DLF PARTNERS

DLF ALLIES

Comments

Please send the DLF Director your comments or suggestions.

Final report of a planning process for the Academic Image Exchange (now the Academic Image Cooperative) Prototype. A digital image resource for students, teachers, and scholars of The History of Art and related disciplines

Based on a document submitted to The Andrew W. Mellon Foundation on 20 August 2000 on behalf of the project Executive Committee.

D Greenstein
30 October 2000

Executive summary
Origin. The Academic Image Exchange
Evolution. The Academic Image Cooperative
Next steps
Statement of expenditure (intentionally omitted here)
Appendices
- Appendix A. AIC business plan (intentionally omitted here)
- Appendix B. AIC Prototype entity model
- Appendix C. Features of the AIC prototype system
- Appendix D. Executive Committee, advisory and review groups
- Appendix E. Technical standards and guidelines for image contributors
- Appendix F. Distribution services. Functional specification
- Appendix G. Draft collection strategy and development framework
- Appendix H. Shared cataloguing tool (forthcoming shortly)
- Appendix I. Statement of expenditure and estimated future costs (intentionally omitted here)

1. Executive summary

This document reports on a planning process for the Academic Image Exchange (AIE) Prototype - a process intended to develop a scaleable database of curriculum-based digital images to be used for teaching the history of art. In the course of the planning process, the project changed its name (to the Academic Image Cooperative or AIC - a name used throughout this documentation for consistency), constructed the prototype, and created a framework for a service capable of launching and sustaining an unparalleled and comprehensive scholarly resource that will promote innovation in research, learning, and teaching in the history of art and other arts and humanities disciplines that depend fundamentally upon the use of visual resources as evidence.

The planning process is still in train. With support from the Digital Library Federation, the AIC continues to assemble the remaining operational details that are required to launch a fully operational image service. Details are available below in Section 4, Next steps. This report is therefore a snapshot of ongoing activity. As such, it takes pains to distinguish between the work that has been completed and that which is still underway.

2. Origin. The Academic Image Exchange

As originally envisaged, the AIC was intended as a web-accessible database of public domain and copyright cleared visual image resources distributed freely or at the lowest possible price compatible with creation of an ongoing service, for bona fide educational use.

2.1. AIC Collection

The collection includes image resources that make up the "concordance core" - defined as the roughly 2,300 art historical objects represented in two or more of the ten leading art historical text books - and is intended principally to support teaching art history survey courses. Collection building was intended as a community exercise. Visual resource professionals and art historians, the hypothesis ran possess images in the public domain or which they hold copyright (many such individuals are themselves exceptional photographers). Further, initial soundings among members of relevant professional communities indicated a willingness to contribute images to a community resource that promised to serve a common educational good.

Achievement: The AIC has assembled a concordance for the ten leading textbooks with brief information about the art historical works that appear in two or more texts (including descriptions of the works as well as their location within texts).
Achievement: The AIC has assembled a collection of about 2,000 TIFF images that relate meaningfully to the concordance core. At present, approximately 800 precisely matching the concordance core, have been processed by Yale University to AIC standards.
Achievement: Brief descriptive records are available for all 2,000 of the TIFF images and indeed, for all works within the concordance core.
In process: The remaining 1,200 TIFF images in possession of the AIC are being processed at the University of Michigan, The entire corpus should be in fully processed form by January 2001.
In process: Full descriptive records (conforming to the Core Categories as recommended by the Visual Resource Assocation - VRA Core, version 3.0) are being developed for the 2,000 TIFF images in the possession of the AIC and for works in the entire concordance core. Full records will be available early in 2001.

2.2. AIC prototype system

Developing the functional equivalent of the traditional slide library and making this functionality available via the web to individuals and institutions with access to only standard web-browsing software was a further aim. The prototype was intended, for example, to support via most web-browsers a range of basic search and retrieval functions as well as image sequencing and side-by-side views as essential to their presentation for teaching and learning purposes. It was also intended to allow the user to navigate images with references to the hierarchy of art historical works as represented in the source textbooks.

Achievement: The AIC developed an entity model to support its intended features. This entity model is available in Appendix B.
Achievement : The AIC developed a prototype supplying the functionality discussed above. Appendix C uses screen dumps to introduce the prototype and its features.
In process: The AIC is developing a metadata model that conforms to the Core Categories recommended by the Visual Resources Association in the recently published version 3.0. The model will be available by the end of November 2000.

2.3. Community building

Regular expert assessment and evaluation was an inherent part of the planning, system design, collection building, and business planning. The AIC has been steered by an Executive Committee and guided by informal advisory and technical groups. The groups included representatives of key user communities (art historians, visual resource professionals, scholarly organizations, universities, colleges, independent scholars, etc.). In addition, the planning process was punctuated by expert workshops convened to review progress and shape developmental trajectories and priorities. The prototype itself was demonstrated at key professional meetings whose participants were invited through question-and-answer sessions, exhibitor demonstrations, and personal follow-up correspondence to reflect on various aspects of the prototype and its aims. Finally, the AIC has maintained a web site through the DLF (http://diglib.isproductions.net/collections/aic.htm) where information about it and the text of a printed promotional flyer ( (http://diglib.isproductions.net/collections/aic/aic.pdf) are made available and kept up to date. Appendix D supplies the names and institutional affiliations of those who have participated on the Executive Committee and the advisory and technical groups. It also lists expert workshops (with pointers to supporting documentation) and public presentations.

Achievement: The AIC has developed a community of professionals that take a deep and meaningful interest in and help inform and review its activities. That community continues to inform the AIC's further development.
Achievement: The AIC has secured the ongoing sponsorship of the College Art Association, the principal professional organization of artists and art historians. CAA also plays a significant advocacy role in the public policy arena, having played an activist role in discussions of copyright and digital images. CAA is itself a scholarly publisher with close ties to trade and university publishers in the art history field, and has facilitated discussions between these publishers and the AIC.
Achievement: The AIC has secured expressions of support from the Visual Resources Association, whose membership is crucial to the AIC's success.

3. Evolution. The Academic Image Cooperative

As of February 2000 the AIC met all its original objectives. It then exceeded them as it turned attention in March 2000 to defining and documenting the policies, practices, and business strategies that could sustain the AIC while developing its collections to include images outside the concordance core. The focus on business planning was a direct response to public evaluation and assessment of the AIC prototype as demonstrated at the College Art Association's 2000 annual meeting and at other professional gatherings (Appendix D). Feedback encouraged numerous changes to the prototype's look and feel as was expected. It also suggested reconsideration of some of the AIC's founding assumptions as enumerated below.

3.1. Commercial assumptions

The AIC rested on an assumption that its cost profile would start out high as the concordance core and the online service capacity were developed and then diminish precipitously as development work shifted into more of a maintenance mode. It was also assumed from the outset that while external support would be essential at startup, the AIC would have to be self-sustaining over the long run. The engagement of the art historical and visual resource communities was seen as the principal mechanism for achieving sustainability. Collections would grow at little or no cost through individual and corporate community contributions of images and data; volunteer domain specialists identified by CAA would ensure the quality of these contributions and legitimize the AIC collections in the eyes of scholars; the effort of image processing and cataloging (all to AIC standards) would be distributed, wherever possible, to contributors; and the proposed use of open standards (XML) for delivery would reduce the AIC's need to create end-user tools and enhance the end user's ability to use local tools to deliver AIC content. It was equally evident, however, that the AIC would have to be a centralized service to some extent, with staff for (re)cataloging and (re)processing images, communicating with contributors and subscribers, and marketing and delivering the service. The project team imagined that this centralized service might most economically be aligned with a large educational institution or professional organization, especially an institution for which the AIC would represent an opportunity to accelerate a "switch" from analog to digital already envisioned.

These assumptions were fundamentally challenged, not because they are erroneous, but because participants in AIC presentations and focus groups emphasized their need for well-managed, richly documented collections of consistently high-quality digital images and their skepticism that minimally processed community-evolved collections were likely to possess these traits. Respondents also recognized the costs involved in the development of high-quality collections and almost universally stressed a preference for paid access to such collections over free access to collections that were more "serendipitously" derived. Driven in part by its respondents' needs but also by the knowledge of costs incurred in documenting, processing, and creating a prototype delivery system for a relatively small image set supplied by only three contributors, the project team began to consider subscription-based models for supporting its activities, and a more managed approach to the collection development beyond the prototype phase.

3.2. Collection scope

Visual resource professionals, art historians, as well as scholars in other arts and humanities disciplines repeatedly stressed the value of the kind of online image resource the AIC was proposing. Indeed, they wanted much more in the way of collections than the AIC was prepared initially to offer. Demonstrations of the concordance core inevitably provoked questions from specialists who wanted to know how their particular interests (e.g., in Americana, in Southeast Asian art, classical art and archaeology, or in architecture of the European Renaissance) would be supported by the AIC. In addition, it provoked offers from individuals and institutions acting as custodians of more specialized image collections and who were willing in principle to consider contributing these to the AIC. In response, the project team returned to the idea of the concordance core as merely a starting point for collection development - a hub comprising one or even several instances of representative art historical works around which whole new collections could be developed. Maintaining its emphasis on coherent collections meeting definable educational needs, it expanded its vision again to include numerous coherent collections, sequentially developed and launched, adding scope to the AIC through addition but also depth, for example, through the cross-collection searching that would become available. A vision of phased addition of coherent collections also underpins the marketing model and pricing strategy as represented in the business plan.

3.3. Service development

The AIC initially envisaged a single initial collection (the concordance core) supported and developed through three online services (online access to the core, a shared cataloguing tool, and an image exchange that allowed individuals to upload digital image content directly into the online corpus as well as to download images for local use). Demonstration and review revealed considerable skepticism about the image exchange, while shared cataloguing efforts were postponed until work on the prototype collection and image delivery service was completed (see Section 4.4. below). Just as some service development aspirations were tabled or put on hold, other new ones were introduced. Indeed, public review and evaluation revealed demand for a far richer range of services that, if included in some phased development path, would diversify the AIC's revenue streams while enhancing its collections. Services are described in detail in the business plan where the shared cataloguing tool is re-introduced as a key feature of the AIC service at launch. Additional services may be included in some phased development.

3.4. Collection strategy

The AIC was launched with the view that a collection could be developed on the basis of images contributed by individuals to fill needs identified on widely publicized want lists that were themselves constructed by subject experts and other domain specialists. Where the concordance core was concerned, the AIC was fortunate enough to find a significant number of the requisite 2,300 images in personal collections. Even with this initial jumpstart, this mode of collection development proved to be slow and cumbersome. There are, for example, uncertainties that surround the copyright status of images contributed by individuals, some of whom may not be abreast of copyright law and its implications. At the same time, as the AIC began to broaden its collection scope beyond the concordance core, it began to identify wholly coherent collections managed by single institutions, and potentially available for inclusion. While maintaining a want list function, therefore, the AIC's principal collection development strategy will rely more on expert advisory groups that define collection priorities in reference to data on research and teaching needs and market potential, and help to select images to assemble priority collections. This more managed and strategic approach to collection development does not necessarily assume a centralized structure for receipt and processing of digital images. It does necessitate in-house technical standards to guide image production and guidelines for those interested in contributing to the AIC. These are supplied in Appendix E.

3.5. Distribution model

Reflecting, in part, the catholic membership of the College Art Association, the AIC initially envisaged a distribution model that focused entirely on web-based services accessible to individuals and institutions with very different levels of computing infrastructure and support. Early on it realized the importance of supporting the growing number of institutions that have interest in and capacity for locally managed image services. Consequently, the AIC has adopted a stratified approach. It will make its collections available online and by distribution (with periodic updates) on portable media. The functional specifications for both distribution paths is supplied in Appendix F.

3.6. Service platform

The AIC set out to develop application software that would supply the full range of web-based access services envisaged for the prototype using open standards (XML) in order to reduce the AIC's need to create end-user tools while enhancing the end user's ability to use local tools to deliver AIC content. The software was developed by Carnegie Mellon University under contract to the AIC. It supplied a large subset of the AIC's specified access features via a suite of Open Source XML files with a configurable API that allows underlying image and data content to be stored in and retrieved from virtually any web-accessible database software.

The software served the AIC's purposes through its public demonstrations (a feature description is supplied in Appendix C) but was rejected by the AIC in favor of a pre-packaged solution when the AIC turned its attention to planning its business future. An evolving market and cost considerations combined to influence this decision. With regard to the market, several solutions suited to the AIC are now available and under consideration. With suitable products now available on the open market, the AIC Executive Committee felt that it could no longer justify the costs involved in developing its own software solution, and the associated long-term costs of maintaining that solution's viability on successive generations of hardware platforms and operating systems.

3.7. Organizational and operational considerations

During its planning phase, the AIC has relied on a combination of volunteer and paid effort at the University of Michigan, Yale University, Carnegie Mellon University, and the Digital Library Federation. Significant contributions and leadership have also been supplied by staff at the University of Oregon, and the College Art Association. Although appropriate to a planning process, this organizational approach is unsuited to the fully operational phase envisaged in the business plan. Accordingly, the AIC seeks development as a 501(c)3 organization responsible to an advisory board with at least 5 members of full-time co-located staff looking after business, collection, and technical development, and marketing and administrative functions. Image processing, data distribution, and web hosting functions may be required in addition but may be supplied by entities acting under contract to the AIC.

In an operational mode, the AIC will also require: empirical data about the use of visual resources in research and teaching, and guidelines for using them to evaluate and prioritize collection development opportunities; a technical framework to guide image processing and management; a functional specification for data dissemination services; and a rights management framework. Empirical data on research and teaching needs will help the AIC to prioritize investment in new collections and enable it to enlist support of domain specialists in their design and development. The latter function is fundamental to the emergence of an educational resource that is trusted by the community for its quality, integrity, and value. Empirical data are presently being generated, and a draft collection development framework has been prepared and is supplied in Appendix G.

A standards framework and technical guidelines have already been referenced above and are available in Appendix E. They will ensure consistency, quality, and reliability across collections sourced from numerous individuals and institutions. They will also enable the AIC to secure some operational services under license from third parties.

A rights framework will reduce the AIC's legal liability while encouraging collection development. The rights framework is one of the least developed operational components. At present it exists as a set of principles (available in Appendix G, section 13) that need will be transformed into templates for negotiable licenses.

4. Next steps

One of the DLF's core aims is to leverage its members' digital library investments and expertise to foster the development of innovative information services that educational and library institutions require but cannot afford or be motivated to supply individually or through limited consortial engagements. The AIC is the best-developed example of what the DLF hopes to achieve in its catalytic role. Other examples include its work on digital archival repositories and on Internet portal services (for further information see "The Digital Library and B2B Services").

In conformance with this programmatic objective, the DLF and the teams it has assembled around the AIC are deeply committed to the development of an information service that will promote and expand opportunities for innovation in research, learning and teaching with visual resources. They feel no proprietary hold over that service but are resolved to deploy their resources and collective expertise to its development. The next steps effectively fall into two categories: those taken to produce the outstanding agreements, licenses, and tools that the service will require; and those taken to initiate the service in an operational form. At present our focus is on the former. The following framing activities are currently envisaged or underway:

4.1. Making a "buy or build" decision about the AIC's delivery platform

That decision will require a further survey of the costs and capabilities of present software solutions. The survey will need to take account of the software's innate capabilities, potential, and costs but also of the detailed functional specification that has been developed for the AIC (Appendix F).

4.2. Turning licensing principles into template licenses

A principal aim of the AIC is to create trusted, high-quality, and legitimate online visual image resources that can promote and expand opportunities for innovation in research, teaching, and learning. Indeed, the failure so far by educational and cultural communities to confront the thorny issues that surround intellectual property and the fair use exclusion has significantly impeded attempts to harness information technology for the benefit of those interested in the scholarly use of visual resources. Intellectually coherent collections of online visual resources are being created by several universities but they tend to incorporate images whose copyright status is questionable at best. Accordingly, they are placed behind institutional firewalls and made accessible to only the smallest audience. They are also inefficient, representing great redundancy of effort while preventing colleagues at different institutions from learning from one another's efforts. More publicly accessible online collections exist to be sure. Relying on the availability of copyright cleared content, however, they tend on the whole to be patchy in their coverage and fail as a result to supply the deep resource that the scholarly community requires, whether for research or teaching purposes. Confronting these issues head on requires close involvement of legal counsel and a robust rights framework that reflects and supports a service's collection development and business aims. The AIC is working toward this framework. It has at present established licensing principles that it wishes to reflect in any licenses that it negotiates with image contributors. The principles are tied to its collection development and business aims. They need to be reviewed by expert legal counsel and drafted into licensing templates. In addition, the AIC needs to work along similar lines to develop an appropriate range of end-user licenses and licenses that may be used with any third party to which the AIC sub-contracts the provision of any of its operational services (archiving, distribution, etc.).

4.3. Establishing priorities for collection development

During the planning process, and particularly in response to public demonstrations of the AIC prototype, the AIC has received numerous offers in principle to contribute image content to its collections. Focusing its limited human resources on the prototype's development and review and then on the collection strategy, business plan and service route map, the AIC has begun to gather the data it needs to set priorities for collection development beyond the concordance core.

The Distinguished Fellows program of the Council on Library Information Resources and the DLF has recently created such an opportunity for the AIC to move in this essential direction by awarding a fellowship to Max Marmor of Yale University. Mr. Marmor has been a guiding light in the AIC's development and is also deeply engaged in a closely related initiative, Imaging America, which is based at Yale University. Mr. Marmor's fellowship will focus the AIC's collection development strategy and then, within that strategy, develop opportunities that can serve the combined purposes of Imaging America and the Academic Image Cooperative. At least five strands of activity are required, the last three of which assume the availability of a prioritized list of collection development opportunities and a robust rights framework replete with template.

Prioritizing collection development effort. The AIC's mission and its business interests are tied to the development of useful image collections that meet definable teaching and research needs. Charting the scope of its collections beyond the concordance core and assessing and prioritizing need for coherent teaching and research collections within that scope are key and immediate tasks. As indicated above, the core is only a starting point for the AIC. Including images representative of art historical work across time and genres, it supplies small sets of images around which other discrete collections may be developed, e.g., in South Asian art, in Impressionism, Romanticism, Byzantine architecture. Collection priorities, although influenced to be sure by opportunity, need to be established on the basis of careful analysis of teaching and research needs and trends. The analysis (which will also supply essential market research) will require methodological innovation. Without the benefit of "Impact Factors" and other data that have helped electronic publishers establish priorities for digitizing scholarly journals, for example, the AIC will need to gather data that illuminate the use of visual resources in teaching and research. In developing its research methods, the AIC will also review the practices employed by competing electronic publishers, focusing in particular on those that seek to develop coherent collections as opposed to large assemblies of image content.
The AIC is uniquely positioned to conduct this analysis. Its long-standing affiliation to the College Art Association will be beneficial - the Association is abreast of such trends and gathers or has access to some relevant data. It can also act independently of institutional, professional, and other special interests that can influence collection priorities on principles that are less empirical and market oriented than those the AIC intends to apply. The College Art Association represents the visual resources community in its broadest form. The DLF is a membership organization but one whose members seek collectively to act as catalysts in developing, rather than as proprietors of, innovative information services. Finally, as a CLIR/DLF Distinguished Fellow, Mr. Marmor will be positioned to act independently of any single institutional interest.
Negotiation with potential contributors for access to rights-cleared images of art historical objects. The research that will inform the AIC's collection decisions will take place over 4 to 6 months. During that time, the AIC will conduct more detailed examination of the collections already on offer while exploring opportunities with the custodians of other substantial collections.
Developing guidelines for expert editorial or advisory bodies. The AIC's collection strategy and business plan both emphasize the importance of having domain specialists and appropriate professional associations involved in developing discrete collections. Experts' work will vary with the circumstances that surround any single collection. Typically, it would involve the development of a want list that includes the art historical works (or categories of works) that need to be represented in a collection to support defined teaching or research aims. Thereafter, experts may find themselves filling the want list through careful review of images on offer from a single institution (whether a slide library, museum, or private or institutional photo archive) or through some more general solicitation to a particular custodial community. Since expert groups will need to surround every collection that the AIC develops, guidelines will supply an essential tool for ensuring that the same intellectual rigor is consistency applied.
Identifying costs and determining distribution of effort in image processing to AIC standards. The work can only take place after collections are defined and in light of the opportunities that present themselves for their development.

4.4. Developing and evaluating a prototype shared cataloguing tool

From its conception, the AIC envisaged a database of records richly describing art historical works. The database was intended to serve several communities. For scholars and students it would act as a reference tool: an invaluable aid to research, teaching, and learning. For custodians of visual resource collections, the database would inform local cataloguing activities while minimizing the extensive redundant effort that they involve. The database could also have a number of business applications, for example, for those concerned in the private market for art historical objects. Focusing its efforts on the concordance core, strategies and methods for developing image collection, and on business planning, the AIC postponed consideration of the cataloguing tool for a second stage of its development. Having reached that stage, the AIC has drafted a statement of the problem a shared cataloguing tool will solve and frames a prototype development initiative. The statement is supplied in Appendix H.

return to top >>

Last updated: