­ ther qualifies this estimate in a footnote saying, “Of course we mean only to indicate some idea of scale. According to Vladimir Loublinsky, production would be somewhere between 12 and 20 million copies.” 5. Ibid., 249. 6. Ibid., 260. 7. An exception to this would be when grant and/or donor funds are available to the academic institution to scan and make available on the Internet educational materials that would benefit an outside constituency or, perhaps, in those cases where the institution of higher education is publicly funded and mandated to provide K–12 access to their collec­ tions. The idea expressed here—that the academic library’s commitment to higher education (to the exclusion of K–12) as part of their mission—is most appropriate to research libraries and is based on a historic and economic perspective that acknowledges the traditional pur­ poses of the academic library. This, however, is an emphasis that could undergo radical changes in the future. See the Library of Congress’s “American Memory Collections” for an excellent example of selection for a broader constituency that includes K–12 and others. 8. Carla Hesse, “Humanities Scholarship in the Digital Age,” in What’s Happened to the Humanities, ed. Alvin Kernan (Princeton, N.J.: Princeton Univ. Pr., 1997), 110–11. 9. J. Periam Danton, Book Selection and Collections: A Comparison of German and American University Libraries (New York: Columbia Univ. Pr., 1963), 34. 10. Ibid. http:curriculum.43 68 College & Research Libraries January 2001 11. William S. Monroe, “The Role of Selection in Collection Development: Past, Present and Future,” in Collection Management for the 21st Century (Westport, Conn.: Greenwood Pr., 1997), 114. 12. Danton, 71. See also Raven Fonfa’s, “From Faculty to Librarian Materials Selection: An Element in the Professionalization of Librarianship,” in Leadership and Academic Librar­ ians, eds. Terrence F. Mech and Gerard B. McCabe (Westport, Conn.: Greenwood Pr., 1998), 22–36. 13. Danton, Book Selection and Collections, 74. 14. Fonfa, “From Faculty to Library Materials Selection,” 29. 15. Danton, Book Selection and Collections, 75. 16. Robert N. Broadus, “The History of Collection Development,” in Collection Manage­ ment: A New Treatise, eds. Charles B. Osburn and Ross Atkinson (Greenwich, Conn.: Jai Pr., 1991), 9. 17. Ross Atkinson, “The Conditions of Collection Development,” in Collection Manage­ ment: A New Treatise, eds. Charles B. Osburn and Ross Atkinson (Greenwich, Conn.: Jai Pr., 1991), 36. 18. John Rutledge and Luke Swindler, “The Selection Decision: Defining Criteria and Es­ tablishing Priorities,” College & Research Libraries 48, no. 2 (Mar. 1987): 125, 128. 19. Hendrik Edelman, “Selection Methodology in Academic Libraries,” Library Resources & Technical Services 23, no. 37 (winter 1979): 34. 20. David Fielding and Carl Lagoze, “Defining Collections in Distributed Digital Librar­ ies,” in D-Lib Magazine (Nov. 1998). Available online at: www.dlib.org/dlib/november98/ lagoze/11lagoze.html. 21. Edelman, “Selection Methodology in Academic Libraries,” 34. 22. Rutledge and Swindler, “The Selection Decision,” 128. 23. Monroe, “The Role of Selection in Collection Development,” 107. 24. Ibid., 106. See also Broadus, “The History of Collection Development,” 3–28, for a deeper treatment on the subject of size of collections and for an excellent bibliography on the history of collection development. 25. Interestingly, John Price-Wilkins takes the opposing view and argues in favor of just- in-case collections, and has applied the just-in-time selection model to the decision to create and store derivatives (i.e., images derived from larger TIFF images) in “Just-in-time Conver­ sion, Just-in-case Collections,” 5 Aug. 1999. Available online at: http://www.dlib.org/dlib/ may97/michigan/05pricewilkin.html. 26. For an excellent overview of selection methods for preservation, see Margaret S. Child, “Selection for Preservation,” in Advances in Preservation and Access, vol. 1, ed. Barbara Buckner Higginbotham (Westport, Conn.: Meckler, 1992). 27. Roger S. Bagnall and Carolyn L. Harris, “Involving Scholars in Preservation Deci­ sions: The Case of the Classicists,” Journal of Academic Librarianship 13, no. 3 (July 1987): 141. 28. Ibid., 141. 29. Child, “Selection for Preservation,” 149–151. 30. Ross Atkinson, “Selection for Preservation: A Materialistic Approach,” Library Resources & Technical Services 30 (Oct./Dec. 1986): 344–48. 31. Ibid., 346. 32. See Christinger Tomer, “Identification, Evaluation, and Selection of Books for Preser­ vation, “ Collection Management 3, no. 1 (spring 1979) and “Selecting Library Materials for Preservation,” Library and Archives Security 7, no. 1 (spring 1985). See also Barclay W. Ogden, “Preservation Selection and Treatment Options,” in Preservation: A Research Library Priority: Minutes of the 111th Meeting of the Association of Research Libraries (Washington, D.C.: ARL, 1987), 38–42. Margaret Child’s summary of the condition and use methodology is excellent in “Selection for Preservation,” 153–55. 33. Tomer, “Selecting Library Materials for Preservation,” 2. 34. Janet Gertz, “Selection Guidelines for Preservation,” in Guidelines for Digital Imaging: Papers Given at the Joint National Preservation Office and Research Libraries Group Conference in Warwick 28th–30th Sept. 1998. Available online at: http://www.rlg.org/preserv/joint/ gertz.html. See also her more recent article, “Selection for Preservation in the Digital Age: An Overview,” Library Resources and Technical Services 44, no. 2 (Apr. 2000): 97–104. 35. Paula De Stefano, “Digitization for Preservation and Access,” in Preservation: Issues and Planning (Chicago: ALA, 2000): 307–22. See also Stephen Chapman, Paul Conway, and Anne Kenney, “Digital Imaging and Preservation Microfilm: The Future of the Hybrid Ap­ proach for the Preservation of Brittle Books.” Available online at: http://www.clir.org/cpa/ archives/hybridintro.html#full. 36. The American Theological Society’s serials microfilming project lasted more than http://www.clir.org/cpa http://www.rlg.org/preserv/joint http://www.dlib.org/dlib www.dlib.org/dlib/november98 Selection for Digital Conversion in Academic Libraries 69 twenty years in which “selection seems to have been mainly by consensus among the theo­ logical librarians who constituted the Board,” see Child, “Selection for Preservation,” 148. For the American Philological Association’s microfiche project, see Bagnall and Harris, “In­ volving Scholars in Preservation Decisions: The Case of the Classicists,” 140–46, and “Who Will Save the Books? The Case of the Classicists,” New Library Scene 6 (Apr. 1987): 17. 37. See Mary Summerfield, “Online Books: What Roles Will They Fill for Users of the Academic Library?” 23 Apr. 1999. Available online at: http://www.columbia.edu/cu/librar­ ies/digital/texts/paper). See also the “Perseus Project: An Evolving Digital Library,” ed. Gregory Crane, Tufts University. Available online at http://www.perseus.tufts.edu, July 16, 1999; Entlich, R. et al, “Making a Digital Library: The Contents of the CORE Project, “ ACM Transactions on Information Systems 15, no. 2 (Apr. 1997): 103–23. 38. Febvre and Martin, The Coming of the Book, 260. 39. Atkinson, “Selection for Preservation,” 346. 40. “Digital Libraries Initiative—Phase 2,” sponsored by the National Science Founda­ tion, Defense Advanced Research Projects Agency, National Library of Medicine, Library of Congress, National Aeronautics & Space Administration, National Endowment for the Hu­ manities, in partnership with the National Archives and Records Administration and the Smithsonian Institution. Final draft version: Feb. 18, 1998; announcement number NSF 98-63 (NEW). 41. Summerfield, “Online Books.” 42. Paula De Stefano, “Use-based Selection for Preservation Microfilming,” College & Re­ search Libraries 56 (Sept. 1995): 409–18. 43. This author queried twenty-five project managers at academic libraries through an informal e-mail survey. Many thanks to those individuals who responded to my question­ naire. Sources used to identify digital projects were: the National Endowment of the Hu­ manities “Top Humanities Web sites” (http://edsitement.neh.gov, 6/24/99); IFLA’s IFLANET Web site document, “Digital Libraries: Resources and Projects” (http://www.ifla.org/II/ diglib.htm, 7/16/99); Rutger ’s “Digital Library Resource Page,” 7/16/99; and the Univer­ sity of California, at Berkeley’s, SunSITE Hosted Projects (http://sunsite.berkeley.edu/R+D, 7/3/99). http://sunsite.berkeley.edu/R+D http://www.ifla.org/II http:http://edsitement.neh.gov http:http://www.perseus.tufts.edu http://www.columbia.edu/cu/librar