User Interfaces for Museums
By Howard Besser
Affiliation at time of Publication:  Contact Information as of 6/1/98
School of Library & Information Science Visiting Associate Professor
University of Pittsburgh School of Information Management & Systems
University of California at Berkeley


Current developments in the computer world will affect information-handling in museum environments.  The development of more powerful computers, more intuitive user interfaces, and a more computer-literate population will likely force museums to take a new look at how they organize the information they gather and how they provide access to it.  This paper contends that we are beginning to see signs of what will eventually become a trend both towards more user-friendly interfaces in museum automation systems, and towards a re-thinking of the traditional separation between systems for handling exhibition and other user-oriented information from those that handle collection management information.

This paper begins with a brief discussion of how the museum community has been historically slow to adopt new technological developments.  It will then very briefly describe the background needed to develop a good user interface.  It will then look at the division of computerized information into two basic camps (the current dominant conceptual frame regarding the use of computers in museum environments).  It then will show how those frames are beginning to change.  It will then describe features found in some of the more advanced user interfaces constructed for the museum environment.  Finally, it will describe some existing prototypes that I think indicate the direction that user interfaces and museum information organization are headed.

New computer technology slowly enters the Museum community

The community served by the Museum Computer Network (MCN) has been very slow to adapt to changes in the outside computer world, particularly in the area of user interfaces.  For much of its life the MCN had acted as a clearinghouse for batch processing punch card-based systems for collection management (particularly SELGEM) and made a great effort to promote one of these (GRIPHOS) that was designed for use in the Humanities (Vance, 1986).  MCN continued to discuss, support, and promote batch systems such as this for  several years after online terminal access had become commonplace in academic environments.  IBM PCs were introduced in 1981 and information management packages (such as Dbase) were extremely popular by 1983, yet museums did not seriously begin to use such systems until several years later.  The Macintosh computer was introduced at the beginning of 1984, yet the first Macintosh-based collection management system did not appear in the museum community until the mid-1989 meeting of the American Association of Museums (and that system, MacMimsy was merely an IBM PC-based system that was just converted to run on the Macintosh).  In each transistion to a new technology that improved user interfaces (cards to line-oriented terminals to PCs to icon-based Macs), museum collection management systems lagged far behind.  And to date, no major vendor is marketing a museum collection management system that takes advantage of a mouse and pull-down menus -- both very commonplace features in business and education-oriented packages.

The museum community has traditionally lagged behind both the business world and the library world in taking advantage of new technological developments.  There are a number of factors which probably contribute to this.  Rogers characterizes these "late" adoptors as cautious, conservative, and needing to feel that an innovation is well-accepted before adopting it.  (Rogers and Shoemaker, 1971) -- characteristics which many museum administrators would use to describe their organizations.  Perhaps because museums see their role as preserving the past (and sometimes the present) for future generations to appreciate, they also tend to hold onto past tools for organizing that information longer than other institutions might.  And they might insist that an innovation be well-proven before before beginning steps to adopt it.

A very likely contributing factor might stem from the concentration of effort that museums and the MCN have put into the organization and handling of the complex set of information that they must manage (Stam, 1989).  Museums have put so much effort into identifying terminology, deciding what fields of description are really necessary, and developing cross-museum standards, that they have not had the time to carefully follow changes in the hardware and software environments that their terminology and description will be sitting on.

But this hesitancy to take advantage of technological developments until they have become commonplace in the business world serves as an indication that museums DO adopt, albeit slowly.  Not only is the historical evidence clear that museum collection management systems have adopted new technologies once they have proven themselves, but economic factors (such as economies of scale) indicate that borrowing from technologies that have been accepted by the business world will, in the future, be inevitable.  Clifford Lynch has made the argument that the library world (which is much larger than the museum community) is just not a big enough market to support systems designed solely for that world, and that bibliographic retrieval systems will have to borrow heavily from software and telecommunications developed for the business world (Lynch, 1987).  We can expect that the museum community will have to do the same.

If the new kinds of user interfaces currently being used on college campuses, in research institutes, and in some businesses are inevitably going to be used in the museum environment, it is important that museum professionals understand what is coming and plan accordingly.  In this case that involves making efforts to understand user interfaces in general (particularly those that are graphically based), and how one might best design them for use in a museum environment.

Unfortunately, even outside the museum community, research into user interfaces is young and "immature".  In order to design an effective user interface, one must first know something about the potential users of the system. In the past, the museum community did not really know enough about its user community and their info-seeking needs to attempt such projects.  But recently, promising progress for art museums does seem to be coming out of the Getty Art History Information Program with their recent study of how Art Historians do their work (Schmitt, 1988) and their study of the image quality needs of art historians (Ester, elsewhere in this issue).

Because the knowledge of the information-seeking needs of the potential users is still sketchy, this paper contains a number of speculative remarks.

What is a Graphic User Interface?

Graphic user interfaces allow the user to interact with the system in ways that are generally considered more intuitive than line-oriented inferfaces.  The popularity of the Macintosh computer is often attributed to its employment of such a graphic user interface.  In most systems, the user employs a pointing device (usually a mouse) to choose from various options on the screen.  A number of devices are employed to prevent screen clutter without limiting the number of options available to the user.

The most common of these devices is the pulldown menu.  For most of the user session, only a small, labeled box or bar at the top of the screen is visible to the user.  When the user wants to view the possible options, s/he points to the box and holds the mouse button down, and a menu of options to choose from appears below the box.  The user then moves the mouse to the desired option, lets go of the button, and that option will be invoked.  Pulldown menus provide users with many options to choose from without cluttering up the screen.
Another very common component of a graphic user interface is the employment of windows.  Windows allow the user to perform more than one task simultaneously, with each task taking place in a separate window on the screen.  For example, a user might want to view object information in one window while checking donor or authority list information in another.  Windows can be layed on top of one another or reduced in size in order to minimize screen clutter.  Windows that have been reduced include scroll bars to help the user view particular sections of the information at different times without filling the screen.

Other common features of these interfaces include: normal buttons that the user can click on to perform specific functions, cycle-buttons that the user clicks on to rotate through a set of options, and slider bars that permit the user to make adjustments across a broad spectrum of choices (such as choosing the level of brightness for an on-screen photograph).

Two basic forms of computer use

Historically, computer use for non-business museum information [footnote 1]  has divided into two rather different kinds of uses: either for collection management, or for exhibition and/or instruction use.  Though there has definitely been some cross-over, by and large past computer use for information purposes has fallen into one or the other of these categories.

Tools for exhibition and/or instruction use are primarily found in interactive exhibits.  This kind of system often incorporates very graphic displays of both visual and text information and typically incorporates a videodisc player to handle much of the visual material.  Most early work on these kind of systems was done in Science Museums, and was oriented almost exclusively towards young people.  But this kind of use has branched out into other kinds of museums and now serves patrons of various ages.  Some of these systems (particularly those for instructional use) display museum info outside the museum setting, and may in fact be designed by educators with little or no museum background.  Systems such as the Renoir disc or the Impressionists <<chk these>> are sold at low cost to schools or individuals.

Tools for collection management, on the other hand, are oriented towards museum staff for internal use or to answer queries by scholars.  They never extend beyond the museum's walls and are designed for a very controlled user population  The Museum Computer Network has traditionally emphasized this kind of computer use over the exhibition and instruction use.

The user interfaces for these two purposes tend to be vastly different.  Those for collection management systems look like interfaces that were state of the art a decade ago, but look old and clumsy now.  Those for exhibition and/or instruction look much more modern and seem to be more intuitive.

Existing interfaces for collection management share several characteristics (see figure A).  They need to display lots of detailed information.  They end up with cluttered screens.  They expect a trained user.  They use older, line-oriented interfaces, usually on PCs or terminals.  They are oriented towards both data input and data output.

Interface for exhibition or instructional use, on the other hand, share a different set of characteristics (see figure B).  Potential users typically have no experience with either the computer interface or the data.  Therefore, the system needs to be easy-to-use, and intuitive.  It also has to be visually pleasing in order to attract users to it (and keep them from just walking by).  Systems like this often includes images.  Displays on these systems must be uncluttered and easy to read.  And these systems are usually found on newer, more graphically-oriented machines (such as Macintoshes).  And, because they are designed for users who are seen as consumers (rather than creators) of the information, they tend to be oriented only towards getting information out, not towards putting any in.
Of course, this is just a generalization, and there are some exceptions to this.  But, by and large, this is the current state of affairs.

Differences will diminish

Changes in technology and users' expectations are taking place that will result in user interfaces for collection management systems beginning to look like those for exhibition or instructional use.  We are beginning to see signs of this, and will continue to do so in the future.  There are a number of factors that will contribute to this.

As machines become more powerful, graphic displays become cheaper, and authoring tools for graphic interfaces become more available, it will become easier to put more visually pleasing user interfaces on top of collection management systems.

As the more intuitive graphic interfaces become more prevalent, users of collection management systems will begin demanding the implementation of these.  (rising expectations)

The ability to incorporate image display into collection management systems will make these systems look more like the systems designed for education and exhibition.

As more information is entered into the computer, and safeguards are developed to protect confidential information, museums are going to see the advantages of linking together the various types of information related to a given piece.  We will begin to see that many user groups (such as staff, scholars, and patrons) really just want different "views" on a large superset of linked information.  For example, much of the museum information that the scholar wants (attributed date, dimensions, provenance) may already be contained in collection management files.  Likewise, much of what the patron wants in an exhibition setting (title, period, creator) is contained in the same files.  There would be strong economic incentives to expand these collection management files (protecting confidential material), create a second user-friendly interface, and allow direct end-user access by both scholars and exhibition attendees.  It is likely that changes in the computer world are going to lead us to this in the not too distant future.

Examples of "Line-Oriented" vs "Graphic" User Interfaces

A typical screen from a museum collection management system is shown in figure A.  (Though these screens come from Argus, it should be noted that a similar user interface is employed in virtually all competing products currently on the market.)  When presented with the menu selection screen, the user chooses the appropriate selection by typing in the number of the selection s/he wants.  Data display screens (figure A) are typically cluttered, and not very pleasing to the eye.  In many of these systems, even when photographic information is displayed on a separate screen, the data display is still visually very cluttered.

But the clutter can be improved through the use of highlighting and arrow keys, as show in an example from a library system at the University of Pittsburgh (figure C).  In this system the available choices are displayed in boldface (or in another color), and the explanatory information is displayed in normal type.  This leads the user's eyes directly to the most important part of the screen.  Highlighting such as this was not possible in early systems, but is now becoming commonplace in the PC world.  <<need better example that incorporates arrow keys>>

The combination of highlighting and menus can be very powerful and easy-to-use.  In this example from XXXX (figure D), the user employs the arrow key or mouse to move to different menu selections, and as soon as s/he points to a new choice, the computer both highlights the choice, and displays some descriptive notes before the user tells it to actually SEEK that item or display screen.  Again, this is becoming possible on the more powerful versions of the PC.

As stated before, examples of more graphic user interfaces are found outside the museum collection management area.  The Guernica videodisc (figure E) allows the user to choose which video segment to play by clicking on iconic symbols designed to be intuitive to the user.  The options available include: interpretation, eyewitness accounts, Picasso’s comments, definitions, examination of the painting, a global overview, time lines, or the story or evolution of a concept.  Choosing one of these options branches the user to still another iconic screen with further options.

Another example of a graphic user interface is the Getty Museum's Illuminated Manuscripts project, produced by Interactive Picture Associates (figure F).  For this videodisc project the user actually touches the screen to select the appropriate option, branching to another screen with still further options.

It should be clear that both these projects, which are designed for exhibition/education purposes, contain a very detailed set of information about the objects dealt with.  Though the level of detail does not approach that of a good collections management system, the overlap should be obvious.

Prototypes of Advanced User Interfaces

Though advanced user interfaces have not yet been implemented in museum collection management systems, a number of interesting prototypes are currently under development that are likely to be adopted for use in this area.  Examining these should give us some idea of what these systems might commonly look like a number of years from now.  Here I will discuss two prototypes being designed at the University of California at Berkeley (UCB).
To help insure vocabulary control, the UCB Architecture department has designed an interactive browsing tool for examining an authority list.  The Art and Architecture Thesaurus Browser, designed by Brian Smith and Maryly Snow, allows the user to search for particular terms, examine cross-references, and see a visual display of each term's place in the hierarchy.

The AAT Browser is divided into three windows: a search window at the top, a synonym window in the middle, and a hierarchy window at the bottom (see figure G).  The user first types in a desired term, such as stairs  at the top of the screen, and the search window will display every term that contains the word stairs.  (The user can use the scroll bars to see terms that will not fit into the window display.)  The user can then click on any term in this window (solid newel stairs  in figure H), and that term will then be highlighted, as well as displayed in its appropriate place within the hierarchy in the bottom window.  At the same time, all the synonyms for this entry will be shown in the center window.  And by clicking on terms in the bottom window, the user can explode to find narrower terms in the hierarchy.

The UCB Image Database System was designed to be a tool that would eventually bridge the gap between a collection management system and one used for instruction and/or exhibition.  In this system, the user employs pulldown menus to look up valid vocabulary and to compose a query in the top panel of the screen (see figures J and K).  Resulting "hits" are displayed in the bottom screen panel, and pulldown menus can be used to see the full record or high resolution image associated with any of the hits (figure L).

An image browser (figure M) allows the user to simultaneously view a large number of images associated with the hits.  Tools are provided to quickly determine brief information about any image.  Clicking on any image highlights that image and the associated record.  Clicking on any text record highlights both that record and the associated image.  Visual sorting tools are also provided (figure N):  one can point to a number of images and save them as a subset to be studied in more detail at a later date.  Similarly, on can pulldown a menu off of any image and ask to see the complete text record and/or a high resolution version of the image.


In the future, we are likely to see now-separate systems for collections management information and user-oriented information coming together, expanding, and overlapping.  And particularly with collections management systems, we are likely to see more powerful, more intuitive user interfaces.

Further Questions

There are a number of further questions which need to be addressed by interface designers even if user interfaces do not follow the directions typothesized here..


Conversations with Clifford Lynch and David Bearman contributed to a number of ideas in this paper, and David Bearman suggested useful changes to an earlier draft.  Alan Newman helped focus my attention on the subject of this paper.  Chas DiFatta provided technical assistance and  Bill Kownacki provided research assistance.  Steve Jacobson, Randy Ballew, and Ken Lindahl wrote the UCB Image Database software.


Lynch, Clifford A. (1987).  Extending Relational Database Management Systems for Information Retrieval Applications,  doctoral disseration, University of California at Berkeley.

Rogers, Everett and F. Floyd Shoemaker (1971).  Communication of Innovations: a Cross-Cultural Approach, New York: Free Press.

Rorvig, Marc E. (1986).  "The Substitutibility of Images for Textual Description of Archival Materials in an MS-DOS Environment," in K. D. Lehmann and H. Strohl-Goebel (eds.), Proceedings of the Second International Conference on the Application of Microcomputers in Information, Documentation and Libraries (17-21 March 1986: Baden-Baden FRG), Amsterdam: North Holland.

Schmitt, Marilyn (1988).  Object, Image, Inquity: The art historian at work, Santa Monica: Getty Art History Information Program.

Stam, Deidre (1989).  "The Quest for a code, or a brief history of the computerized cataloging of art objects," Art Documentation 8 (1), pages 7-15.

Vance, David (1986).  "The Museum Computer Network in Context," in Richard B. Light, D. Andrew Roberts, and Jennifer D. Stewart (eds.), Museum Documentation Systems: Developments and Applications, London: Butterworths, pages 37-47.

B Voyager or Renoir
C highlighting & arrows, boldface
D new menu comes up when you move to new selection
G AAT Browser (stairs)
H AAT Browser (solid newell stairs)

M browser & hit list correspondence
-End users will be able to search certain kinds of unprotected c.m. info.
-C.M. info can be of definite use to scholars (ie. Provenance)

-LC--experience of user-oriented vs catalogers -- 2 diff user interfaces
-More intuitive interfaces are more device-independant
-Our system--good for people already familiar with SUN-type interface


footnote 1: Museum business office and development systems are not mentioned here because they do not differ substantially from systems used in almost any office environment.  Word processing, financial projections, and maintaining databases of names and addresses are activites in no way unique to museums.

Howard's Home Page
Back to List of Papers

This paper originally written in Microsoft Word for Macintosh 4.0 in 1989.
It was converted to HTML 4 using Netscape Composer 4.04 on 5/18/98.