Long before the invention of computers museums exchanged information with each other and their researchers. The introduction of computers has made the practice both easier (because it is faster and potentially more usable) and harder. The prospect of computers supporting exchange of information among museums and construction of shared catalogues motivated the formation of the Museum Computer Network in the U.S. in 1967 and continues to be the foundation for many collaborative projects today. But exchange of information among institutions is not the only reason for interchange standards.
Museums will increasingly acquire their holdings with machine readable information already associated with them whether from site surveys or auction records. They need to incorporate this information and other data from biographical and historical files into their local information systems.
Museums must also be concerned with whether computerized documentation they painstakingly build up around their collections will be accessible and meaningful to anyone else long after the original collectors of it - and the systems they worked with - have vanished. Therefore, they need to be able to move the data from one vendor's system to another, because of system obsolescence and because multiple applications within the museum will require access to the same information.
Museums also want to deliver information from a variety of sources and formats including print, oral history, still and moving pictures in their interpretation of holdings and to document the contexts of an objects's creation and use. They want to publish their interpretations in exhibitions and reference works, in print and multimedia, and to distribute information electronically.
Each of these needs is dependent on information interchange: information interchange to capture data when collections are acquired; information interchange to move documentation from one system to a replacement and to integrate multiple applications within the museum; information interchange to merge multimedia into unified databases; and information interchange to publish and disseminate museum interpretations.
Making information interchange work smoothly and transparently is difficult because: different makes of computers and different software products, are not designed to "talk to each other", and different professions within the museum have their own intellectual perspectives, and different processes within the museum require information in different forms. As a consequence, the structure and content of information in different computer applications varies considerably. Making everybody use the same system and do everything the same way is not a solution, because we value the differences in perspectives, because the differences in uses for information reflect the actual requirements of day-to-day administration, and because competition between hardware and software vendors helps stimulate additional functionality and keeps prices falling.
The metaphor of human communication is a way of explaining computer to computer communication. Humans use the formal properties of language, grammar, syntax and vocabulary to provide the basis for communication and exploit the physics of transmission of sound waves to carry the message, and the technical apparatus of human vocal chords, ears and brain to decipher it. Generally we are not confused by the sounds we hear, if we speak the language, because we understand the "standards" for successful exchange from our earliest learning experiences.
The case would be the same for computer interchanges if we had the same degree of familiarity and standardization. But, we need to become familiar with the new equipment, learn new rules, create the formats and syntax and agree on the meanings and vocabulary.
During the past decade the international museum community has made numerous efforts to standardize aspects of its practices including description of museum objects and methods for their accessioning, storage, loan, and exhibition (Bearman,1991a). Museums have also adopted standards for ethics, for collections plans and for financial accounting. Much effort has been expended internationally on standardizing the content of museum information within particular shared systems, especially by the Canadian Heritage Information Network (CHIN) and the Museum Documentation Association (MDA). However, museums have not yet collectively defined an information management standards framework within which to view the content, structure and terminology standardization efforts in context.
The CIMI initiative was designed to advance the use of computers in museums by defining a framework for interchange which, if followed, will reduce costs and risks to individual institutions. Eventually, these standards will increase the number of opportunities which museums will have to participate in effective information interchange as a result of acquiring hardware and software with built in capabilities for supporting the needs of museums.
The CIMI Standards Framework presented in this report specifies the interchange standards that should be used by different museum applications to transfer data so that it will be usable in receiving systems independent of their hardware, software or network vendor. It rests on some simple premises:
In 1986 the International Council on Museums Committee on Documentation (CIDOC) endorsed the standard ISO 2709 as the basis for all museum information interchange. Their action went virtually unnoticed and it was evident that it would not have a practical impact on museums unless national organizations took up the challenge of implementing software and networks based on the standard. The CIMI initiative was proposed to the Board of MCN in 1988 as a means of taking the CIDOC action forward in the United States. It fit neatly into MCN's newly adopted mission of increasing the role of computers in museum management and collection documentation through promotion of standards.
In endorsing the CIMI initiative, the Board of MCN acknowledged that the absence of standards significantly inhibited the use of automation by museums and the creation of museum oriented products by vendors. In 1989 MCN issued invitations to U.S. museum associations to send representatives to serve on a body which became known as the CIMI Committee. The invitation was accepted by the American Association of Museums (AAM), the American Association for State and Local History (AASLH), the Association of Living Historical Farms and Agricultural Museums (ALHFAM) and the Association of Systematics Collections (ASC) which together represented most museums in the country. The MCN itself sent a representative, and extended invitations, which were also accepted, to the providers of networked museum services in North America: CHIN, the Conservation Information Network (CIN) and the Research Libraries Group Inc. Funding for the meetings of the CIMI Committee was secured from the National Endowment for Humanities and the Pew Charitable Trusts, and in-kind support was provided by the MCN for a period of two years. Before the Committee met additional invitations to participate were accepted by the Argus Users Group, the U.K. Museum Documentation Standard project, the U.S. National Park Service, and Willoughby Associates. The chairman of CIDOC also participated throughout. In addition, the CIMI Committee meetings were open to observers and several observers played quite active roles and came to most meetings.
Prior to the first CIMI Committee meeting in the fall of 1990, it was anticipated that the effort would lead directly to adoption of a format for interchange based on the MARC format, known to the standards community as ISO 2709 because this standard had been endorsed by CIDOC. In its applications for CIMI funding however, MCN noted that there were problems in using ISO 2709 (which had been designed for text only), and suggested that either the ISO 2709 would need to be modified or that an alternative format would need to be devised. The grant proposal envisioned that in the end there might not be a single data interchange format for museums, but it authors (Bearman and MCN Executive Director Deirdre Stam) assumed that an institution might select one format from among several approved by CIMI. Neither the proposal nor any of the early discussions about the interchange standards prepared us for adopting a suite of existing standards. In retrospect it is clear that the deliberations of the CIMI Committee radically changed, and valuably modified, not only the CIDOC recommendation but the original assumptions of MCN sponsors of CIMI.
When the CIMI Committee first met in October 1990, it recognized that the concept of what museum data, and interchange services, had evolved considerably since 1986. Not only was it evident that digital image bases, sound bases, and compound digital multi-sensory and multi-media documents were soon going to play a major role in collections documentation, but also that services other than collections information database construction, such as exhibits development and loan, photo fulfilment, and conservation assessment would be a focus for data interchange in future museum networks. Therefore the first CIMI meeting focused on establishing a process for involvement of the broadest representation from the museum community to define requirements for museum information interchange because it assumed, correctly, that such a process would bring new requirements to light. By the end of its first meeting, the Committee had laid a foundation for a statement of mission and guiding principles for conducting its work shown in Figure 1.
+----------------------------------------------------------------------------+ | Mission Statement | | | | The mission of the Computer Interchange of Museum Information | | committee is to identify a technical framework to support the computer | | interchange of all museum information deemed relevant by the museum | | community. | | | | Goals | | | | Develop a technical framework | | | | Develop a technical framework, known as the CIMI Standards | | Framework by combining data interchange formats, interchange | | transport protocols, transfer media options, and other requirements to | | coherently support the interchange of museum data. | | | | Assist in the development of museum data interchange formats | | | | Work with task groups from the professional community to map specific | | requirements into the CIMI Framework by defining standard | | interchange formats that are tailored to the needs of particular | | interchanges. | | | | Create support and acceptance | | | | Create the conditions to encourage widespread acceptance of the CIMI | | Framework by assuring their effectiveness, by communicating | | extensively with a broad professional community, by working with | | organizations representing the information needs of the professional | | community, and by implementing a legitimate decision making process. | | | | Accommodate evolution and maintenance | | | | Work towards a mechanism to support the maintenance and continued | | development of the CIMI Protocols in response to the ongoing | | elaboration of community interchange needs and the evolving technical | | environment for providing interchange. | | | | Guiding Principles | | | | Conduct the business of the committee in an open, consultative manner. | | Fundamental to this is the participation of standards consumers, | | producers and others with relevant interests; decision making by | | consensus; the opportunity for broad public review of work; and the | | adoption of other relevant national and international standards. | | | | Work with Task Groups constituted from the professional museum | | community to ensure that the CIMI Framework neither dictates the | | content or purpose of interchange, but assures the accommodation of the | | information needs desired and expressed by the profession. | +----------------------------------------------------------------------------+ Figure 1: CIMI Mission Statement
The Committee also asked Project Manager John Perkins to prepare two briefing papers in advance of its next meeting: one on the standards development process, and a second on museum requirements for information interchange.
The briefing papers prepared for the second meeting held in Spring 1991 introduced the range of standards bodies and standards (discussed in Chapter II of this volume) and began to reveal the extent of museum needs for interchanging information. They, and summaries of all other subsequent briefing papers, were sent to a growing mailing list of museum professionals interested in CIMI. By June 1992 this list included more than 600 individuals and organizations.
At the meeting, the Committee was briefed on the concept of standards frameworks, particularly the OSI Reference Model (discussed in Chapter II and shown in Figure 4) and the OSE Model (Figure 5) as ways of making sense of the plethora of standards. The committee decided, as a consequence, to advise museums, to look to external standards such as the level 1-5 OSI standards or equivalents [2]. The focus of CIMI Committee work, therefore, became standards relating to the Application layer, or level 7, and possibly on some level 6 standards relation to the Presentation layer of the OSI Reference Model.
The Committee held a brainstorming session to identify several dozen museum applications. It began to discuss a number of standards about which individual members were knowledgeable and which were thought to be germane to specific applications. At this point, in the absence of more complete information about either the standards or the applications, the Committee again turned to the Project Manager and requested a briefing paper for the fall 1991 Committee meeting in which specific museum applications would be correlated with specific standards and recommendations would be presented. Before adjourning, the Committee drafted the definition of its requirements for a CIMI standards framework discussed in Chapter 3.
In the process of writing the paper on "Options for the Computer Interchange of Museum information" (Perkins 1992), it became clear that clusters of museum applications shared requirements for information interchange which were satisfied in turn by clusters of standards. In some cases only a single standard method of satisfying a requirement was available; in other cases several standards seemed potentially applicable.
At its meeting in the Fall of 1991 the Committee decided to locate additional standards to add to the OSI level 1-5 foundation of the evolving CIMI Standards Framework. For example, it decided to adopt Electronic Data Interchange (EDI) for those business functions of museums that are identical to business functions of other organizations, such as acquisition of materials and supplies and sales of museum products (including ordering, billing, receiving, and paying), and for those functions which are analogous to business functions of other organizations but peculiar to museums such as exhibitions loan (including steps like insuring, custom brokering, and shipping).
The Committee also decided to adopt standards for electronic mail which conform to existing international messaging standards and for information retrieval corresponding to ways in which data is retrieved in other environments. The Committee further decided that the interchange of data files could be handled by existing ISO file transfer, access, and management standards (FTAM).
The Committee was not able, on the basis of the information it had, to decide between three standard methods of interchanging collections-related data records. After defining a set of criteria by which it would evaluate the adequacy of each of the three different approaches, the Committee again asked its Project Manager to illustrate the way in which the data might be handled by each standard and describe the benefits and drawbacks of each.
The National Gallery of Canada and the Canadian Heritage Information Network provided sample museum records which were then mocked up in each of the three standards (ISO 2709, SGML, and ASN.1) to show which would best serve as a framework for interchange of museum object identification, description and analysis information. How well each standard met the criteria set by the Committee was then researched and the results presented to the Committee for its next meeting.
At its final meeting in the Spring of 1992, the CIMI Committee assessed the sample transactions displayed in each of the three possible standards to determine: 1) How much of the presumed data required by museums could each convey? 2) How much further specification would each require in order to define an implementation for museums? 3) How much overhead would each incur in a theoretical model of museum interchange? 4) How intelligible would each be to museum staff? 5) What software and network services were already available in the market to support each method? 6) How difficult it would be for application vendors in the museum software marketplace to develop input/output routines which format data according to the standard? And secondly, whether computer staff of institutions using in-house systems would be able to write such routines. 7) How complex it would be for the community to effect the standard? Whether existing implementations could be readily changed without significant museum community investment? 8) What other benefits or drawbacks were associated with the use of each protocol? Whether the protocol(s) would assist museums to achieve their missions after data was interchanged?
The Committee did not initially come to a unanimous decision. All but two members felt that SGML provided on the whole the greatest advantages for collections related data interchange. One member felt that ISO 2709, the MARC format framework, best supported some types of data, especially object surrogates or citations. One member felt that ASN.1 best supported some types of interchange, especially scientific and research data.
In the end the Committee unanimously adopted a compromise statement that it had examined all three options for museum collections and catalog information use. Each might have relevance depending on the type of information being interchanged and its intended uses. The differences between the three approaches were expressed as how well they might handle a particular type of data and how easy it might be for the museum community to use them.
It was agreed that more research was necessary and to further clarify options the CIMI Project Manager would continue working with Task Groups, once they finished defining their data requirements, to develop interchange services using any or all of the three options.
From the outset, the CIMI Initiative was designed to involve partnerships with representatives of specialized museum professions (registrars, conservators, curators) and discipline-oriented museums (art, history, science) who would be responsible for defining the specific data requirements of their community. CIMI encouraged existing groups to join the initiative and helped form new working group. These Task Groups were to define the information they needed to interchange, and the CIMI staff would develop the technical specifications for the desired service. By working together each partner would contribute necessary expertise. Computer experts would not end up dictating content and museum professionals would not be forced to make decisions about technical protocols.
The first Task Group to become affiliated with CIMI was the Art Information Task Force (AITF), a working body of the College Art Association and the Getty Art History Information Program, whose special interest was in the interchange of scholarly information about art objects. The second Task Group was formed with the American Association for State and Local History (AASLH) to follow up on the AASLH Common Agenda Database Task Force and to find ways of fulfilling an interest among cultural history institutions in information interchange. A third Task Group arose from the interests of registrars of several large art museums that regularly lent exhibitions and wanted to interchange data about the loans with each other and agents in the process such as shippers, brokers and insurers.
These Task Groups were encouraged to describe their interchange service requirements in their own terms, specifying the data they needed to send/receive and the functions of a systems which would serve them. As of March 1993, one Task Group has completed this process and one is very near completion. The CIMI Project Manager is assisting the Task Groups to express their requirements formally and technically in an "interchange service definition" or in the structures provided for content definition in an existing protocol whose services are already defined. Once this formal definition is complete, the interchange service can be implemented by museum software vendors and network service providers.
Go Back
Go to Index