This paper describes work that has the goal of developing an Internet-based learning environment for enabling the remote engagement with art objects. Our aim was to provide a multisensory experience, incorporating information, sound, media, images, data and hyperlinks relating to art objects, in order to provide a richer sense of the artifact to remote users. To help us achieve our goal we have employed software engineering techniques and developed the Customisable Multisensory toolkit (CM toolkit), an authoring system for creating environments. The paper describes the design considerations for the learning environment, the selection of technologies and system design of the CM toolkit. We then describe the prototype learning environment that was created using the toolkit, 'ARTyFACTS'. This prototype was evaluated with 2 groups of users, and the results of this evaluation is presented and discussed. We conclude the paper by outlining areas for further development work, and some of the considerable challenges that remain if we are to develop full multisensory systems for remote learning and engagement with art objects.
Virtual learning environments, World Wide Web, art objects, multimedia, multisensory, software engineering, user profiles
6,825 words
Within the context of the Internet and World Wide Web and the explosive interest in the arts, our project aim was to begin a process where an art object would become accessible and provide a multisensory experience within an online learning environment. Educational environments and centres for informal learning like museums and art galleries are moving away from transmitting information and knowledge via textual information. The application of more readily available and cheaper technology, including the Internet, makes this a reality, and allows a more engaging and fulfilling experience to be created. Text-based delivery modes are also at odds with the requirements of young people today, who are enormously fluent and skilled in other modes of communication, which are increasingly becoming very visual and very auditory (Morice, 2000; Stewart, 2001).
The move towards online environments or repositories of digital content for remote viewing or learning is therefore a very attractive proposition, both for the owners of collections of artworks and the audiences interested in them. Collections can be made available for audiences unable to travel to the physical location, for example because of distance or disability, and ephemeral artworks and artworks no longer in existence in the physical world can be preserved in a digital form.
Whilst the creation of a digital reality is a real possibility, what will viewers make of virtually represented objects in the virtual learning environment, virtual museum, gallery or collection? Early attempts to put collections online often resulted in badly digitised photographs, small and badly rendered 3D models or lots of descriptive text. Objects were presented with no communicated sense of their own materiality, presence, history or politic. We felt that these objects could offer a multisensory experience through less passive, more creative and innovative uses of digital technology, representing them through simultaneous layers of meaning: literally getting under the skin of culture (Baker, 2000). In addition to being able to view the objects themselves, our perspective is that users should have access to contextual information, such as contemporary images and music, as well as historical, cultural and socio-political texts to help provide a richer experience.
This paper describes work that aims to provide such a multisensory learning experience. The ARTyFACTS project (Cook, Baker and Oriogun, 2000) is an on-going research project. It was made possible by a small grant awarded by the Arts and Humanities Research Board in the UK. The work is a collaborative venture between lecturers, artists, software engineers, multimedia designers and developers, musicians and pedagogical scientists. Software engineering techniques were intentionally employed to design and build a robust software system. A major part of the work was the creation of a customisable tool for creating environments for the exhibition and exploration of fine art objects. This was then used to create a small prototype learning environment containing just two art objects, to allow us to evaluate the approach and get feedback from users. The paper describes the design of the Customisable Multisensory authoring tool (the CM toolkit), the design concepts behind the ARTyFACTS learning environment, and a description of the prototype. We then outline the user evaluation process and its results. The paper concludes with discussions and an outline of future research and directions that could be taken in light of the user evaluation, and the interests of the project team.
One of the underlying research questions used to guide the design phase was 'can extensive information, sound, media, images, data, and hyperlinks relating to an object be successfully interfaced to give a sense of that object to a remote viewer?' Whilst our target user base was broad, we wanted to facilitate learning experiences that could support both formal and informal learning situations. Our intention was to create more than a virtual exhibition - we wanted users to be able to learn about the art objects from a variety of perspectives, including contextual knowledge to broaden the learning experience.
The ARTyFACTS learning environment has been designed to draw upon constructivist multimedia learning environment principles that emphasise rich user-centred interactions. Constructivism in our context enables the social construction of knowledge through the use of multimedia learning environments (e.g. see Boyle, 2000). Furthermore, our multisensory approach has much in common with the concepts of modalities and channels in interactive multimedia communications (Elsom-Cook, 2001). There are five modalities: touch, taste, sight, hearing and smell. Many channels of communication can exist within a single modality. So, for example the modality of hearing can have different channels such as receiving spoken language, listening to music, or hearing noise. A key distinction is that in multisensory communications we are, ideally, employing multiple-modalities and channels of communication in order to get under the skin of an art object, and to subsequently convey this through various channels to the user.
A key part of the development process was to consider how these design aims could be realised through the application of state-of-the-art digital and online technologies. Below we discuss some of these technologies from the perspective of developing multisensory learning environments that are capable of displaying art and cultural objects.
Digital imaging techniques can be utilised to illustrate the layers of information and detail of how artworks were constructed, detail which is now unveiled to conservationists and scientists through the application of modern scientific techniques. For example, infrared photography is commonly used to analyse the layers of material and pigment that make up a painting, and digital scanning techniques can reveal the cracks on a painting's surface. This level of detail can now be made available to wider audiences.
The digital preservation of ephemeral artworks, or artworks now obsolete, is a key area of concern for our project. By combining available records with constructed models, an artwork can be preserved over time. Photographs, video clips, press cuttings, critiques, personal reviews of the experience can not only convey the original work, but combined in an multisensory medium, can bring it back to life in a rich and engaging way. For example, Pierre Vivant's temporary installation 'The Space of Use' at the Millais Gallery in Southampton, UK, stands as a marker for what simple video surveillance and software technology can do to re-exhibit a work that no longer exists to a global audience (Vivant, 1998).
Virtual reality models, reconstructions and simulations can be created to represent what an object looked like or how it was constructed, if the actual object is unavailable. Three-dimensional digitised representations of the original object can be created, that can be viewed from different angles, or animated to be viewed one stage at a time. An example is the analysis of perspective and composition through animation in Pierro della Francesca's work produced by The National Gallery in the UK. The locations of artworks and their environments can also be more realistically portrayed. Computer created panoramas from photographs or constructed models can represent the real environment of an object and maps can show geographical location. Accompanying sounds or music can help to convey the atmosphere of the environment, for example the sounds visitors would hear if they visited the location, or typical music of the period or location.
Databases and knowledge-based systems enable objects and information to be classified in many ways, and then recalled according to a user's choice. If the user does not know what it is they want to see, a computer-based agent could make suggestions for them, either randomly, or prompting them to enter some parameters, i.e. a date, country of origin, portraits, animals. Furthermore, from an initial selection by the user, related choices could be offered by the agent, e.g. from a date the user could be offered information or hyperlinks on contemporary historical or political events, fashion, architectural style and maps. A new class of user is emerging for whom such support facilities will seem the norm. For example, Morice (2000) has reported that Nintendo games like "The Legend of Zelda" use animated pedagogical agents to provide advice to players.
The Internet is the ideal delivery platform for the project to fulfill its objectives. Unlike CD-ROMs, which need physical distribution to identified users, Internet-based resources are available to all. Recent developments in online technologies allow comparable delivery of multimedia content, for example via streaming media, QuickTime VR panoramas, Flash animated and interactive movies. Once online, further resources on the Internet can also be accessed. Back-end technologies such as database repositories, search engines and software agents can help to provide a transparent but user-tailored selection of content for users to choose from.
A key design feature we wanted to include in the system is to base the presentation of content and its retrieval from the database upon user profiles. This would mean that when a user enters the environment, they are presented with a range of data components closely related to their interests. Art objects are of interest to a broad audience, and we wanted to create a system that could accommodate a range of users with both formal and informal learning intentions. Initially, 7 user types were identified: practicing artist, art student, art historian, teacher, researcher, art lover/enthusiast and tourist. However, for the prototype we restricted the range to just three types of user to make development more manageable: the art historian, the artist and the tourist. Each would have different requirements of the system, and would be interested in different information and data from the available repository. Each user type was profiled in a specification, for a picture of their requirements to be developed, for example, age, likely other interests, what other requirements they may have e.g. access to travel information, other galleries or library collections.
The user profiles were seen as an important part of making the system more sensitive to different types of users. A parallel can be drawn with cookies. When we shop online at, for example, Amazon the vendor downloads cookies onto our computer that indicate our interests. This approach is used to notify us of similar products that may interest us when we next go shopping at Amazon. With art objects, viewers are becoming more and more demanding and come from more varied backgrounds. So, although our user profiling system is rudimentary, we feel it is an important design feature that is trying to fit into the trends as outlined above.
The project team spent some time on the design of the architecture of the system, utilising software engineering techniques to provide a solid foundation for the development of subsequent and larger systems. The early work used techniques such as object oriented design (Bruegge and Dutoit, 2000, Hawryszkiewycz, 2000), and will allow future systems to be built more easily. The main objective at this stage of the project was to devise a system design that allows components to be re-used, a concept that can provide value for money and reliability as systems get larger and more complex. To facilitate this, the decision to create the CM toolkit was taken.
Six modules were identified to comprise the CM toolkit: web page and service manager, web services, user manager, object handlers, search engine and database management system. Following the identification of these main modules and related databases, a technical design was generated for the general architecture of the toolkit. This included data flow diagrams to model the inputs to the system and the required output from the system to the user, State Transition Diagrams for the time dependent behaviour of the system through identified Java classes. The CM toolkit is thus a generic architecture for generating multisensory learning environments. A full discussion of the design of the CM toolkit is beyond the scope of this paper, but for a detailed design specification see Cook, Baker and Oriogun (2000).
As mentioned earlier, the prototype of the ARTyFACTS system was intentionally kept small in its scope, to enable our research aims to be explored, and then evaluated with users before further development was carried out. The context diagram in Figure 1 below displays the highest level overview of the ARTyFACTS prototype.
It illustrates the two possible inputs provided by the user:
The project concentrated on working with just two art objects for the construction of the prototype. Two very different artworks were chosen to illustrate a wider range of opportunities. The first is a 16th Century Italian painting, the other a 20th Century British sculptural work. Bellinni's portrait of Leonardo Loredan the Doge of Venice was painted in 1501, and is part of the collection of the National Gallery in London. Rachel Whiteread's 'House' was created in 1993-4, and is no longer in existence and can only be seen through reproduction and archive material.
The first stage was to obtain permission to reproduce images and information relating to the artworks, and then conduct research into them. We worked with The National Gallery, using their archive to accumulate background information about Bellini, the painting, and the purchase of it by the National Gallery. We also negotiated rights to display the image within the learning environment on a large scale, but in the end did not require this for the small-scale prototype and evaluation we undertook. As Rachel Whiteread is still alive and a practicing artist, we were able to gain permission from Rachel herself via the Anthony D'Offay Gallery, which provided sources, bibliography and supporting information for our research.
An interface was developed to illustrate what the system could look like, as shown in Figure 2 below. This front-end does not have the full system behind it. That is the focus of ongoing work. What the prototype allows us to do is obtain user feedback at the early stages of development. This is very important, as all too often systems are developed that do not meet users' needs. Our approach can thus be described as a user centered systems design approach (Long, 1996; Norman, 1986; Norman, 1996). The prototype was optimised for personal computers running Internet Explorer version 5 as a browser and a screen size of 1024 by 768 pixels. This screen size was chosen to provide a viewing area for the art objects which was as large as possible.
When a user enters the ARTyFACTS learning environment, they are requested to choose a user profile. The selection of the profile (currently restricted to art historian, artist or tourist) takes the user to the main interface. The two artworks are shown at the top right of the screen (Bellini's painting of The Doge or Rachel Whiteread's House). When a choice is made, the primary image of the artifact is displayed in a larger scale on the left-hand side of the screen. The selected user profile is displayed at the top of the screen, 'art historian' in the example shown in Figure 2.
The large spiral on the right-hand side of the screen displays a number of thumbnail icons representing the media files available to the user, relating to their chosen user profile. Selecting a thumbnail icon retrieves the text, movie, sound or image file from the database. A spiral was used because it offered a simplistic visual representation of available data and data files most relevant to the current user type. In the centre of the spiral is the icon depicting the object being explored. When artifact files are added to the database, they are categorised according to their relevance to the user types. Visually, this translates to its position in relation to the centre of the spiral. The closer to the centre, the more relevance it has. The further away from the centre, the less relevance it has. So for example, in Figure 2, clicking on the icon at the bottom of the spiral, which indicates a music clip, provides contextual material for the art object displayed. The user can also point the mouse over the icon to see a text label describing it.
The horizontal toolbar provides additional user functionality. It displays a text label indicating the data file selected, and has buttons for users to control the files. Magnification buttons for zooming in and out of images to view more or less detail, and play and stop buttons for audio and movie files.
Both of the artworks selected for the prototype enable different remote learning experiences for the user. For Bellini's painting, created several centuries ago, users may want to understand and experience what 16th century Venice was like. In an attempt to convey the atmosphere of the period, the spiral for the Doge contains two contemporary musical clips. The bottom icon on Figure 2, for example, plays Venetian Bells played on Ascension day celebrations in the 16th Century, and a related text box appears that provides some historical perspective about this piece of music. Whiteread's House provoked much controversy at the time, even though it was a temporary installation and was subsequently demolished. As an artwork it no longer exists, but there is a wealth of press and archive material which can convey its presence and surrounding controversy to future audiences. Figure 3 illustrates a newspaper article, included within ARTyFACTS to provide contextual information.
By providing contemporary descriptions, images and samples of music, we are able to provide context to art and cultural objects, providing a virtual and multisensory experience. By combining these with archive materials, we can make some attempt towards the digital preservation of ephemeral artworks, and work which is not easily experienced in the physical world.
The prototype of the learning environment was evaluated with users to provide feedback on its use, and to inform future development. A two-stage evaluation process was conducted with two diverse user groups. The first stage was to get some initial feedback from experts, using staff at the University of North London. The second stage involved a larger number of users who were students, and more typical of one of the intended user groups (Henderson, Oriogun, Cook and Baker 2001). Evaluators were given no instruction on how to use the environment, other than an overview of its intended purpose and that it was a prototype. They were given the web address and asked to engage with the environment, and then complete a simple questionnaire. The questionnaire asked questions related to usability, in terms of ease of use and simplicity (Nielson, 2000; Faulkner, 2000), the design (look and feel), loading time (which is key to usability), and future developments. It also asked for comments and suggestions to be made, to draw out qualitative data. The results and analysis of these two evaluations are presented below.
This first stage of the evaluation process was conducted with staff at the University of North London employed within the Learning Technology Research Institute and School of Informatics and Multimedia Technology. They were all experienced in using online learning materials and environments, and some were designers of software products and multimedia learning materials themselves. The aim of this approach was to get some initial, and what we assumed would be very critical feedback, on the interface design and usability of the system by experts in the field, before more typical users were brought in at the second stage. The evaluation was completed by 10 members of staff, and the results indicate that we have more work to do to improve the overall usability and design of the prototype, but that the work done to date is worthy of further development. Table 1 summarises the responses.
| Interface | Very poor | Poor | Good | Very good |
| Usability | 0% | 50% | 50% | 0% |
| Design | 0% | 20% | 60% | 20% |
| Loading time | 0% | 10% | 50% | 40% |
The issue of usability produced many comments, and evaluators were split in their views. 50% rated usability as poor, and 50% rated it good, but no users gave a rating of very poor or very good at the extremes of the scale. Several users found the initial selection of objects and the overall navigation difficult at first, but once they had explored the environment for a few minutes, they overcame these difficulties. One user commented that they liked the break with traditional navigation and would like to be offered the possibility for even greater discovery within the environment. In terms of design and look and feel, 60% of the population rated the design as good, with another 20% rating it very good. Only 20% rated design as poor, and comments from users giving lower ratings indicated that this was due mainly to usability issues, such as size and positioning of icons and graphic elements and problems caused by the required screen size. There was a definite tension between the overall design and its ease of use, with a strong feeling that the user interface is a bit awkward, but also that the spiral motif with its "retro-modernist" look and feel gives the prototype a unique identity of its own.
The issue of user profiles raised many comments. The questionnaire only asked for suggestions for additional profiles that could be incorporated in future developments, rather than a rating of their usefulness. In spite of this evaluators were split in their response - they either gave suggestions for other types of user, or they commented that they disliked or were confused by the profiling system. Some disliked the fact that a user would be provided by content pre-determined by the system (essentially defined by someone else classifying the information entered into the system) and because of this certain types of information would not be available to them. Two evaluators suggested alternatives to user profiles, allowing users to make their own selection from those presented for each art object, or by offering a 'more' or 'less' information option.
Nine of the evaluators (60% yes, 30% definitely yes) said it would be a good idea for the ARTyFACTS learning environment to not only search for information within its own database, but also throughout the web too. But comments were made that users should be warned first that a web search would initiated, that the response time in undertaking the search should not be too long, and that it would be important to create a custom search facility that would only find items pertinent to the user profile. The majority (67%) felt that ARTyFACTS would not be better as a CD-ROM, as a web-based environment offered the possibility for further online information to be accessed, and enabled information in the database to be more easily kept up-to-date. However, many commented that making it available on CD-ROM as well would broaden its use, making it more portable, providing access to the content without the need and cost of an Internet connection, or ideal for use within galleries and museums.
In terms of technical performance, loading time was rated good (50%) or very good (40%), but some commented that optimising the system for PCs using Internet Explorer 5 as a browser with a screen size of 1024 by 768 pixels might restrict its use. Within the University, Netscape Navigator has been adopted as the standard browser, and the screen size was larger than the average monitor size supplied for most users.
All the evaluators bar one would like the prototype to be developed into a full system (10% no, 70% yes, 20% definitely yes). Unfortunately, our questionnaire did not probe users to indicate their reasons, so we don't know why one user gave a negative response. Whilst the evaluators had been very critical of the usability, navigation systems and some of the design features, they felt there was potential for improvement and it was worthy of further development. Improvements suggested were to provide some description of the environment, its purpose and user guidance at the front end to avoid initial confusion to uninformed users, and simplify the initial process of selecting one of the art objects. One user wanted to see how the design of the interface would expand with the inclusion of more art objects, and two users would like more contextual information about the objects included. It was also suggested that "the key to the success of this prototype learning environment will be its usability, and that reconciling surfability and serious inquiry, whilst making it coherent and innovative will be a great challenge".
In the second stage of the evaluation, the prototype was used by 5 third year Architectural students, and 13 first year Architectural Technology students at the UNITEC Institute of Technology in New Zealand. Three members of the teaching staff also evaluated the prototype, giving a total of 21 users in this group. The methodology used for this second group was identical. They had an initial briefing on the aims and scope of the environment, and were asked to complete the same questionnaire. Table 2 summarises the results.
| Interface | Very poor | Poor | Good | Very good |
| Usability | 0% | 13% | 68% | 19% |
| Design | 0% | 19% | 75% | 6% |
| Loading time | 0% | 38% | 31% | 31% |
The users in this group were not as critical of the usability (ease and simplicity) of the prototype compared to the participants from the first group. Only 50% of the user population at the University of North London gave the prototype a good rating for usability, whereas 68% of UNITEC population rated the system as good, with a further 19% rating it very good. Comments included "how simple and uncluttered" the screen was, that it was "straight forward and easy to use", and that "instructions were clear". However, several participants were unsure how to return to the original screen to change their user profiles.
The students, especially the Architecture students, were impressed with the layout of the spiral. A Computer Aided Design lecturer commented after using the system, that although there were no instructions, the spiral gave an intuitive feeling that icons toward the centre were more important or relevant to the inquiry. Other students found the second screen "very stimulating". However, because of the resolution available on the students' PC monitors, the icons on the spiral were difficult to decipher.
The students were more critical than the University of North London evaluators concerning the loading time. Although not indicated by the quantitative evaluation, comments regarding the lengthy loading time indicated that students were unsure if the web address was correct or if in fact the internet was still responding. The fact that files were being downloaded from the UK to New Zealand would also inevitably create some time lag. Staff wondered if content density equated to content richness, and was it appropriate at the time of loading, especially when the user was not prepared for the long wait. After the user profile selection, the next screen loaded so quickly it took some by surprise. Staff observed that the design of the two screens was disparate. When the first screen offers a drop down menu, there is an expectation that this protocol will be followed in subsequent screens. Another comment that echoed several student comments was that there is no obvious way to return to the home page.
In conclusion, the students indicated that they were interested in such a tool becoming available to facilitate course-related learning, and that the prototype should be developed further. There were guarded comments that students would not like to be bounded by the information contained within the ARTyFACTS environment, and that they would use this facility in addition to a regular web search.
Whilst the results of the user evaluation focussed purely on the usability, interface and functionality of the prototype, rather than our conceptual aims and the design of the CM Toolkit and software specification, it provides much useful feedback which can be used to direct further work on the project. There is potential for future work to be conducted in a number of areas.
Firstly, based on the feedback from the user evaluations, the usability of the ARTyFACTS learning environment will be improved. The expert group was more critical of this aspect than the student users, who are more typical of the intended users envisaged for the system. Whilst one could expect this group to be more critical of key design and usability issues as many of them are involved in software system and web-based learning materials design themselves, a number of the suggestions for improvements they raised can be explored. The difference in the two views however does suggest that we need to carry out further evaluations with a small number of galleries and museums and users more representative of our user groups, particularly art historians and art lovers to ascertain their user requirements and views.
The CM Toolkit should also be tested through the creation of a further prototype containing a larger number of objects, and its ease of use evaluated. One of the intended outputs from this research is a tool that can be used by a wide range of users. For example by galleries and museums to offer a richer experience of materials in their collections, or it could be made available to the Visual Arts Data Service (in the UK). Because of the development of the CM Toolkit, it should be a relatively easy process to create additional learning environments populated with data relating to other collections of artifacts.
To facilitate the scaling-up of the learning environment, and make the multimedia artifacts within it re-usable, we could incorporate metadata and ensure that developing standards for learning objects and learning environments are adhered to. The current state of play in terms of standards for the repositories of learning objects is problematic. Content is currently described using a wide range of standards and specifications, including a confusing mix of emerging international specifications (Miller 2001). Important developments in this field include the IMS Metadata Specification [URL 1] and closely related LOM [URL 2] from IEEE, the Dublin Core Metadata Element Set [URL 3] and proposed educational extensions [URL 4], and locally developed alternatives. However, these technical problems are obscuring a more fundamental pedagogical problem, namely that these standards have barely considered how to underpin developments in a theoretically sound way (Boyle & Cook, 2001). Future work in the project will examine ways to incorporate pedagogical theory into multimodal learning objects.
Another area of future work is to develop intelligent agents that can individualise advice and content based on user profiles and preferences. Users are becoming more and more demanding in how information and knowledge is presented, and come from more varied backgrounds, and this is particularly true within art appreciation and e-learning. As Web publishing continues to grow exponentially, and the amount of information and data at our fingertips increases with it, there is a greater need for selective filtering of material to make it digestible and narrow the choices presented when conducting searches. We have endeavored to construct a system that individualises learning by building upon user profiles that can help in making decisions about what intervention to make for a particular learner. Whilst the ARTyFACTS learning environment's user profiling is simplistic and requires further development to make it more effective, as reflected in the evaluation comments, we feel it is an important design feature and one that we want to continue to explore. One of the author's of this paper has already produced an agent called MetaMuse (Cook, 2001) that is able to engage with undergraduate students as they consider musical composition ideas. The techniques developed for MetaMuse could also be applied to the development of an agent for the ARTyFACTS learning environment.
We also intend to extend the functionality of the learning environment to make wider use of Internet technologies to enable interaction and communication amongst users. For example we could incorporate synchronous and asynchronous communication to allow users to discuss and debate topics of interest or issues. For more formal education settings, for example if ARTyFACTS was to support courses in art history or appreciation, experts could be available for planned online chats or consultation through discussion groups or by email. These kinds of remote interpersonal exchanges can be important stimulants and enhancements, and in e-learning are important ways of overcoming isolation amongst remote users.
We also want to re-examine our conceptual aims of trying to bring art and cultural objects to life, allowing multisensory experiences within a virtual environment. Budget restrictions within the project played a large part in preventing the development of more multimedia content, with development work concentrating on system design and architecture and prototype development. This is evident in the prototype, and several users expressed a desire for more contextual files to be included. The limited number of data files within the prototype does not fully explore this concept at present. A number of creative solutions could be developed, for example to include interviews and reviews from critics, artists and art historians, video clips, 3D models and panoramas. In the longer term, we would like to explore how we could integrate developments like haptic input technology (e.g. see Stanford University, 2000). This would expand the multisensory range of communications, allowing future users to actually feel and manipulate objects of interest in an ARTyFACTS environment which could be more akin to a virtual playworld.
The work described here has been exploratory in a number of areas of interest to the project team, but essentially the goal has been to develop online systems and multisensory learning environments, which are vital to allow the enhanced learning experiences required for engaging with art objects.
We envisage that there will be three primary audiences for the results of this project. Firstly the art gallery and museum communities, especially those charged with the responsibility for developing both wider and deeper appreciation of their collections. Secondly, computer scientists with an interest in artificial intelligence, visualization and multimedia in relation to art, design and creativity; and thirdly, educational organisations that are interested in using our system as a learning resource. It is evident from the evaluation comments from our student users that the ARTyFACTS system has the potential for use in the teaching and learning process with the specialists involved (from architecture and construction), especially in relation to distance and e-learning. With respect to other educational applications, students of art history at undergraduate and postgraduate levels, plus their teachers, should benefit from the use of our system, and we feel it also has potential for use within the context of problem-based learning (Grabinger, Dunlap and Duffield, 1997; Oriogun and Georgiadou, 2000).
Our evaluators raised a number of suggestions for further work, and the project team has a number of interests it is keen to explore in more depth. However, we feel that further work should be based on a more solid analysis of the requirements and needs of intended users, and future development should use this as a starting point. It is difficult to ask uninitiated users about what they would like to have in a computer system or learning environment, but with a working prototype to demonstrate the concepts and a series of probing questions it can be possible to draw out a range of needs. There are a number of directions that the project could take, and many possibilities for further research and development work. It will be necessary to take some decisions on which directions should be pursued.
The ARTyFACTS prototype has attempted to preserve the sense and the experience of art works that can be difficult to re-create in the physical world. To 'get under the skin' of them. By harnessing and applying computer technology and the capabilities of the Internet, we can more successfully create these multisensory experiences for remote and disparate audiences, providing opportunities that are not possible in the physical world.
Baker, A. (2000). Skin Jobs. Paper presented at The Symposium in Digital Creativity, Greenwich University, January 14 2000.
Boyle, T. and Cook, J. (2001). Towards a pedagogically sound basis for learning object portability and re-use. ASCILITE 2001, Melbourne, Australia, December, 2001.
Boyle, T (2000) (Professorial Lecture), Learning Technology: Vision and Reality, University of North London, November 22 2000. Slides available at the LTRI site: <http://www.unl.ac.uk/ltri/pubs/t_boyle/tb_proftalk.ppt>
Bruegge, B. and Dutoit, A. L. (2000). Object -Oriented Software Engineering: Conquering Complex and Changing Systems. Prentice Hall.
Cook, J. (2001). Bridging the Gap Between Empirical Data on Open-Ended Tutorial Interactions and Computational Models. International Journal of Artificial Intelligence in Education. 12, to appear.
Cook, J., Baker, A. and Oriogun, P. K. (2000). ARTyFACTS: End of project report to Arts and Humanities Research Board. Learning Technology Research Institute, University of North London, November 2000. PDF copy available on request from the first author.
Cubitt, S. (1998). Catalogues, Charts and Cabinets: Problems of Order in Multimedia. Computing & Visual Culture Representation and Interpretation Chart Conference Proceedings, p.12, London.
Deutsch, D. (1998). The Fabric of Reality. London.
Elsom-Cook, M. (2001). Principles of Interactive Multimedia. London: McGraw-Hill.
Faulkner, X. (2000). Usability Engineering. London: Macmillan Press Ltd.
Grabinger, S., Dunlap, J. C., and Duffield J. A. (1997). Rich Environment for Active Learning in Action: Problem-Based Learning. Association for Active Learning Journal, 5(2) 5-17.
Hawryszkiewycz, I. (2000). Introduction to Systems Analysis & Design, 5th Edition, Prentice Hall.
Henderson K J, Oriogun P K , Cook J, Baker A (2001). Towards a multimedia Learning Environment for Undergraduate Students Studying Architecture and Construction, IV2001 Conference, IEEE Computer Society, ISBN 0-7695-1195-3, pp 621-623
Long, J. (1996). Specifying relations between research and the design of human-computer interactions. International Journal of Human-Computer Studies, 44, 875-920.
Miller, P. (2001). Towards consensus on educational metadata. <http://www.ariadne.ac.uk/issue27/meg/>.
Morice, J. (2000). Skills and Preferences: Learning from the Nintendo Generation. Proceedings of International Workshop on Advanced Learning Technologies, 4-6 December 2000, Palmerston North, New Zealand. IEEE Computer Society Press.
Nielsen, J. (2000). Designing Web Usability: The Practice of Simplicity. New Riders Publishing.
Norman, D. A. (1986). Cognitive Engineering In D. A. Norman & S. W. Draper (Eds). User Centred Design. Hillsdale, NJ:Erlbaum.
Norman, D.A. (1996). Design as practiced. In T. Winograd (Ed.) Bringing design to software. MA: Addison-Wesley.
Oriogun, P. K. and Georgiadou, E. (2000). Towards Making Learning Agents Think of Their Roles in the Learning Process Through a Problem-Based Learning Grid. Problem-based Learning Conference, Temasek Polytechnic, Singapore. Copy of paper available from the conference Web site at <http://www.tp.edu.sg/pblconference/6.htm>
Stanford University (2000). Haptic Input Technology <http://www.stanford.edu/~jennchen/>
Stewart, W. (2001). Keynote talk. CADE 2001 Computers in Art and Design Education Conference Glasgow, Scotland 9 - 12 April 2001.
Vivant, P. (1998). See <http://millais.solent.ac.uk/The%20Space%20of%20Use/Vivant01.html>
URLs
1. The IMS Metadata Specification. < http://www.imsproject.org/metadata/>
2. The IEEE LOM. < http://ltsc.ieee.org/wg12/>
3. The Dublin Core Element Set. < http://uk.dublincore.org/documents/dces/>
4. The proposal of the Dublin Core Metadata Initiative's Education working group.
<http://uk.dublincore.org/documents/education-namespace/>