Berkeley Digital Library SunSITE

Redesigning Library Services: A Manifesto

6: Supplement

What Will Collection Developers Do?

In chapter 6, in our discussion of the shift from paper collections to electronic collections, we noted a major change in the dispensing role role: a much reduced premium on local storage compared with storage at a distance. Local storage (and local ownership) is no longer a necessary condition for convenient access. Local storage becomes optional rather than necessary. So what will collection developers do if and when the emerging environment of networked electronic resources were to lead to smaller local collections?

J. P. Migne's Patrologiae cursus completus, (Latin series), is a nineteenth century compilation of the writings in Latin of the early Christian fathers in 221 large volumes. This work has become available in electronic form, so it would appear that almost any material of interest to library users could become available in electronic form and offer, in principle, a credible alternative to the assembling of local collections. In the special case on indexing and abstracting services, the searching of remote databases has already tended to replace subscription to paper editions.

We are not concerned here with how completely, or under what circumstances, access to networked resources will replace local collections in paper, only to probe some consequences of a such a replacement.

The mere idea that access to networked electronic documents could substantially replace local collections raises intriguing questions about the purposes of local collections and the work of collection developers. What do collection developers do for library users? At the superficial level of process, what they do is clear: They make a stream of decisions telling the library's technical services staff which items should be acquired and cataloged for the local collection and which should be discarded or not acquired. But the purpose, as distinguished from the process, is less clear. There are large literatures on collection development, on how new library-related technologies are evolving, and about a shift in emphasis from ownership to access (e.g. Rutstein, DeMiller & Fuseler 1993). Important though these topics are, they are primarily concerned with process (with "how") and can distract attention from examination of purpose (of "why").

There has been a significant shift away from viewing technical services as being a grand apparatus for establishing usable local collections and towards a notion of technical services being the user's gateway to the bibliographic universe. (See, for example, Norgard, Berger, Buckland & Plaunt 1993; Buckland 1994). A comparable change in the perception of collection development seems likely.


Lionel McColvin's classic The Theory of Book Selection for Public Libraries (1925) begins with the following words:

"Book selection is the first task of librarianship. It precedes all other processes--cataloguing, classification, or administration--and it is the most important. No matter how thorough and efficient the rest of the work may be, the ultimate value of a library depends upon the way in which the stock has been selected." (McColvin 1925, p. 9).

McColvin starts with the need to have available the materials that will be in demand, not only because of the practical need to have in stock what users ask for, but also because there will be no benefit derived from acquisitions that are not used. No demand means no use and, therefore, no benefit, McColvin argues. But he also argues that a passive adapting of collections to demand would betray the mission of the library.

"If, however, we consider the library as a social force with the power to direct to some extent man's demand, (or, to use the usual expression, if we consider the library as an educational force) we will not be content to leave demand our only consideration." (p. 37, emphasis in original).

In brief, McColvin provides a frame with two dimensions--Demand and Value--which correspond to the dispensing and advisory roles of collections noted above. (For a convenient summary of McColvin's arguments, see, for example, Broadus (1973, chap. 2). For a broader historical background see Lunati (1975)).

There is a small, specialized literature by Lister and others on the quantitative techniques that can be used to adapt collections to patterns of demand. (See Note 2.) But much discussion of collection development, especially that concerning "balanced" collections, has to do with librarians' value-laden departures from purely demand-driven collection development. It is a matter of deliberately privileging some documents over others as a separate matter from the anticipation of demand. The books placed in the shelves, in the reader's face, so to speak, carry an implicit endorsement: These you should read; these are good books for you; or these are books you will like. Other materials, those not selected for (or weeded from) the collection, are actively (though implicitly) treated as less suitable for readers. They have been deliberately judged less satisfactory for whatever reason. It is important to recognize that this function, the privileging of some works over others, is different in kind from the narrowly logistic task of anticipating demand to maintain high immediate availability. The specialized literature on quantitative collection development techniques developed by Lister and others for ensuring high immediate availability is essentially irrelevant to this second task of privileging of some documents over others.


With materials on paper, the development of well-selected local collections dominates the quality of service, as McColvin stressed. There are gradations of local availability (on reserve, out on loan, in storage, etc.), but there is a basic, binary distinction: What is held locally is accessible and what not held locally is inaccessible. Recourse to interlibrary borrowing is an unsatisfying substitute for local holdings. Put simply, by the process of selection the collection developer is imposing a partitioning of the universe of library materials into two broad ranks: those that are to be made more accessible by being held locally; and those that are to be kept less accessible because not added to the local collection. Structurally, the effect is the same as the compilation of a selective bibliography and of an online search: Some items are selectively brought forward to the reader's attention--acquired, listed, retrieved respectively--the others, those not acquired, listed, or retrieved, still exist but they are left in less-accessible obscurity.

With networked, electronic resources, however, this binary distinction between local (and accessible) and non-local (and only inconveniently accessible) becomes less clear. In principle, all resources become equally (more or less) accessible. The distinction between locally-held and not-locally-held loses significance and, in a sense, all collections can become locally-accessible collections. A consequence relevant to technical services is that the local catalog, essentially a guide to what is locally-held, loses its past pre-eminence relative to union catalogs, remote catalogs, and bibliographies of network-accessible resources.


"In the library of the future there will be no library. All will be navigation!", we have heard from an academic acquaintance. This declaration does not imply an end to library service or to librarians (all good navigators, of course). Some of us may prefer "bibliography" to "navigation" as the term of choice and, obviously, some repositories (collections) of electronic materials must be located somewhere. What is placed in question is the future of the local collections that have hitherto dominated library service and have accounted for most of libraries' expenditures once the full costs of selecting, acquiring, processing, and housing local collections are properly attributed to the development of local collections.

Clearly there will continue to be local collections of two kinds:

  1. Materials on paper, microfilm, and other localized media for which the reader and the document must be in the same place; and

  2. "Caches" of electronic documents that are used often enough to justify keeping locally so long as demand remains high.

The population of documents in the localized caches will be transient and transparent. In general, users need neither know nor care which documents are stored locally and which are not at any given time. Automatic algorithms (essentially the same as those developed by Lister and others, noted above) can be designed, based on expected frequency of use, unit storage costs, and the costs of obtaining a copy from remote storage, to adjust the cache dynamically. It is a task for industrial engineers rather than subject specialists. Site licenses will need to be negotiated, but that seems likely to become more like negotiating a blanket order than traditional title-by-title, copy-by-copy book selection.

There are localized electronic media, notably CD-ROMs, that need to be selected and acquired like books or microfilms. But CD-ROMs can be put on networks and their contents can be stored in repositories. CD-ROMs seem a transitional technology or, at best, a temporary storage device.

As for the local collections in localized media (paper, microform, etc.), indications are that they will be a diminishing portion of what is used and certainly not what defines a library's ability to serve as they were in the past. So the question remains: What will collection developers do as local collections diminish in significance relative to networked electronic resources? Will their professional lives be enriched by the assignment of other, different duties? The answer lies in the purposes of what they do now, in the distinction between demand and value.

In traditional collection development, in building local collections of localized media, the single act of acquisition is the one response to both demand and value. One procedure addresses both concerns and one cannot know, by examining a book on the shelves, whether its acquisition resulted from an expectation of demand, from belief in its value, or from some combination.

In an electronic environment, however, considerations of demand and of value diverge because they require different courses of action. The logistics of catering to high demand, in detail a matter of a hierarchy of caches, can be delegated. It is a mechanical task and one could dispense with collection developers if that were all they did.

But what of the concern for value in the electronic library environment, for the privileging of some books over others? If there is more to collection development than responding to demand, then the value-laden role of privileging some resources over others needs to be continued unless the purpose of library service is to change. How is this other, remaining task of collection developers to be done? If it can no longer be done obscurely, combined with the logistical task of meeting demand by placing copies on the shelves, it will need to be addressed directly and separately.

Collection developers and technical services staff have been more closely co-conspirators in achieving the library's mission than libraries' organizational charts have indicated. Acquisitions departments and catalogers process only those items that the collection developers select. Collection developers and technical services staff play complementary and interdependent roles in establishing an ordering of the universe of documents, in determining the relative accessibility of different documents, for their local users. There is no clear reason why that purpose and that partnership should cease. Quite the reverse. As electronic resources multiply, the need for a convenient ordering, of differentiated accessibility, increases.

The privileging of the better and, by default, the non-privileging of the rest, remains a significant needed service. If we accept McColvin's statement for library service, written in relation to documents on paper, on what grounds would we deny it for library service using documents on diskdrives? Should the choice of technology for the storage medium determine the mission of the library? Let us re-read McColvin in relation to networked electronic resources: "If, however, we consider the library as a social force with the power to direct to some extent man's demand, (or, to use the usual expression, if we consider the library as an educational force) we will not be content to leave demand our only consideration."

Technical services have been evolving away from being the grand apparatus that constructs the local collection and has evolved to becoming the provision of navigational (alias bibliographic) tools to the universe of documents, electronic and non-electronic. But value judgements are still needed concerning which resources are most suitable for any given user group. It is contrary to common sense and to the central traditions of library service to make all material equally accessible. There is simply too much of it and, while we may agree that it ought all to be accessible, it would be unhelpful to make it all equally accessible. Some items are demanded more frequently than others, some may be regarded as more valuable than others. It has not been the purview of technical services staff to select which items should be privileged over others, but rather to implement that privileging for (and only for) the documents designated by the collection developers.

What collection developers have done in the past is to select items for local acquisition. The purpose of that process is to manipulate the universe of documents so that some--on grounds of demand and/or value--are made more accessible than others. Some documents are made visually prominent by being placed on the local shelves and others are deliberately not. As paper documents on shelves cease to be the technological medium of choice, different procedures are needed for a different technological medium, but the objective remains.

The design of a gopher service provides a simple example. Not all items can be equally accessible at the same, highest hierarchical level. It makes for efficiency if items that will be looked for often are given a privileged place high in the hierarchy of gopher levels. It makes for effectiveness if items that we regard as valuable for library users and that we wish to have seen by them are placed high in the hierarchy. Other items remain accessible, but can be left to deeper, less convenient levels of storage. If collection development is seen as deciding which items to privilege, then the need for those with that ability would appear to increase as a local paper collections diminish relative to networked electronic collections--and the traditional partnership between collection developers and technical services staff should be just as close.


What collection developers will do in the future with the new technology can be expected to differ in various procedural ways from what was done in the past with the old technology:

  1. Hitherto the privileging of documents has been dominated by a binary division: Items acquired for the local collection and those not acquired or not retained. In the environment of networked resources any such abrupt division seems improbable. A much finer gradation of degrees of accessibility and privileging seems likely.

  2. Hitherto all users of a given library have been supplied with one and the same collection. This "one-collection-for-all" approach has been technologically inevitable, but it is Procrustean rather than egalitarian. Different users have different needs. Users are unlikely to be equally well served by what the collection contains or by the way it is arranged. The popularity of branch and departmental libraries arises not only from geographical convenience but also from their being designed for a smaller group of users. With the new technology, different forms of access (multiple "clients," multiple "views") can be designed for different interest groups within the local population served.

  3. Because of the inherent localness of local collections, collection development work has been specific to each location and has resulted in massive geographical inequalities in library holdings. Library users with similar interests but located at different sites have received radically different service. With the new technology it may well be that the task can and will become more specific to topical areas than to locality. This would open new opportunities for cooperative efforts. Similar forms of access could be shared by those who have similar interests but who are at different locations.

  4. Because the evaluative, privileging role will no longer be combined with catering to demand, it will become a separate task and, therefore, a performance with greater visibility and accountability--as has already happened for catalogers (Hafter 1986).

  5. The notion of "materials budget" will evolve. Historically a component of the cost of making privileged documents more accessible, a different deployment is inevitable if the traditional purpose of library service is to be sustained in a changed environment.


What collection developers will do, depends on how one regards what they do now. At the superficial, procedural level, it seems that there will be a much reduced need for them. But if we are to take seriously the purposes underlying the procedures of library services, then, while new and different technology brings new and different procedures, the fundamental purposes and the expertise needed for selection (as opposed to acquisition) remain crucial. So, too, in this redesigned environment will the dependence of technical services on what collection developers do.

 Go to Chapter 7

Notes on the Supplement to Chapter 6 Collections Reconsidered

    1. An earlier version of this Supplement was published as: What will collection developers do? Information Technology and Libraries 14 (1995): 155-159.

    2. For example, Winston C. Lister, Least Cost Decision Rules for the Selection of Library materials for Compact Storage. Ph.D. dissertation, Purdue University, West Lafayette, Indiana, 1967 (ERIC ED 027 916); Richard W. Trueswell, "User satisfaction vs. s of holdings at three academic libraries," College and Research Libraries 30 (1969): 204-13; Michael K. Buckland, Book Availability and the Library User (New York: Pergamon, 1975); Stanley J. Slote, Weeding Library Collections. 4th ed. (Littleton, CO: Libraries Unlimited, 1997); John A. Urquart & N. C. Urquart, Relegation and Stock Control in Libraries (Stocksfield, U.K.: Oriel Press, 1976); Gary S. Lawrence, "A Cost Model for Storage and Weeding Problems," College and Research Libraries 42 (1981): 139-47; R. E. Staynor and V. E. Richardson, The Cost-Effectiveness of Alternative Book Storage Programs (Clayton, Victoria, Australia: Monash University, School of Librarianship, 1983).


    Broadus, Robert N. (1973). Selecting materials for libraries. New York: H.W.Wilson.

    Buckland. Michael K. (1989). The roles of collections and the scope of collection development. Journal of Documentation 45:213-226.

    Buckland, Michael K. From catalog to selecting aid. ALCTS Newsletter 5, no. 2 (1994): Insert A-D. ("From Catalog to Gateway: Briefings from the CFFC," 2).

    Hafter, Ruth. (1986). Academic Librarians and Cataloging Networks: Visibility, Quality Control, and Professional Status. Westport, CT: Greenwood Press.

    Lunati, Rinaldo. (1975). Book selection: Principles and procedures. Metuchen, NJ: Scarecrow Pr.

    McColvin, Lionel Roy. (1925). The theory of book selection for public libraries. London: Grafton.

    Norgard, Barbara A., Michael G. Berger, Michael Buckland & Christian Plaunt. (1993). The online catalog: From Technical Services to access service. Advances in Librarianship 17:111-148.

    Rutstein, Joel S., Anna DeMiller, & Elizabeth A. Fuseler. (1993). Ownership versus access: Shifting perspectives for libraries. Advances in Librarianship 17:33-60.

Copyright © 1997 Michael K. Buckland.
This text may be used for non-commercial personal and non-commercial education purposes provided that copyright ownership by Michael K. Buckland is acknowledged. All other rights reserved.
Document maintained at by the SunSITE Manager.
Last update July 2, 1998. SunSITE Manager: