The Visual Assistant: institutional perspectives on the development of software

Colin Beardon


Abstract
It is argued that institutions have been conservative in the uses to which they have put computer technology. Two projects which involve the development of new applications software are presented. The Virtual Curator and the Visual Assistant share certain design objectives, mainly the desire to give users tools with which they can construct new meanings using digital resources held in archives. Feedback from potential users indicate that, while some objectives have been met, users can have problems using archive material for some practical purposes. An open systems model is proposed as an alternative to the institutional model for software application.
 
 

When a new technology first appears there is a natural tendency for institutions to see it only in relation to their contemporary problems. Computers, for example, have been seen by many institutions as arriving just in time to bring them necessary power in order to overcome some particular administrative burden. We might, for example, wonder how academic libraries would cope with present day demands if they were still using manual circulation systems.

As understandable as this perspective is, it promotes an inherently conservative face of new technology. Joseph Weizenbaum questioned the view that the computer arrived "just in time" to save our institutions from some kind of apocalyptic collapse.

"Yes, the computer did arrive 'just in time.' But in time for what? In time to save - and save very nearly intact, indeed, to entrench and stabilize - social and political structures that otherwise might have been either radically renovated or allowed to totter under the demands that were surely to be made upon them."[ 1]

 Weizenbaum also warns that, while the computer may not be as indispensable as we like to think, once it is introduced into an institution the process "may constitute an irreversible commitment" [2].
 
 

Consider our contemporary libraries or museums as an example. They emerged mainly during the Nineteenth Century to fulfill a number of social, intellectual and cultural objectives of the Victorian era. They have successfully evolved throughout the Twentieth Century and have had to meet the challenges of the consumer society and of rising individualism. They are, as I am sure all readers of this journal are aware, institutions whose relationship to society is extremely complex. At times their continued existence seems quite precarious and, of course, computers can be used to help them cope better within this tradition. We can plan for the retrospective digitisation of collections and their management and we can use CD-ROMs to disseminate encyclopaedic information as part of educational and marketing programmes. These are probably all good things, but we should at least pause and reflect.

When a new technology becomes available its full potentiality is rarely understood at once. It is more likely interpreted through the conceptual structures we have constructed in order to understand the existing dominant technologies. Hence the motorised vehicle was first conceived of as a "horseless carriage" and television programmes were first made by setting a camera before a traditional stage. Eventually, after much experimentation, new ways of seeing the potentiality of the technology began to emerge and the camera was fixed to the front of a moving train, people drew directly onto blank film, and a whole new medium began to emerge. Interestingly, in both these examples, the key element in discovering the true potential of the technology was disrespect, an element that is so far almost absent from our attitudes towards computers with the consequent paucity of really new ideas about how they may be used to change our lives.
 
 

For example, deep down inside most of us, when we use a library we would really like to steal the contents and write all over them. How shocking! (But how else can we explain the popularity of photocopiers in libraries?) If our library only holds one copy of a work, or if it costs money to reproduce it, then we cannot be allowed to do this, and quite rightly so. Our collective response has been to civilise ourselves into accepting our social responsibility when we use a library. We know that we should not steal, or deface material, or put it in the wrong place, or commit any of a number of social 'sins'. Libraries (and museums) are exemplars of civilised behaviour and, for many people, that is part of their charm.
 
 

But the technology of digitised libraries can make that act of social responsibility completely redundant. The contents of a library can be stolen and defaced without cost to anyone: in fact no-one may ever realise that it has happened. There is no need to be 'civilised' in this way any more. The user of the library can now possess all the contents as if it were a private collection (with an inexhaustible capacity for photocopying). This was surely unthinkable when libraries were first formed, so it would seem necessary to try to reformulate what a library (or museum, or university) might be.

The problem is exemplified, in the UK, by the way that we have come to describe the new digital technologies. We call them "information technology" which presumably means that we see them primarily as a technology for the collection, processing and dissemination of 'information'. But what is "information"? I think I know what "knowledge" is (in its various forms) and I think I can separate its usage from that of "belief". I find the word "wisdom" useful on occasions and feel I can use it with some degree of accuracy. The word "data" is more problematic but I can relate it to patterns of bits as they appear in digital media. But what is "information"?

The question has been asked many times, yet I know of no satisfactory answer. Nevertheless, we are basing our entire thinking about how to use the technology around this one idea. In practice it seems to mean that we concentrate upon the formal aspects of what is held.

"No compilation of data ... can be called information unless it has been organized in a meaningful way..." [3]

 and,

 "unorganized facts, or raw data, are transformed into an arranged, ordered, usable form called information." [4]
 
 

The main problem with this formulation is that it is either trivial or it assumes that some degree of arrangement and ordering are already done for us. An alternative conception of the role of the technology is in relation to our working practices in which the arranging and ordering are done by us and it is this conception that I want to develop further.
 
 

By looking at practices (i.e. what it is that people do and want to do) we get a diversity of purposes reflected in a variety of tools. The emphasis switches from 'information' to 'tools' and from the observation of structures created by others to the creation of structures oneself using digital resources.

In the remainder of this paper I want to describe two related software projects that have, in part, been motivated by these considerations. They attempt to break the mould of traditional applications software and are, therefore, more like seminal pieces, which can be 'read' in their own right, rather than potentially very useful products (though the second of them will be released for distribution later this year). They aim to provide a new perspective on the fundamental question, "What might we meaningfully do with the contents of a digital collection?".
 
 

The Virtual Curator project

 The Virtual Curator was a collaborative project between Colin Beardon and Suzette Worden which ran from 1991 until 1996. It was primarily concerned with the development of educational software to support students of Design History. After studying various types of educational and institutional software it was realised that none addressed in a satisfactory way the need to go beyond the passive delivery of digital material. Writing in 1992, we said,

 "The practical task we set ourselves was to devise a piece of computer software that enables the user to be as free as possible from the preconceptions of traditional institutions. The author should be in control of all aspects of the collection of information and its arrangement and be able to make statements in the form of exhibitions or other displays. The starting point is the idea of the museum. The central processes are those of collection, selection, order and arrangement which in themselves give meaning, but the appearance of their product as universal total knowledge is questioned." [5]
 
 

In practical terms we wanted to build software that would enable students of Design History to gain the intellectual benefits from curating a small exhibition, without all of the resources required to put on a real one. The solution required that we turn the concept of the "Virtual Museum" on its head. The Virtual Museum was a recreation of the experience of visiting a traditional museum. Not only was the selection of material replicated, so were the physical properties of the space, along with all of its limitations. In some cases these even extended to a recreation of "antique" picture frames.

In our system, the user was not a museum visitor but a curator. The store room was open and the user could bring in their own material. More importantly they could annotate it and arrange it how they thought fit and then select any of it to be arranged in the form of an exhibition. Visitors to our virtual museum would arrive in the store and leave having created an exhibition.
 
 

The main elements of the Virtual Curator, which has been widely reported on, including in a previous Issue of this journal 6, are as follows. The user is confronted with two spaces, the 'Store' represented by a blue screen, and the 'Exhibition' represented initially by a white screen. The Store is the place where material is accumulated and organised. Material can be in the form of images or text files (it was intended to add video and sound files later). This organisation is achieved by the user distributing the material over several screens, organising it visually within a screen and drawing lines between objects to show associations.
 
 

Having organised material in the Store, the user copies selected items to the exhibition space, which is initially configured as if it were a two-dimensional sheet of paper. Once an image is in the Exhibition it can be resized, flipped, moved around screen, and brought to the front or sent to the back. Text objects can also be introduced, appearing like small title captions which, when clicked, reveal the contents of the file in a scrollable window.
 
 

The background to the Exhibition space need not be a simulated two-dimensional paper, however, but one of a number of possible "intelligent spaces". Several spaces were conceived, representing rooms, display cabinets, etc. In the Room setting, which was implemented, images pasted to the various walls or the floor or ceiling could either act as a repeating pattern, or be made to fit the available space. For example, an image of a whole carpet could be pasted to the floor and would appear in perspective, while an image of a wallpaper sample could be pasted to a wall where it would be repeated (at whatever scale requested) to fill the whole wall before it was also shown in perspective.
 
 

The Virtual Curator was intended to be used the by a tutor with a group of students who would work on a project to mount a simulated exhibition on a given theme. The tutor could provide all, or some of the original source material but may decide to provide none at all. It was hoped that the students would follow the process of collecting and analysing material so that the tutor could check that the right balance and connections were made between the visual material and the supporting textual information. Some comments on the experience of using the Virtual Curator in a real teaching situation can be found on Suzette Worden's www page (http://cccw.adh.bton.ac.uk/ctiad/saw/conf/TITLE.HTM).
 
 

Adaptation to theatre studies

 The idea of the Visual Assistant software began when a group of people from theatre studies saw a demonstration of the Virtual Curator and became interested in its possible use within the theatre as a technique for the rapid visualisation of stage settings in conjunction with some planned digital archives of theatre history.

Initially, the existing Virtual Curator software was examined from this perspective and a list of amendments produced which would make it more useful for these purposes. The principle changes involved the addition of a third 'intelligent space' which was called the 'Stage'. In this space objects are restricted to maintain contact with the floor and their position within the floor space corresponded to a position within the imaginary three dimensional world of the stage. The appearance, therefore, was of two-dimensional objects in a three-dimensional space, which can be thought of in terms of cardboard cut-outs containing images of real things standing upright on a stage.
 
 

By moving an object upwards on the screen, this represented moving it backwards on the stage and the object would get correspondingly smaller. It may also move behind an object that was previously behind it. A number of different viewpoints were also introduced, including top view (a view from directly above the stage in which objects appear as lines); zooming in and out; and viewpoint (which allowed the user to select a point within the auditorium from which to view the stage).
 
 

The resulting software enabled historical material, such as scanned images of stage settings, images of standard props or images of famous actors and actresses, to be quickly added to a Store and then assembled on a stage which could be further defined with a backdrop and covering for the floor and side walls if desired. It was found to be generally more effective if the side walls and ceiling were set to black. The resulting output remained within the tradition of the Virtual Curator. The images did not attempt to be photo-realistically generated from three-dimensional models but, rather, appear more like three dimensional collages which had to be 'read' by the user.
 
 

The expectation, in making these changes to the Virtual Curator, was that theatre people who had acquired only a basic knowledge of the use of computers, could rapidly learn to use the software. Furthermore, if they had access to an archive of digital images and texts, they could assemble material relevant to a play and create visualisations of particular scenes. It was also hoped that these visualisations and the method for creating and amending them would enable constructive real dialogues to take place between the various parties involved in a theatrical production (i.e. the director, producer, lighting manager, etc.).
 
 

The software was given to eight potential users within a number of different institutions for evaluation and feedback. The main points to arise from this were as follows.
 
 

The software was found to be easy to learn and easy to use by new users. One user even decided to use it for the first time in a live demonstration at an international conference and, afterwards, casually reported that there were no problems. Others confirmed that it took no more than five minutes to get some interesting results from using it. It was deliberately provided without a manual or any operating instructions, on the grounds that very few people use them anyway. The only support material was an on-line User manual (in html format on a web site) which gave short descriptions of each menu option.

The graphical representations that were created were considered very good at re-creating the "atmosphere" of a possible stage setting. By contrast, three-dimensional modelling packages were found to require too much detail and to frequently produce "clinical" representations (i.e. without "atmosphere").
 
 

Whilst forcing objects to remain on the stage floor was a convenient way of controlling their movement in three dimensional space (i.e. it gave no possibility of vertical movement), it severely limited the representational possibilities in terms of the theatre. The outcome was that the type of stage was restricted to something we might call "the Village Hall", i.e. it is a flat stage with a flat back and sides, with no possibility of the actors using the height. This was not a suitable model for considering more contemporary performance spaces, or for theatre education in general.

Some interesting comments were also made on the starting point for set design. These may concentrate upon any aspect of the production, for example the movement, tension, choreography, light, mood, plot, etc. The first approaches to set design are not necessarily always concerned with physical objects in a space and can be quite abstract in nature.
 
 

There was no representation of lighting. This was not a simple matter, for the intelligent space is an imaginary one. It is not clear how two-dimensional objects in a three-dimensional space should be lit. It is not clear whether they should be given shadows, for example, as it would begin to favour a photorealistic interpretation of the space, which has been deliberately avoided.
 
 

There was no Print option available. It was intended to provide one but there simply was not sufficient time to implement it for the first test version. Interestingly, no-one complained about its absence or requested it as an amendment. The general feeling was that it was a piece of software for producing images or virtual worlds on the screen and printing did not seem particularly relevant. There is now no intention to include a Print option in later versions.
 
 

Most users did not use material from historical archives, though there was not a great deal available in a suitable form. Specific images were felt to be too specific and to have too many artistic connotations. An image of Laurence Olivier playing Hamlet, for example, was such a specific interpretation of this character that it became overpowering. Its use seemed to immediately determine a large number of factors about a scene that should have remained open in the early stages of design.
 
 

Users found little difficulty seeing their creations as temporary "sketches" which could be changed or painlessly discarded if not successful. This lack of preciousness about their creations was seen positively by users who did not feel that anything dreadful could happen while using the software. Even if it 'crashed' (which it sometimes did) users had not spent a great deal of time with their latest creation and, because they used in relation to conceptual design, the main points were still clear in their heads.
 
 

Some users began to experiment by using the software for purposes other than set design. It was used to make a general public presentation of the HaMLET project (in a similar way to PowerPoint). Another use was to model the conceptual structure of one of the other pieces of software being developed, the Director's Assistant (see below).
 
 

In addition to these comments, there were technical problems concerning the methods used to access files and to save versions of completed work. The problem is a common one when dealing with a mixture of stand-alone and networked file access methods. The Virtual Curator used a network approach to file access in which the address of a file was permanently held but a copy of the file's contents was only taken when it was needed. This had some big advantages: the program ran in a relatively small amount of memory, even though it delivered quite complex graphical scenes and functionality, and when work was saved the file size was extremely small (as it contained only text). The disadvantage was that it became difficult to move a creation from machine to machine, or became unreliable if any of the contributory material was moved to a new folder or renamed. This became a problem for teaching student groups, who needed to be able to take their work away on a disk and restart on any machine.

The Visual Assistant project

 The Visual Assistant is a new piece of software, designed for theatre use but with much wider potential application, which has been written in 'C' for the Apple Macintosh. Version 0.1 of the Visual Assistant became available in May 1997. Its development is taking place within the HaMLET project, which is funded by the EU under its Leonardo da Vinci programme. This project, which involves institutions in Finland, France and the UK, aims to create a multimedia tool kit to support theatre education (http://vconf.hut.fi/hamlet/). The components of the tool kit are: a Writer's Assistant (a training package for script writing); a Director's Assistant (to assist in planning for a production); a Producer's Assistant (to assist the producer) and the Visual Assistant (for visualisation in general).

The Visual Assistant is conceptually simpler than the Virtual Curator. It makes no distinction between the store and the exhibition, but there is the possibility of multiple screens, each of which can be configured as a two-dimensional or a three-dimensional space. There is no conscious separation between foreground and background, operations have an effect on whatever object is selected. If the user clicks on a foreground object then this is recognised first, but if not then the corresponding part of the background will be selected.

Objects are no longer restricted to the floor, which means that two types of move operation are needed: 'planar' move interprets movement as being horizontal and vertical (i.e. on the same plane as the screen), while 'depth' move interprets movement as being horizontal and in terms of depth (similar to locking objects to the stage in the previous version). This means that objects can be moved anywhere within the three-dimensional 'box' represented by the positioning of the walls.
 
 

Every object can contain an image, a text, a link to another screen and a label (but is only required to have at least one of these). An image of a door, for example, can contain a text file describing some features of the door, plus a link so that if it is clicked (with the appropriate option) it will transfer the user to another screen.
 
 

Objects (background walls or images) can now be coloured using the standard colour picker. In the case of walls they are simply painted the appropriate colour. In the case of images, all their non-white areas are painted with the chosen colour, resulting in conversion into a coloured silhouette.
 
 

Two developments that are currently being implemented are stretchable cubes (which allow the three-dimensional box which represents the stage to be populated by simple solid shapes) and light filters which are editable whole screen filters which can be used to simulate the effects of lighting.
 
 

A fully working stand-alone version of the Visual Assistant (version 0.2), with all the features listed above, is planned to be available from a web-server in July 1997. The next stage will be to rewrite the application in Java as a platform independent plug-in for web servers. This will handle images in JPEG format and text files in HTML format. It will also produce an output version of the stage in VRML 2.0 format on request. It is hoped that this version will be available by July 1998.
 
 

Conclusions

 The Visual Assistant project, like the Virtual Curator project before it, is an attempt to develop software that it facilitates the constructive use of digital resources. In doing so it attempts to bypass the traditional institution and to look afresh at some of the reasons why we have such institutions and some of the more liberating roles that may be developed within the context of networked digital resources.
 
 

In judging how far the attempt has been successful, it is possible to identify at least three sets of criteria. Firstly, is it usable as software? The preliminary answer must be positive, but this has only been demonstrated at a relatively trivial level of complexity. That is to say, a really useful product may need a great deal more functionality. I feel confident that this could be achieved without compromising the simplicity of software design and ease-of-use, but it would require more professional, and possibly commercial, development to do so. There are technical issues concerned with three-dimensional rendering, memory management, etc. which are difficult to address properly within the resources of a speculative research project.

Secondly, can it be used beneficially in real life situations? This has yet to be shown in practice, though hopefully it will be tested in the next phase of the project, where there is an intention by a number of theatre studies lecturers and professional producers, to use it for real.
 
 

Finally, does it challenge our traditional ideas of technology and its relationship to institutions? I think that, in a sense, it does but not in the way that was intended. The original aim of liberating the archive by making the software available for people to use in constructive tasks, was not achieved in the particular case of the theatre. There was a tendency to avoid specific historical material when planning particular stage settings. Whether this might change if the images were held differently (e.g. an image lifted from its background and appearing with a white surround is much more useful than a standard rectangular image with a background) we do not know. What we do know is that users are willing to be far more imaginative in seeing the software in terms of a general visualisation tool, of which the stage setting is just one manifestation.
 
 

In the longer term, it is envisaged, within the HaMLET project to agree upon a set of general net-based standards for all developments so that not only can all the applications that are being currently developed interact, but any proposed new "assistant" can, with the minimum of adaptation, simply exist alongside whatever exists in the toolbox. Within this context of truly open systems, we see the possibility of evolving sets of useful software that do not conservatively reproduce the properties and problems of past institutions, but provide a new forum within which new applications will evolve on the basis of users' needs.
 
 

References

1 Weizenbaum, J. (1976) Computer Power and Human Reason, p.31. San Fransisco: W.H.Freeman & Co.

2 ibid - p.28

3 Arnold, R.R., Hill, H.C. & Nichols, A.V. (1978) Modern Data Processing. 3rd edition, p.2. Santa Barbara: John Wiley.

4 Frates, J. & Moldrup, W. (1983) Computers and Life: an integrative approach, p.10. Englewood Cliffs, NJ: Prentice-Hall.

5 Beardon, C. & Worden, S. The Virtual Curator: multimedia technologies and the roles of museums. In: Barrett, E. & Redmond, M. (1995) Editors. Contextual Media: multimedia and interpretation, pp. 63-86. Camb, Mass: MIT Press.

6 Worden, S. (1995) Computers and Empowerment: Connecting Cultural Theory and History Making, Computers and the History of Art, 5( 2), 69-78.

7 Beardon, C. & Worden, S. (1995) Multimedia in education. In: Earnshaw, R. & Vince, J. (1995) Editors. Multimedia Systems & Applications, pp. 173-91. London: Academic Press.