Stories

D-Lib Magazine
March 1999

Volume 5 Issue 3
ISSN 1082-9873

The Getty Information Institute

A Retrospective

blue line

Eleanor E. Fink
eefink@earthlink.net

 

Publisher's note: Since the early 1980s, the J. Paul Getty Trust has been a pioneer in applying computing to art history information. Recently these activities have been carried out by the Getty Information Institute, directed by Eleanor Fink. Following a change of leadership, the trust has decided to close the institute in June this year. To commemorate the end of its very successful initiatives and programs, D-Lib Magazine invited Eleanor Fink to write the following retrospective survey of the institute and its achievements.

Introduction

"If you donít think that L.A. has any culture, itís time to get connected. Faces of L.A. is an electronic gateway to the vast collections of over 20 artistic and educational institutions in Los Angeles, available 24 hours a day through the Internet. This Ďdigital libraryí is a huge resource for creativity and research, inviting you to explore over 2.5 million books, articles and works of artÖ"    Öexcerpt from Faces of L.A. Brochure1

For almost two decades, the projects and activities of the Getty Information Institute (GII) have been devoted to strengthening the presence, quality, and accessibility of art and cultural information via computer technology. More recently, in this era of networking technology, GIIís focus has been to promote the concept of digital libraries and the collaborative development of standards, tools, and guidelines needed to build and link digital art resources for ease of access globally. With the rise of the Internet as a universal medium of communication, cultural heritage information in the form of texts, sound, and images are being digitized and placed on the Web at an accelerated pace. Yet the current effect of universal access is like a weed patch that lacks any effective means of locating and retrieving specific items2.

To bring order to this chaotic situation, and to leverage the efforts of many geographically dispersed groups working separately, the Information Institute played a catalytic role in shaping an information infrastructure -- a coherent, synergistic infrastructure that would make access to electronic information about our cultural heritage much easier. The Information Institute believed that the cultural heritage sector should develop data structures and vocabulary tools that assist in making connections across repositories of information. Well-designed structures and vocabulary tools would allow students and researchers to search an enormous variety of the worldís electronic databases of texts and images as if they were one. GII promoted the concept of digital libraries or a virtual database and worked with many institutions to develop research databases, data structures, standards, guidelines, vocabularies, and demonstration projects3.

As part of the J. Paul Getty Trust, GII occupied a strategically advantageous position nationally and internationally4. Unlike the limitations placed on many governmental or for-profit agencies, GII represented neutral ground at the intersection of many vested sectors and was, therefore, able to address an entire spectrum of issues and policies. Through its diverse range of staff expertise, including specialists in information and library science, art history, museum studies, database design, computer networking, and multimedia, GII bridged art information and technology and was well prepared to achieve its goals. But now, as a result of a radical reorganization taking place at the J. Paul Getty Trust, the Information Institute will cease to exist as of June 1999.

Established as the Art History Information Program (AHIP) in 1983, GIIís original mandate was to explore ways of applying computer technology to improve scholarly access to and use of art-historical information5. Initially, GII began to (1) build a critical mass of electronic information essential to art-historical research: catalogs of works of art, visual reproductions, bibliographies, biographies and provenance of art objects, (2) identify model approaches and act as a catalyst among other institutions with similar interests to foster standards and procedures for sharing information on an international scale, (3) analyze the needs of the art-historical community and tailor emerging technology to the requirements of humanistic research and, (4) influence the development of a technical structure through which others could contribute and retrieve data.

Laying the Groundwork: Model Research Database Projects

It is ironic, with all the hyperbole about the information age, bandwidth, wiring schools, and portals, that the need for content -- a fundamental prerequisite -- is not receiving more attention. Works of art and their collateral information are dispersed throughout the world in many countries and institutions. Research materials exist in diverse visual and textual media, often without systematic organization.

GII contributed significantly to a critical mass of high-quality content by collecting information in more accessible electronic form that allowed manipulation for scholarly use. This was achieved by building model research databases in collaboration with partner institutions: The Avery Index to Architectural Periodicals (with Columbia University)6, the Bibliography of the History of Art (formerly RILA, with the Clark Art Institute in Williamstown, Massachusetts; see Illustration 1)7, the Census of Antique Art and Architecture Known to the Renaissance (with the Warburg Institute in London and the Bibliotheca Hertziana in Rome)8, the Witt Computer Index (with the Courtauld Institute in London)9, and the Provenance Index10. All of these efforts were collaborative and continue today with the goal of assembling information in automated form. Two other projects, the Museum Prototype Project and the Architectural Drawings Advisory Group (ADAG), later Foundation for Documents of Architecture, were consortia of institutions that produced useful experience and guidelines, but did not go beyond the prototype stage11.

These data collection projects, begun in an era of stand-alone computing, posed many of the challenges that remain fundamental to databases in the humanities. First, their data collection involved original cataloging, abstracting, and recording from primary sources. This labor-intensive and highly specialized work needed to be performed by professionals with either advanced degrees or an in-depth knowledge of the subject matter. Second, because unrecorded source materials are boundless and can absorb a limitless supply of resources, questions related to sustainability arise. Who is responsible for continued refreshing and updating of the data? How is the long-term maintenance to be funded? And what limited objectives need to be defined? These questions seem particularly relevant today with the increasing interest in forming portals and subject specific networks.

Additionally, the complexity of art-historical information pushed computer technology to its most sophisticated limits. Because this type of information has inherent idiosyncrasies that cannot necessarily be reduced to equivalencies without losing critical data, rigid categories failed to adequately describe even what may appear to be an ordinary object. For example, a "simple" architectural drawing may:

These challenges were eventually addressed by building production platforms in-house. No commercial off-the-shelf software could handle the range and complexity of relationships these resources required. Contracting production was seldom cost-effective because of the enormous learning curve involved in understanding audience, points of view, specific practices and needs.

Preserving and Connecting Digital Resources: Vocabulary Tools

One of the most fundamental lessons learned from building the model research databases was the challenge of dealing with a technology that was constantly changing. GII was quick to appreciate that the real investments in developing a database are the data and the standards used to formulate and manage the data, and to migrate data successfully to new platforms as hardware and software evolve. This investment demanded that consistent standards and terminology be developed both to insure the long-term viability of electronic information and to enhance the possibility of searching across databases.

Since research involves making connections and art information does not reside in one type of institution, be they libraries, museums, archives, or private collections, a wholistic approach to developing standards involving many levels of practice and expertise was needed. There are good reasons, for example, why different kinds of objects -- buildings, contemporary art, traditional paintings, ethnographic works, decorative art -- use different categories to record similar information. Cataloging expertise also varies, making different descriptive terminology appropriate. Creating such complex connections in large databases demands innovation. It was in this area that GIIís Art & Architecture Thesaurus and the Vocabulary Coordination Group made the greatest strides.

In the case of the Art & Architecture Thesaurus (AAT), GIIís departure from the traditional approach to vocabulary "authorities" took advantage of the new flexibility offered by technology12. The AAT proposes that there can be several acceptable terms for the same object, idea, or technique and links related terms so that the use of any one of them points to the same descriptive concept (see Illustration 2). The methodology gained from developing AAT indicates that additional descriptive languages covering various subjects could vastly reduce the cataloguing efforts of individual institutions and simultaneously ensure vocabulary consistency across disparate databases.

GIIís Vocabulary Coordination Group (VCG), founded in 1987, developed two additional vocabularies: the Union List of Artists Names (ULAN) and the Getty Thesaurus of Geographic Names (TGN)13. Like the AAT, both of these resources are based on the flexible principle of mapping together multiple name forms rather than providing just a "preferred name" (see Illustration 3 and Illustration 4). They address international requirements with respect to incorporating vernacular terminology and have consequently become de facto standards in many museums and visual resource collections internationally. This flexible approach is significant in a field such as art history where, for example, two art historians might record an artistís name differently, each version being "correct" and with its own traditions of use. The result of mapping together variant names is a robust set of vocabularies that have proven as useful in locating and retrieving information as they have in providing guidance in cataloging: A search for Domenico Theotocopuli will locate records cataloged under El Greco; a search for Constantinople will locate records cataloged as Istanbul and Byzantium.

Establishing a Voice for the Arts and Humanities: Advocacy and Education

By 1993, it had become clear to GIIís new director, Eleanor Fink, that the shift from a stand-alone computing environment to a networking and interworkability environment could, with the help of standards, provide global and simultaneous access to the worldís art and culture14. Anticipating the feasibility of erasing geographic boundaries and simplifying access across institutions, Fink determined to promote the vision of the "virtual database" -- bringing together images and contextually related information about the worldís art into one virtual space for purposes of education, research, and enjoyment15. GII began to encourage cultural institutions, at both the policy and practice levels, to look beyond the limits of their physical locations and explore the opportunities communication technologies offer to make connections across collections.

To begin, several requirements and issues needed to be addressed. Museums, libraries, archives, other cultural heritage institutions would have to see the value of a collaborative effort that could lead to a digital library or virtual database concept. Additionally, their participation in developing the data structures, standards, and guidelines that create deep linkages and ease of access across collections would be critical. Intellectual property-rights issues would also need to be demystified and solutions for educational use of digital resources would need to be better understood. Because the commercial or entertainment interests driving the rapid pace of technological change could not be counted on to make these issues a priority, there was an urgent need to ensure a voice for the cultural heritage sector in the planning and development of global networks.

In order to address these opportunities, visions, and requirements, GIIís mission centered on strengthening the presence, quality, and accessibility of art and cultural information across networks worldwide. Three strategic areas evolved:

As a prelude to a strategic focus on advocacy and education, GII and the American Council of Learned Societies (ACLS) co-sponsored a national conference in 1992 with financial assistance from the Council on Library Resources, the Coalition for Networked Information (CNI) and the Research Libraries Group16. The "Technology, Scholarship, and the Humanities" conference brought together a highly representative group of policy makers to discuss a number of issues, including the need to articulate requirements and establish standards if the arts and humanities were to be served by technology.

In 1994, with heightening concerns that scholarly information in the humanities was being left out of the emerging National Information Infrastructure, GII again joined with ACLS and CNI to publish Humanities and Arts on the Information Highways: A Profile (1994) . The Profile stimulated the interest of the White House Economic Council and a broad spectrum of organizations in the arts and humanities by making a case for the importance of providing all people with electronic access to the nationís cultural heritage. The publication also articulated the special challenges and opportunities associated with digitizing humanities and arts information. It included the findings of two working groups, one on technical requirements for arts and humanities computing, the other on electronic resources.

Two additional GII strategic initiatives grew from the interest generated by the Profile. The first was a collaborative effort with the ACLS and CNI to form a broad coalition of arts, humanities, and social science organizations. The resulting National Initiative for a Networked Cultural Heritage (NINCH) was designed to serve as an on-going voice in the planning and development of the National Information Infrastructure, the much-publicized plan for a national telecommunications system. Today, NINCH consists of nearly seventy member organizations and serves as a pivotal resource for information on issues and policies across its constituency.

The second initiative, The Research Agenda for Networked Cultural Heritage, was undertaken by GII to offer public policy makers and private foundations the information they needed to direct support for arts and humanities computing. The project was designed to (1) achieve consensus among technology and information experts on the research needs for arts and humanities computing and (2) to articulate and publish a research agenda. Eight papers were commissioned to outline research needs for specific domains. These papers were made available on the Internet and electronic discussions ensued. The final report, containing the papers and a summary of these exchanges, was published in 199617.

In addition to ensuring a voice for the arts and humanities in planning global networks and addressing research needs to help policy makers and foundations direct support for arts and humanities computing, other challenges commanded GIIís strategic planning: inadequate understanding of, and the lack of agreements related to, issues surrounding Intellectual Property Rights (IPR) were major barriers to universal access to images and art information. In response, GII launched the Museum Educational Site Licensing project (MESL), a three year initiative aimed at demystifying IPR by exploring and identifying educational uses of digital images and recommending models for site licensing. MESL participants included seven museums and seven universities18. The museums provided a testbed of digital images to the universities that was made available over campus networks. Two publications summarize the findings of MESL: Images Online: Perspectives on the Museum Educational Site Licensing Project and Delivering Digital Images: Cultural Heritage Resources for Education. As a pioneering initiative designed to increase understanding of the mechanisms needed to support digital libraries or virtual databases, MESL inspired two independent business models: the Art Museum Image Consortium (AMICO) and the Museum Digital Licensing Collective (MDLC).

Erasing the Boundaries of Time and Place: Global Standards and Tools, and Demonstration Projects

With the objective of promoting the concept of simultaneous access to the worldís art and culture, GII intensified its efforts in developing global standards and tools, and began a series of state-of-the-art demonstration projects19. As the glue that enables building and interconnecting digital art resources and providing ease of access globally, standards are essential. Because their development requires the participation of the community or communities using them, exploration of productive models for cooperation and consensus building are critical.

To leverage the investment needed to advance interchange standards for art and culture and to stimulate international cooperation, GII, together with The Canadian Heritage Information Network and the Research Libraries Group, became a sponsor of the Consortium for the Computer Interchange of Museum Information (CIMI)20. CIMIís purpose is to develop and test applications for interchange formats such as Z39.50 and to promote metadata standards such as the Dublin Core. As a membership organization that includes several European institutions, CIMI is an ideal incubator for long term standards development and testing.

While helping to sponsor and direct the work of CIMI, GII produced two key data structures utilizing a consensus building approach. The first was the Categories for the Description of Works of Art (CDWA), the product of a project called the Art Information Task Force (AITF). AITF was formed by GII and the College Art Association of America. It united an interest in creating a data structure for art objects that was shared by the Museum Computer Network, the Visual Resources Association, and the Art Libraries Society of North America. AITFís goal was to define the categories of information about works of art that scholars use and would want to access electronically. With a grant from the National Endowment for the Humanities, the project brought together academic scholars and museum curators, art information professionals, and visual resource specialists to address a number of questions: What categories of information are useful for research? What is the purpose of the information? How is it used? How do scholars want to be able to access it21?

The outcome, Categories for the Description of Works of Art (CDWA), provides a standard for documenting art objects and reproductions that serves as a structure for distributing and exchanging information via such channels as the internet22. CDWA has been widely adapted, particularly in Eastern Europe and Latin America where computerization of cultural information is just beginning. It also serves as the framework for CIMIís demonstation project Cultural Heritage Information Online (CHIO).

The development of the second standard, Object ID, focused on a data structure for uniquely identifying cultural objects to help recover stolen art and help stem the illicit traffic of art objects. If CDWA represented a maximum number of categories of information required by the scholarly world, Object ID was a distillation of the minimum information needed for identification of an object. Its purpose is to provide a core record, nesting within the diverse databases of many communities that will enable broad and swift distribution across networks. In addition to joining the interests of the Council of Europe, UNESCO, ICOM, and USIA, over 1,000 organizations in eighty-four countries contributed to the development of Object ID. The collaboration included museums, heritage centers, law enforcement and customs agencies, the art trade, appraisers, and the insurance industry23.

Both CDWA and Object ID originated out of a desire to join disparate efforts that were being undertaken separately by a variety of professional groups. By convening constituents who were unaware of each otherís common goals, GII was the catalyst for both projects. The Instituteís methodology centered on surveying common practice among a broad range of organizations, then identifying areas of agreement and hosting focus group meetings to inform, discuss and systematically build consensus. The value of these structures and efforts such as the Dublin Core are summarized in GII's publication Introduction to Metadata: Pathways to Digital Information24.

To address global networking, GII also produced guidelines in several languages on how to build and use vocabularies and thesauri for the documentation of art and culture. It held workshops in conjunction with programs of partner institutions such as the Council of Europe, the Documentation Committee of the International Council of Museums, UNESCO, and others. Both the Union List of Artist Names and the Getty Thesaurus of Geographic Names incorporated multiple language equivalents, paving the way for common access across global digital resources. The Art & Architecture Thesaurus, in addition to incorporating U.S. and British equivalents, has been adopted as a tool by several countries where translation into the national language of the country has been taking shape with guidance from GII.

GIIís demonstration projects explored how universal access might take form and provided a better understanding of issues and requirements involved in that process. A.k.a., one of the first projects, was based on a Web interface, combined with GIIís vocabularies. This project demonstrated simultaneous access to 1.5 million Getty records consisting of texts and images that resided geographically in different Getty institutions and on different computers. In performing a search, information was gathered into one space for browsing. A click on a citation in the Research Instituteís Library or in BHA brought up the full catalog record. The addition of GII vocabularies made it possible to find and match disparate terms for a concept, name of an artist, or geographic place name among the various databases.

Eventually, the a.k.a concept was applied to 2.4 million records, drawing on the holdings of twenty partner institutions in the greater Los Angeles area. The project, called Faces of L.A., is publicly available on the Web. It provides simultaneous access across these many collections and brings all of the digitized resources together into one virtual space for purposes of research, online exhibitions, and curriculum development for schools. One of the interesting findings of a.k.a, as demonstrated in Faces of L.A., is that GIIís vocabularies help in searching even when applied to disparate collections that have not used them as descriptors. Because the vocabularies map together so many variations, they improve access.

The Faces of L.A. project has served as a model for other initiatives: California Culture Net, sponsored by the California Arts Council, and the New Mexico Culture Net. In Washington, D.C., GII formed a partnership with the Institute for Library and Museum Services and the American Association of Museums to create American Strategy, a cooperative of twelve agencies that seeks to make the cultural content of federal collections more widely available for research and education. GIIís technical development team has produced a demonstration of the concept which illustrates the ability to search across several federal collections.

In 1998, GIIís demonstration projects continued to explore and pave the way for common access across global digital resources. A partnership with NEC USA resulted in a visual Web tool called ARThur (see Illustration 5). By harvesting information on art objects and related information available on the World Wide Web, ARThur combines the ability to search by image with the GII vocabularies25.

All of these initiatives serve to illustrate the value of making access to art and culture, and exploring connections, as easy as possible. In many respects, the current interest in forming portals, or vertical markets that aggregate information across like-minded sectors, echoes the path that GII pursued. Last year, in conjunction with members from the California Young Presidentís Organization (YPO), GII began to explore a business model for a portal for art and culture. A demonstration was developed.

GIIís intention was to act as a catalyst for forming a portal in partnership with other cultural organizations, consortia, and the private sector. The portal would have offered online access to high quality digital resources for learning and enjoyment. Revenues from private sector investment, via advertising or subscription, were to subsidize the participation of smaller institutions with limited ability to attract support on their own. This kind of model would also have explored interworkability in the networking environment through a cooperative, distributed approach to building and maintaining the types of vocabulary tools a portal for art and culture would need to ensure access across different types of collections.

The dissolution of GII is the result of a change of leadership at the Getty Trust. The research databases and vocabularies have been transferred to the Getty Research Institute where their future will be evaluated. Demonstration projects such as a.k.a and Faces of L.A. that served as models and pointed to what is possible if members of the cultural sector cooperate will be replaced by an expanded grant program. As the staff of GII leave to seek new challenges within the cultural community, they take with them a determination and considerable talent to serve the arts and cultural information community. At a moment when the sociology of networked collaboration is just beginning, the loss of an organization like GII, positioned at the neutral intersection of many vested interests, is yet to be calculated.

Acknowledgements

Special thanks to Laurie Levin and Marilyn Schmitt for their editorial review, to Patricia Harpring for the illustrations, and to the GII staff for their outstanding accomplishments.

Illustrations

Illustration 1: The Bibliography of the History of Art

[Bibliography of the History of Art]

The Bibliography of the History of Art (BHA) indexes and abstracts all writing on Western art from 1973 to the present, regardless of the country or language of its publication. Produced in English and French, the BHA is a collaborative publication of the Getty Information Institute and the Centre National de la Recherche Scientifique. It currently contains 125,000 bibliographic records, to which approximately 24,000 records are added annually. It offers researchers access to more than 2,000 periodicals published in 45 languages, as well as books, exhibition catalogues, conference proceedings, and bibliographies.

 

Illustration 2: The Art & Architecture Thesaurus

[The Art & Architecture Thesaurus]

The Art & Architecture Thesaurus (AAT), consisting of nearly 120,000 terms, is the first comprehensive thesaurus for art, architecture, and material culture from antiquity to the present. Although its focus is currently on the Western world, the scope of the AAT extends to other cultures. The AAT is not organized by discipline, but in facets and hierarchies independent of any one subject area. The focus of a record in the AAT is the concept. A single concept is represented by a cluster of terms, one of which is established as the preferred term, or descriptor. Each descriptor in the AAT is graphically displayed to show its relationship to broader and narrower terms.

 

Illustration 3: The Getty Thesaurus of Geographic Names

[The Getty Thesaurus of Geographic Names]

The Getty Thesaurus of Geographic Names (TGN) is a structured vocabulary of over 900,000 place-names for documenting and retrieving cultural heritage information. It is organized in hierarchies representing all nations of the modern world, including vernacular and historical names, coordinates, place types and other relevant information. The TGN can be used as an authority file in the documentation (cataloging, indexing, and description) of cultural heritage information. Because of its wealth of variant names, the TGN can also be an effective retrieval tool for searching across a variety of databases that may use different name forms for the same place.

 

Illustration 4: Union List of Artist Names

[The Union List of Artist Names]

The Union List of Artist Names (ULAN) is a database of over 200,000 names for artists and architects, which also includes bibliographic and biographic data. The ULAN was originally produced from the merged authority files of seven Getty projects and two other Getty programs. Its coverage spans ancient times to the present, with emphasis on post-medieval artists and architects from Western Europe and North America. Variant names for one individual that would be separated alphabetically are clustered, allowing the user to find all records for an individual regardless of which name is catalogued or requested. The ULAN does not advocate a single "preferred" form, as in library usage, but presents a range of choices that reflect art-historical usage.

 

Illustration 5: ARThur

[ARThur]

ARThur, a tool developed in partnership with NEC, is designed to demonstrate image-retrieval technologies over the Web. Based on NEC's Amore image matching engine, ARThur can find images on the Web based directly on visual attributes, such as shape, color and texture. Visual retrieval, combined with enhanced text searching tools, allow users of ARThur to find visual materials collected from over 1,000 Web museum, library, and archive sites.

 

Referenced Web Sites

[1] A.k.a. < http://www.gii.getty.edu/vocabulary/aka.html >

[2] American Strategy < http://www.americanstrategy.org >

[3] ARThur < http://www.gii.getty.edu/arthur/ >

[4] Art & Architecture Thesaurus (AAT) < http://www.gii.getty.edu/vocabulary/aat.html >

[5] Avery Index to Architectural Periodicals < http://www.rlg.org/cit-ave.html >

[6] Bibliography of the History of Art (BHA) < http://www.rlg.org/cit-bha.html > and < http://www.gii.getty.edu/bha/index.html >

[7] California Culture Net < http://www.californiaculture.net >

[8] Categories for the Description of Works of Art (CDWA) < http://www.gii.getty.edu/cdwa/ >

[9] Census of Antique Art and Architecture Known to the Renaissance < http://www.gii.getty.edu/index/census.html >

[10] Consortium for the Computer Interchange of Museum Information (CIMI) < http://www.cimi.org >

[11] Faces of L.A. < http://www.facesla.org >

[12] A Guide to the Description of Architectural Drawings < http://www.gii.getty.edu/index/fda.html >

[13] Getty Thesaurus of Geographic Names (TGN) < http://www.gii.getty.edu/vocabulary/tgn.html >

[14] Humanities and Arts of the Information Highways: A Profile (1994) < http://www.gii.getty.edu/haif/index.html >

[15] L.A. Culture Net < http://www.lacn.org/ >

[16] Museum Educational Site Licensing Project (MESL) < http://www.gii.getty.edu/mesl/ >

[17] National Initiative for a Networked Cultural heritage (NINCH) < http://www.ninch.org >

[18] New Mexico Culture Net < http://www.nmcn.org/ >

[19] Object ID < http://www.gii.getty.edu/pco/objectid/index.html >

[20] Provenance Index < http://www.gii.getty.edu/provenance/ >

[21] The Research Agenda for Networked Cultural Heritage < http://www.gii.getty.edu/ranch/ >

[22] Union List of Artist Names (ULAN) < http://www.gii.getty.edu/vocabulary/ulan.html >

Copyright © 1999 Eleanor Fink

URL for the Research Agenda for Networked Cultural Heritage corrected at the request of the author, The Editor, March 16, 1999, 10:12 AM. The name Art and Architecture Thesaurus was corrected to Art & Architecture Thesaurus; the name Union List of Artists Names was corrected to Union List of Artist Names; and an addition was made to the Acknowledgments statement at the request of the author, The Editor, March 19, 1999, 8:20 AM.

Top | Contents
Search | Author Index | Title Index | Monthly Issues
Letters | Next story
Home| E-mail the Editor

D-Lib Magazine Access Terms and Conditions

DOI: 10.1045/march99-fink