Museums and the Web 2004, Selected Papers from an international conference
Edited by David Bearman and Jennifer Trant.
Pittsburgh: Archives & Museum Informatics, 2004.
Museums and the Web: Maturation, Consolidation and Evaluation
David Bearman and Jennifer Trant,
Archives & Museum Informatics, Canada
Perhaps the stringent times limited risky new investments. Or did management realize that significant resources were now being directed towards Web-based programming and examine its purposes and methods more carefully? Or did the third iteration of a site encourage its designers to wonder which strategies worked better? Maybe Web-savvy users returning with experience of other sites and options, demanded features they had seen elsewhere? Or was it that the growing body of literature slowly infiltrated the realm of pure invention and led to reflections on practice? It is likely that all of these causes have influenced the measured reassessment of the museum Web presence, and prompted a fuller integration between the virtual presence of the museum and its traditional, physical manifestation. As the Web is accepted as a method of reaching targeted audiences for defined purposes, developers of Web-based museum programming are being held accountable for how well they achieve their objectives.
Refreshingly, there was hardly a paper, proposed or given, at MW2004 that did not allude to evaluation as a planned component of the project being reported. Rarely were technology choices simply related, without being justified. In growing numbers of reports, alternative technologies are discussed and explicit rationales - related to programmatic intent - are invoked to explain the final selection. Institutional objectives, more than technological capability, have become the driver for new Web initiatives. Technical success is not enough; museum Web projects must produce something of value. All this is encouraging as a sign of maturation. As much as we all loved the heady adolescent days when pure technology urges were sufficient for skunk works to launch bravado endeavors, we can see that the true promise of the Web for museums lies in its full institutionalization and acceptance as a major venue in which the full range of museum programming takes place. Fuller accountability is a small price to pay for greater relevance and long term support.
A subtle novelty in this emphasis on measurable results was the increasingly common search for an accepted yardstick. Whereas five years ago those measuring success were as likely to report a rapid rise in number of visitors, or to quote gushing exit interviews, today’s demands are more rigorous and more meaningful. The goal now, and what we seek to measure, is usability- and not just usability by some preferred, and easily satisfied, audience. Museum Web site usability is being judged more broadly; communities with a voice include those who have otherwise been difficult to reach – teenagers, the handicapped, indigenous communities, serious researchers, school teachers and students. We’re beginning to push the definition of usability beyond ‘can the user make the site work’ to ask, Does the user find what they need? And are they learning anything?
Gone are the narrative case studies – ‘how we did it here’ – without an explicitly experimental component. Gone are the pure academic experiments without a museum application. Gone are the applications developed with no survey of the prior state of the art, or state of the art surveys undertaken without attention to how they might shape the next project. The full cycle of research and development is now visible in every serious project. Though any given report may focus on one aspect, the iceberg of a planned undertaking is just below the surface in most of the projects represented.
One subtext of these articles is that their exploration of a particular means of Web interaction is no longer to the exclusion of others. Designers acknowledge the interplay between storytelling, game playing, simulation and database-driven retrieval environments, and the interpretive power of interactive audio files, screen readers and location-aware devices. Where several years ago only one interface was explored in depth, today we see implementations in which many methods co-exist; the paper in this volume may focus on how one of them has been implemented, evaluated and refined so readers are advised to look closely and consider the effectiveness of the alternative approaches.
Similarly, these papers cluster around professional issues – design, metadata, technology selection, standards, community building – without excluding museological issues. Papers reporting on a single technology, such as handhelds, Flash applications or content management systems, come out of projects in which programmatic expectations are high and other technologies are deployed simultaneously.
This intertwining makes organizing a linear print volume into discrete sections sometimes seem artificial. Just as we had a category set, a paper would leap the bounds of one section and reposition itself within another cluster. Fortunately, our introduction, like the virtual representation of objects in the museum, is freed from the physical constraints of having to address each paper in one and only one place. Thus liberated we can explore themes that run through the volume, returning to papers previously seen in a different light in order to examine familiar features from a fresh dimension.
Darren Peacock and his colleagues lead with a review of the issues facing us as we develop state-of-the-art museum web applications. The National Museum of Australia sought to make their collection accessible, which Peacock and his colleagues define broadly as “usable and valuable”. As historians, in the midst of a redesign of their collections searching, they take us back to the first Museums and the Web conference (1997); we are reminded that it was by no means a given that museums and this new medium, the Web, were going to develop a viable partnership. This historical tour and a thorough review of a hundred Web sites (reported at last year’s Museums and the Web conference) show that perspectives gained from the - albeit recent -- past can be crucial to a future implementation. This group shows the power of locating searching in concepts drawn from the historical context of the objects, and from the attributes of the objects themselves. Achieving this demanded a consistent knowledge representation of relations between objects and the contexts of their discovery, collection, use, exhibition and curation. The retrieval environment enabled by these ontologies, History Browser, is now undergoing the kind of rigorous and comparative evaluation they wished had been available for other systems when they began, and provides a baseline for future work.
Formal knowledge models, particularly ones built on facetted and object-oriented foundations, provide a means to move from the explicit link Web to the “Semantic Web” dream of Web inventor Tim Berniers-Lee. In the path-breaking work of Eric Hyvonen and his colleagues at the University of Helsinki, we begin to see what an RDF based Semantic Web could mean to museums; the knowledge representations of cultural collections could indeed provide for retrieval of contents not explicitly indexed. Through a combination of searching on semantically related concepts and browsing a semantically linked (recommendation) system, users can find related materials, actions, locations, times, events, and objects (if it sounds like Ranganathan, it is). While many technical hurdles remain before such knowledge-based self-description replaces machine addresses and directory structures underlying current Web addressing schemes, the experimental implementation described here is an encouraging start. It’s available as a WAP 2 application (mobile phones) - a tantalizing reminder that even fairly far-out applications will soon be embedded in everyday devices.
The Gernsback Machine, as its author’s have dubbed it, is the most ambitious effort reported here to implement a fully facetted knowledge representation and support retrieval along any encoded intellectual dimension of cultural heritage. By deconstructing our cultural categories of time, space and matter and presenting natural language terms as representations on a each of these axes, the Gernsback Machine makes it possible to ‘pivot” on concepts in one facet and view nearby terms within the same idea-cluster. Navigational maneuvers such as stepping, zooming, and flipping reflect the contextual frameworks that envelope the objects of interest while providing exploratory pathways through numerous collections in context. If the bento box - a media resource with a ‘metadata halo’ of grouped concepts - envisioned by Hobbs et al. is a multi-dimensional version of Minsky’s frames, it might serve as the navigation engine for scaling metadata from collections in n-dimensional museum space.
Communication and interaction requires more than textually-driven retrieval and presentation. Video, multimedia and audio enliven experiences of culture and heritage. But the question of how to retrieve, “display” (e.g. play) and manipulate these “non-text” files is of considerable concern. It is not enough to find relevant information (on a related topic). We must present the information in a way that makes sense to listeners, even if it is audio or multimedia. At Simon Fraser University in British Columbia researchers working with Ron Wakkary are exploiting insights from the domain of storytelling to augment the visitor’s experience of an exhibition, identify audio relevant to expressed or inherent interests, and present it in a natural way. Structuring data to satisfy these requirements involves rigorous data collection and metadata labeling protocols as well as concept mapping (in this case to CRM). Audio objects must be designed to be flexibly delivered and the whole system must be governed by user-based rules, conversation-based rules and a “soundscape model”. Designing and meta-tagging media files so that the aural information is encountered in a way that “feels right” to the user is a challenge of knowledge representation, retrieval and information display highly suited to the needs of museums developing multi-modal interaction environments. On-going evaluation of this project will, we hope, help museums exploit multimedia content as a databased resource rather than one requiring handcrafting of narratives.
Just Like Me
At its most granular, the relevance of a cultural institution is measured by its ability to connect with an individual visitor, virtual or physical, and communicate something of lasting value. We augment the reality of our exhibition environments, following expressed or tacitly indicated interests, and tailor our Web spaces to support personal needs, all in the hope of engaging individuals in the ideas inherent in the objects in our care. Meeting diverse visitors with widely varied expectations on their own terms drives us to individuate the experiences we deliver. Fortunately, richer knowledge representation and better understanding of human interaction styles (mediated by smarter delivery services) can enable just this kind of personal touch.
Computer-individuated communication has become an important area in human-computer interaction research. In museum applications, the ability to personalize (or even to adjust interaction based on group characteristics, such as learning styles) has been a feature of advanced design since the advent of the Web. In their survey, Jonathan Bowen and Silvia Filippinni-Fantoni take us for an historical tour presenting a variety of customization, adaptation, individuation and personalization strategies and technologies that support these functions. As Amazon.com has demonstrated using collaborative filtering, each user may have a unique profile and sense of why things are of interest, but they typically share their sense of important relationships with others whose interests overlap. Thus a degree of what feels like personalization can be achieved based on the predictive power of prior visitor choices. It is somewhat tautological to remind ourselves that at a given time in the history, most people within one culture share a common worldview. Personalization and associative technologies provide ways to help enable the micro-associations of individuals construct a macro-message.
The challenge of interpreting cultural collections derives from the distance between historical and contemporary world-views; museums are faced with a constant task of explaining the ways in which people from other cultures and time periods saw the world, to enable contemporary visitors to understand, and perhaps even appreciate, those perspectives. Gundestrup and Wanning introduce us to the Royal Danish Kunstkammer, an artifact from a world that collected objects based on very different criteria than our own. To travel from Apartment to Apartment in this world is to visit familiar artefacts in a new (to us) context. By inviting us to deconstruct this Renaissance “housing” and put the collection together in our own exhibition, they have created an application that enables us to personalize that world, thereby exposing how our Kunstkammer is based on our own, implicit, categorizations. This engaging “meaning-making” activity has the potential of making the objects once trapped in the contexts assigned to them in the Renaissance, accessible to us.
Conceptual structures such as historical categories are barriers to accessibility, but not the type that most interest Di Blas, Paolini and their colleagues. Access by the visually impaired is the focus of their research and test implementations. They believe that the inherent limitations of screen readers could be partly overcome by standards of design beyond those reflected in the W3C Content Accessibility Guidelines. This team suggested additional guidelines, based on a model of an interaction dialogue, and then implemented them in an on-line version of an exhibition about Edvard Munch. Their evaluation of the experience of blind visitors to the Munch exhibition suggests further areas for improvement. Though this work is in its early stages, its exploration of rule-based understanding, and the interaction between explicit presentation frameworks, object metadata and intelligibility to screen readers may be a promising direction for designers and standards makers.
As Gillian Wilson reports, visitors to the Tate Gallery now have the option of a handheld computer-based guide as an adjunct to their experience of the galleries. Evaluation reveals such interactive guided tours enhance the gallery experience for many kinds of visitors. Content presented at Tate goes beyond traditional audio tours, and includes multimedia, games and a British Sign Language interface for the deaf. In addition to interactivity, depth of information (from the collections database), inter-visitor communication, and other extensions of the gallery-going experience, some adaptive features are also introduced as is a social component enabled by ‘texting’ between visitors. Continuous evaluation is allowing Tate to redesign the systems, programs, content and interactions and build a robust framework for the future.
Infrastructures for Educators
Museums make considerable investments in educational resources and programs to augment the visitor experience. Designing such resources to enable re-use of materials, and encoding files to ensure that they can be readily integrated into a multiplicity of publishing frameworks, is a wise practice with a long horizon. Few organizations have taken this approach to building educational resource libraries as far, or consciously enabled as many different uses, as the American Museum of Natural History in New York. Steve Gano and his colleagues explore the rationale and benefits of this program and detail the standards and technology choices they made before creating content for their now impressively broad repository of educational resources. By building modular content, they have created resources that can be integrated into McGraw Hill published kindergarten through sixth grade textbooks and a wide range of museum education products, and can continue to grow over the years without requiring a change in fundamental architecture.
The National Gallery of Art 2003 Teachers Institute explored the pedagogical methods for museum-based digital storytelling. Unlike the reusable educational resources of the AMNH, the NGA storytelling approach values the “voice” of an engaged educator or artist, who weaves a very specific and often highly personal pathway through a topic. This personalized presentation is enabled by content management systems, such as that used at the Chicago WebDocent program (also described by Borst-Brazas, Lorch and York), that assist the teacher writing the story, giving them access to the knowledge inherent in museum artifacts and enabling them to share it with their students (while and allowing administrators to manage resources effectively).
Besides attending to the structure of content in an educational repository, museums developing large scale educational programs must deal with a range of infrastructural issues. Topp and Dickey detail how the University of Alaska Museum has built the kinds of community relations that enable the content being created for the web to be used effectively in schools throughout the state. Attention to the limits of rural telecommunications helped structure the range of multimedia resources and determine how they were distributed; teacher training programs built familiarity curriculum objectives and the ways that objects could be support learning; community curation programs enabled the museum to form partnerships with native peoples to provide grounded interpretations. Importantly, these apparently disparate activities were all understood as an integrated whole, needed, along with a formal program of evaluation, to ensure that the educational venture was successful and perceived to be valuable.
Learning objectives predominate when designing programs for children, but in developing Web-based interactions for adult visitors to the virtual museum, the strictly educational takes a back seat to serendipitous exploration. Brad Johnson looks for experiences that are engaging, that empower audiences to make new connections and discoveries, and that invite visitors to take different paths through the collections, revisiting the museum regularly to retrace their steps if they wish, or take new routes to unexpected ends if so inspired. Like the knowledge representations that supported the facetted retrieval systems featured in the first section of this book, Johnson’s underlying data model allows connections between objects on many dimensions and supports viewing other objects linked by common features. ARTscape supports such searching, but then goes beyond it by giving the visitors power to create their own collections (like the virtual Kunstkammer), personalize their visit, and take home a souvenir record.
Empowering the user, when the user is also the subject, empowers the subject. And in Southern Africa, with few learning resources and limited IT investment, making the oral record of the experience of black South Africans under apartheid available to the South African people themselves (through the Web) is a critical means of engaging the population in a vital project of remembering, and of bringing these memories to the rest of the world. The divergence of these accounts from “published history” is an experience shared by other indigenous peoples such as the Quickan (Nevile 2004) and Burrabura (Crane 2004).
Reconstructing scientific knowledge, no less than reconstructing popular knowledge, requires the mastery of methods and technologies that require new toolsets. In order to enable visitors to visualize dinosaurs and their behaviors (and appreciate recent research findings of the paleontologists at the Senckenburg museum in Frankfurt Germany), Sauer et al. created an application platform, and developed rendering tools that go beyond what is available in the marketplace. Then, they delegated interpretive power, and gave the visitor the opportunity to “create” a dinosaur, in a simulation program. The user can retain “snapshots” as a record of their exploration. Unlike educational programs that confine themselves to recalling the current images and documentation of dinosaurs from a database, the experience-based approach gives the user the ability to create their own organism, within the confines of recent scientific knowledge, and own a new understanding of a scientific problem.
Mixing fun and learning isn’t confined to prehistory – the future on Mars can be just as exciting. Aided by a huge public interest in the current Mars explorations, the Space Science Institute’s highly interactive and experiential MarsQuest Web site is drawing tens of thousands of visitors daily to a range of edutainment activities including driving the Mars Rover (doubtless the slowest interactive video game racing vehicle of all time). The way in which the designers used, and overcame, the structure and interactive style of the physical exhibit, and the choices they made in technologies for implementing their Web-based activities were evaluated on an on-going basis and adjustments made throughout the design. Real-time access to Mars imagery from the NASA rovers was added along with additional materials for teachers to bring the exhibition alive and enable them to integrate current events. In an embracing of the scientific method, most of the interactive elements pose problems that visitors are invited to solve – ongoing evaluation is determining whether the targeted audience, visitors with an interest in science, is actually learning.
Going to Mars is obviously something we can do only on the Web, but museums encounter similar constraints in imagining physical exhibitions all the time. Kirk Alexander and Janet Temos explore two concepts beyond the frame of traditional museum exhibitions, ultimately developed into full rich interactive educational resources for university study. The Nolli Map Interface explores the relationship between the architecture of a city, Rome (over time), and its depictions in art (also over time and with different representational biases), documented in deep primary sources of evidence. This 1748 map of Rome and contemporary “database” of 1320 monuments created by the mapmaker, serve as the interface to the city and provide access to mid-18th century images of these same monuments by Piranesi and Vasi. The two-dimensional becomes three- and contemporary documentation can be viewed in context -– a true example for the Gernsbach machine. Similarly, a large-scale exhibition of the works of a widely dispersed group of artists defied the cost and organizational limits of physical exhibitions but was well-suited to the “digital collection” and representation. German Nazarene artists of the early 19th century have never enjoyed a full retrospective, but an individual scholar could create a virtual retrospective. Furthering their impact, both projects were incorporated into distance education courses, showing how the institutional infrastructure of Princeton’s Educational Technologies Center enabled teachers to interpret cultural resources far beyond museum walls and confines.
Are we there yet?
Most major museum programs of web development today have built in evaluation components. But few commercial developers, working with museums, have structured such a significant piece of research as that presented here by Educational Web Adventures (David Schaller and Steven Allison-Bunnell) and their academic partners at the University of Florida. The importance of the work goes beyond the collaboration of academic research teams with commercial designers, though it would be a pleasure to see more such partnerships. The focus of this study is, arguably, the crucial design question of the day – does the animation supported by Flash technology, designed to engage students, lead to better learning than traditional HTML presentations? These kinds of practical choices face designers all the time, and usually there is nothing besides the current fad and personal preferences to guide the museum administrator making the decisions. Here, carefully following the guidelines for best practices for both HTML and Flash, the designers produced both types of site and conducted in-depth qualitative evaluations with middle-school and college students. While the evaluation does not land a knock-out blow to either camp, it begins to reveal when one presentation style might be preferable to the other and suggests the need for additional methods of evaluative research.
Tackling the implementation of a faddish technology takes a certain amount of courage; reporting on its evaluation more so. The Italian team led by Professors Salmon-Cinotti and Garzotto has been developing and testing location-aware mobile applications in museums, and is now reporting on their initial evaluation. Acknowledging the controversies in this young technology and some of the problems that arise from its simply not functioning as the user expects, they identify simple ways to make users understand that they are not causing the problems, and begin to suggest how users view the decisions designers have made regarding content and interfaces. The challenge here is a significant one, as the complex concepts, rich content, faceted knowledge systems, multiple dimensions of time and space and adaptive interfaces all come together in a small, portable device, designed to be easily used.
If our goal in creating museum Web sites is a successful personal experience, then an understanding of visitor motivations is a prerequisite to assessing our success. Typically, however, evaluations of physical museum visits and visits on the Web fail to ask individual visitors why they came in the first place and what they sought to achieve (except as part of an overall profile of the audience which is then not analyzed discretely with respect to other responses). In their study of motivation factors and on-line museum visits, Kate Haley Goldman and David Schaller put this question front and center. Grounding their research in classical learning theory, they first identify the failure of prior museum studies to use knowledge of visitor’s motivations in their assessments. They gathered significant numbers of visitor responses from a range of institutions and exhibit types, although because they focused on educational activity sites they got input mostly from students and teachers. Perhaps the most interesting thing to come out of this study is recognition that our questions about why people visit Web sites and about the sense they make of their visits are still too crude to reveal insights into the relationship between motivation and meaning-making. That leaves us with fertile ground for a future evaluative research agenda.
Museums and the Web 2004 features reports from mature projects, structured in sophisticated ways, with an awareness of their environment and other activities – including those reported outside the museological literature. We have space here for a less than twenty of the nearly 100 presentations given at MW2004 (these and all other papers are on the CD-Rom and also the Web). But in this selection, the breadth of the field is evident, as is the international flavor of contributions (authors of these papers come from eleven of the approximately thirty-five countries represented at the conference). In selecting fewer than one in three of the papers published in the Proceedings, which in turn represent one in four of those proposed to the MW2004 Program Committee, we are clearly proving just a taste of the extraordinary range of museum activity on the Web.
As suggested by Haapalinen et al. and Marty in papers in the full Proceedings, this field is growing, professionalizing, and developing, in part using the methods and practices of distance education on the Web. We are delighted to see, in the citations in these papers and in our own research, a growing body of literature on the Web, published in journals, reported at other conference, and appearing in books, directly addressing issues at the confluence of Museums and the Web. We hope this taste of some of the richest and most rewarding work inspires further use of the Web as an effective tool for communicating and creating culture.
David Bearman and Jennifer Trant
Toronto, Ontario, Canada
March 8, 2004.
All references in this article are to other papers given at Museums and the Web 2004 and published in the full conference Proceedings (CD-ROM) and on the Web site at http://www.archimuse.com/mw2004/
Julia Bost Brazas, Benjamin Lorch and Sean York “Chicago WebDocent: Bringing Museums to the Digital Classroom – On-line Content Management and Presentation System (CMAP)” http://www.archimuse.com/mw2004/papers/brazas/brazas.html
Geoffrey Crane, “Burarra Gathering: Sharing Indigenous Knowledge” http://www.archimuse.com/mw2004/papers/crane/crane.html
Marjo Maenpaa, Slavko Milekic, and Riika Haapalinen, “Collaborative Teaching & Learning Between Continents: A case study” http://www.archimuse.com/mw2004/papers/haapalinen/haapalinen.html
Paul Marty, “The Changing Role of the Museum Webmaster: Past, Present, and Future” http://www.archimuse.com/mw2004/papers/marty/marty.html
Liddy Nevile, Jutta Treviranus, Sophie Lissonnet, Vera Roberts, “ Rich Experiences for All Participants” http://www.archimuse.com/mw2004/papers/nevile/nevile.html