The Art Museum Image Consortium (AMICO) Library, a collection of more than 50,000 catalogued images of art works held in 26 North American museums, is a wonderful thing—but perhaps not the thing everyone might want it to be. A teacher might want such a collection to supply the images needed for standard art history courses. A researcher might want it to provide a catalogue to the vast holdings of these museums. When something is new and different, we often attempt to understand it using familiar models—in this case the slide collection or the catalogue of a museum’s permanent collection. However, unlike a good slide collection, the selection here would disappoint anyone attempting to teach using AMICO images exclusively. Not surprisingly, architecture is represented only by a few salvaged fragments. The art of some periods and cultures will never be well represented in the AMICO Library, as long as it remains an effort of museums in North America, a continent that came late to the business of conquest and pillage and even private art collecting. The database offers a much better selection to someone teaching Impressionism or photography, but even so an instructor would probably need to add images from other sources. So far only a tiny fraction of the participating collections have been included. The database does not yet comprise a thorough catalogue for any of the museums involved. Yet it does grow rapidly. In its first “testbed” year only 20,000 images appeared. In this second year we have 54,000, and annual infusions of 20,000 images are planned. The number of museums also has increased.

Having explained what the AMICO Library is not, let us consider what it is. AMICO is a non-profit organization with the purpose of facilitating virtual access to museums. (The acronym is meant to be pronounced with the accent on the second syllable, as in an Italian friend rather than on the first, as with the major oil corporation. The list of members and plenty of other information may be found at: www.amico.org.) The dues-paying museums that comprise the membership benefit mostly from a coordinated delivery system and the opportunity to obtain technical expertise from the consortium. They receive none of the subscription fees. Dues and subscriptions pay for the collation and distribution of the AMICO Library. This product can be partly understood as an experiment in museum cooperation and an expansion of the Museums Educational Site License (MESL), a similar image sharing experiment that preceded AMICO. A third party, presently the Research Libraries Group (RLG) for most users, distributes the AMICO Library. Their role includes mounting the database on its computers, providing an interface for it, and authenticating users. As a subscription service from RLG, the AMICO Library may also be understood as an electronic publication, comparable to others distributed by RLG such as the Avery Index to Architectural Periodicals, and therefore may be held to the standards of quality expected from such publications.

As a publication, the AMICO Library looks very attractive; the packaging, in the broadest sense of that word, is well designed. This is especially important in electronic publishing. The web interface developed by RLG shines. Users find it easy to navigate, and this is no small feat. (RLG adapted their Eureka interface, adding image browsing and electronic shopping cart features.) The Consortium made some excellent decisions about how to provide access to this database of reproductions. Images in a variety of generous sizes (up to 1,024 by 768 pixels, ample for current projectors) display and download quickly. The categories of description accompanying an image have been well constructed to suit most uses without bogging the project down in minutia. Skillful planners have accommodated widely varied museum practices and emerging international standards. Having established this record structure, there must have followed a Herculean task of coordinating data from many museums, which doubtlessly contained ten variations of a name like Francesco Jose de Goya y Lucientes and an endless supply of similar problems. Most decisions about how to package the data have wisely balanced quality against efficiency, but a few may need to be revisited as the project matures. Color and gray-scale reference bars have seldom been included in the digital images. Though distracting, they do help users cope with variations in the quality of scans and displays. Although the system can apparently accommodate more than one picture of a work of art, a single photo is the norm, even for sculpture and decorative art objects. Links to closely related works (for example, two paintings on opposite sides of one panel) might be more systematically identified. But the design of the database usually succeeds. Its success results from an enormous amount of work that will be invisible to most users, just as excellent book design often makes itself invisible so that the reader’s focus can remain on content.

As regards content, the AMICO Library resembles a collection of essays of embarrassingly uneven quality. Even though users at most testbed institutions expressed satisfaction with the overall fidelity of images, dim scans can be found easily. The amount of description provided ranges from the briefest indications of the basics (creator, title, dimensions, media, ownership, the contact for publication rights, and accession number) to the lengthy discussions of subject matter, provenance, or other topics typical of full catalogue entries. One could argue that museums have been arbitrary in selecting works to digitize. Frequent browsing suggests that few, if any, of the museums have systematically chosen their most important holdings. Instead, curators, registrars, and photographers admittedly selected works because documentation was needed for other projects such as exhibitions, preservation work, or photograph requests. Museums made these selections independently and, in the rush to create a large database, there was little or no filtering. All member museums have not contributed equally. In fact, two organizations have supplied more than half of the data (The Library of Congress Prints and Photographs Division, with more than 20,000 works, and the Fine Arts Museums of San Francisco, with over 10,000). Five museums offered fewer than 100 works each and less than one percent of the database. Someone will surely delight at finding more than 1,200 Lewis Hine photos or the 742 Edward Hopper paintings and drawings. Others will be disappointed at finding only two Clarence White photos and five Warhol’s. Some will enjoy virtually rifling through the extensive Cabinet of American Illustration from the Library of Congress. In its present state, the AMICO Library functions more as a grab bag than a systematic corpus.

Users sometimes complain about such a haphazard assortment, and the AMICO members are themselves uncomfortable with it. Museums have been cautious about including recent art because of copyright and similar concerns, but AMICO has recently reached an agreement with the Artists’ Rights Society, that should facilitate inclusion of contemporary art. (CAA.Reviews readers may also want to check another review, <a href=http://www.caareviews.org/reviews/341
>A Museum Guide to Copyright and Trademark, by Michael A. Shapiro and Brett I. Miller and published by the American Association of Museums.) Now that museums have loaded their preexisting data into the AMICO Library, they have begun to show a greater interest in selecting works that users want (which was one of the important recommendations of the MESL project report). Usage logs, maintained by RLG, help identify user needs by listing heavily used works in the AMICO database and by recording the search statements that met with few or no results. When users at subscribing institutions have bothered to write with specific requests, they have been carefully considered and usually filled. In contrast, the Academic Image Consortium or the Society of Architectural Historians take a more orderly approach, by collecting images of works in standard textbooks for their websites. Yet a different complaint, of reinforcing an antiquated pedagogical canon, troubles those groups. Clearly, these very different types of digital image databases complement one another. But even if both kinds of project flourish, some amount of local image production will remain necessary, since the imagination of most people teaching art and art history defies containment.

Subscribing to the AMICO Library is more complex and expensive than subscribing to a magazine or newspaper. As is typical for electronic data services, subscription rates depend upon the number of people who will have access to the data. Generally, the costs seem high to faculty and slide curators, but impress librarians as slightly below average (as compared to business or science resources). Nearly all of the thirty subscriptions to universities have been purchased with library rather than departmental funds. This may be more than a matter of price. The MESL report predicted that digital image users would be more demographically diverse than slide users, including more students and faculty from a broader array of disciplines. One of the testbed studies of AMICO usage supports that supposition. AMICO requires a licensing agreement, which is also a normal practice with electronic subscription services. Potential subscribers will notice two, alternative licenses for universities posted on the AMICO web site. One is very long and requires that a subscriber monitor and report activity in ways that would be difficult in most environments. It is the earlier version of the license, based upon the thoughts of some of the MESL participants, and written with the guidance of a number of museum administrators and their counsel. The shorter license was developed during the first year of subscription activity—the result of negotiations with actual subscribers—and can be recommended more highly. This less restrictive version of the license does not provide access to images larger than 1,024 by 768 pixels. But these huge image files must be delivered offline and were seldom requested by the first year subscribers.

What is the value of the AMICO Library? For “e-teachers” it can be an important, but seldom their only source of digital images for classroom display. For those who want a few handy facts or to see an example of an artist’s work, it is a quick, but not a comprehensive, source of information. For the casually curious, AMICO offers some of the same benefits as a stroll through a museum. For those seeking unusual examples of a theme, or seldom published illustrations, the AMICO Library can provide a surprising view of the storage rooms of an assortment of North American museums. For example, a search for pictures of Pittsburgh reveals a number of unfamiliar views by a prolific Canadian draftsman C.H. White, who sketched during visits to several Midwestern cities at the turn of the century. Interesting, fresh material. AMICO is a good idea, and deserves whatever support we can give it. The idea of museums helping each other into their digital future has value, as does the goal of extending access to the educational experience of museums. Even notions of membership dues and subscription fees have merit, if they can keep digital stewardship of our cultural resources out of crassly commercial hands. As we navigate through web pages laden with the animated commercials of e-commerce start-up companies, it is difficult to remember that the Internet began as a means for a few educational and research organizations to cooperatively advance their missions. Cooperative image projects such as the AMICO Library and others explore that original, and still very real, value of a World Wide Web. They have a long way to go toward perfection, but lead us in a good direction.