Historically, automation for exhibition and for collection management have developed along very distinct and independent paths, with different sets of vendors, software, tools, and platforms as well as no possibility of integration. The Worldwide Web offers the possibility of linking collection management information to interactive exhibitions, overlaying the narrative structure of exhibitions onto the item-based rich collection of information found in collection management systems. This paper describes the concepts behind such an integrated system, explains what activities will be needed to get there, and reviews some existing Web activities that begin to approach such a future integration.
Historically, most museum automation efforts have been driven by the need for record-keeping and inventory control, and have resulted in collection management systems. A smaller set of automation efforts has centered in museum education departments, and has focused on interactive exhibits. The vendors, software, tools, and platforms for computers used in collection management have been very different from those in interactive exhibitions. (For a summary of the differing characteristics between the two, see figure #1.) This split was due to technological limitations, but recent technological changes are likely to increasingly cause a convergence between these two camps.
Collection management systems have historically handled very complex information but have had poor user interfaces and been extremely difficult to learn to use. User interfaces for these systems were very slow to even incorporate underlining or font changes, and not until the mid-1990s did these systems make widespread use of a mouse and a graphic user interface. Multimedia exhibitions, on the other hand, have used graphic, point-and-click interfaces since the mid-1980s. Because there is no chance to give even a short "how-to-use" lesson to a museum visitor, these user interfaces had to be intuitive and easy to use.
Collection management systems are object-based, and most data within them refers to a particular object. Though they can reflect complex relationships between objects, this is usually limited to situating an object within a set (a group or a collection). The relationship between an object and other objects, people, or theories, (usually a key element of an exhibition catalog) is seldom reflected in a collection management system.
Multimedia exhibition packages, on the other hand, primarily offer a narrative structure; they essentially tell a story and situate the objects in some relationship with other objects, people, and theories. They also offer a coherent view of some domain, unlike collection management systems which focus on individual objects. In this sense they are much closer to an exhibit or exhibition catalog than to a collection management system.
|Interactive Multimedia Exhibition Packages||Collection Management Systems|
|Designed for explanation and access||Designed for inventory control, record-keeping|
|Good user interface||Poor user interface|
|Narrative-based; Offer coherent view of some domain||Object-based; No overall view of domain|
|Records are limited and carefully selected||Unlimited records and all are available|
|Frequently no database||Usually a powerful database|
|Limited user navigation||Unlimited user navigation|
|Always single-user||Single-user or multi-user|
|Closed-box system||Open or interoperable systems|
Collection management systems attempt to create and store records for every object in the museum. Interactive exhibits, on the other hand, usually store only a selected group of records and images -- just enough to tell a story. And while the records in the exhibition system may be derived from records in the collection management system, they are often enhanced and have always been stored in a separate location.
Collection management systems usually incorporate a powerful database, and give users the opportunity to pose just about any possible query to that database. Interactive exhibit systems seldom contain a database. Though on the surface these exhibit systems appear to give users a variety of choices, usually the choice consists of multiple-choice branching, and the user can only follow branches anticipated by the system designers.
Collection management systems have tended to be relatively open; for the most part vendors have recognized the need to import records into these systems and export them to other systems, and in recent years have provided tools so that even naive users can perform these functions. Interactive exhibit systems, on the other hand, have almost always been "closed-box" systems. They see themselves as stand-alone systems that have no need to share information with other applications (except perhaps during the early stage of a project when developers might take text or images from other sources).
While collection management systems have always been dynamic (with system tools oriented towards handling data that is constantly growing and being updated), interactive exhibits have tended to have static contents. Like a publishing media, interactive exhibits have mostly captured growing and changing information and frozen it at a particular point in time, and systems have seldom provided tools for updating that information.
A number of technological limitations have led interactive exhibit systems to evolve in this particular way. Until recently, the vast storage needs for digital multimedia had led most developers to favor analog videodisc systems for exhibits incorporating images or multimedia. Because in the mid-to-late 1980s good toolkits for developing graphic user interfaces were only available on single-user systems (like the Macintosh at that point in time), most interactive exhibit developers conceptualized their projects as single-user systems. Developers chose to store their text and images in whatever format would help them achieve the highest performance, ignoring standard formats (which frequently diminished multimedia performance). They also discovered that they could simplify their work if they chose to limit the total amount of information, and control the types of questions that could be asked. And because most developers saw their work as ending with the installation of the interactive exhibit, few considered implementation of a function to export images or text.
Collection management systems were first implemented on mainframe computers, then on early PCs -- both platforms that were slow to adopt graphic user interfaces or incorporate images or multimedia. System vendors have seen their role as handling complex information about and relationships between objects (in groups), and until recently ignored technological developments bringing graphic interfaces and multimedia to their platforms.
The advent of robust graphic user interface toolkits and software for handling multimedia are making it easy for collection management systems to develop better user interfaces and incorporate forms of multimedia. As they do this, on some levels they will begin to resemble the interactive multimedia systems.
But the recent development which holds the greatest promise for the convergence between interactive exhibition systems and collection management systems is the WorldWide Web (WWW). The widespread deployment of WWW browsing tools essentially creates tens of thousands of multimedia-capable "clients". This large "market" will provide incentive for interactive developers to abandon closed-box, single-station systems in favor of producing "servers" designed to deliver to multiple clients -- something that just wasn't possible before the number of clients had reached a critical mass.
The WWW holds the promise of combining an overall narrative view of some domain with a database of individual items. A museum could create WWW documents out of narrative essays from an exhibition catalog and link these directly to object-level text and images from a collection management system which is constantly being updated. Thus, instead of the "frozen snapshot" that characterizes most catalogs and interactive exhibits, the WWW exhibit would be dynamically updated as collection management records changed. Furthermore, designers of a WWW interactive exhibit could allow users to view any non-protected text or image within the collection management system and to ask almost unlimited questions -- a radical departure from the limited questioning available within current interactive exhibits.
Most Web-based exhibits today follow the paradigm established a decade ago by interactive multimedia exhibition packages: any images or text to be used (within a Web exhibit) is copied from its normal storage spot and inserted into a Web-based narrative. But recently we have begun to see applications that are likely to eventually lead to the break-up of this paradigm.
The 2 critical features that museums and similar institutions are beginning to use are database queries and collection-level views over individual items. The ability to link a Web page with a database is a powerful tool that allows the user to change the focus of the interaction and to request more detailed information about a particular object. Though few (if any) museums have yet used their collection management information as a database linked to the Web pages of an online exhibit, a number museums and similar collections have taken the first step in that direction by putting item-level detailed object information into a database and linking that into a Web-based query (see for example the Perseus Project [www.perseus.tufts.edu], the US Holocaust Museum [www.ushmm.org], and the Jepson Herbarium [http://ucjeps.herb.berkel ey.edu/smasch.html]). Today we are at the stage of stand-alone queries to an independent database, sometimes linked to narrative exhibits. In the future we are likely to see queries integrated into narrative text (where instead of a hot-link to a pre-computed site, the hot-link is created on the fly in response to a user query). We can also expect to eventually see these queries directed towards some more permanent storage of object-level information (like a collection management system) rather than a database that is excerpted from the permanent storage area.
The second critical factor moving in the direction of this convergence is the recent movement in some domains to connect an overall view of a collection with detailed item-level descriptions. The archival community, particularly the Encoded Archival Description Project (Finding Aids 1996), has been the most active in building this type of system. Like an online museum exhibition, these projects have created online text (finding aids) that describes groups of items and their contextual relationship to one another. But these systems then link this to the full description of individual items so that the user can ask for item-level detail. Interestingly, in most of these systems the item-level information is not excerpted from another storage area, but is itself the definitive item-level description that is updated whenever new information is noted. These Finding Aids are encoded in SGML rather than a collection management or database management system, but some employ software that effectively turns SGML marked-up text into a database.
One potential new application area is to allow users to create their own narratives over the source material. An example of this can be found in the Valley of the Shadow Project at the University of Virginia (Ayers 1993-). Edward Ayers has created a narrative over a wide range of source material in the form of personal papers, government documents, images, and newspapers. But he has also opened this up to other narratives that access the same digital versions of primary source materials. Teachers and students at a number of other institutions have used the Web to create their own narratives over this same set of source material.
A number of technical impediments still need to be overcome before we see widespread adoption of this convergence. Among these are database interfaces from HTML documents which thusfar have required specialized knowledge (such as familiarity with cgi-bin). Utilities to link static Web pages to dynamic databases are beginning to become available, and will allow museums to more easily produce exhibits linked to object-level records.
Another area that has not yet been adequately explored is security: how to make available object-level records without providing access to information that the museum may not want to make public. This is a serious issue even for museums that excerpt records from their collection management systems and put them into some publicly-accessible space. Certain fields are obvious candidates for exclusion (including donor information, insurance and valuation). But other areas are more problematic. For example, at least one museum contributing records to the Museum Educational Site Licensing Project had to revisit several fields in most records where curators had added information that was "true" enough for themselves to look at, but might mislead the public which expects that online museum information carries the authoritative weight of the museum behind it.
In the future we are going to see the worlds of collection management and online exhibition increasingly converge. The WorldWide Web holds great promise for making detailed object-level records available to users from within a narrative or contextual description of a particular domain.
Portions of this paper (accompanied by other material not dealt with here) will appear in the chapter "The Transformation of the Museum and the Way it's Perceived" in Katherine Jones-Garmil (ed.), Museums and Emerging Technologies, Washington: American Association of Museums, 1997.
Copyright Archives & Museum Informatics, 1997