About the
Doors of Perception 1
CD-Rom
In this article we describe the editorial-, design- and production process of the Doors 1 interactive proceedings.
1 Introduction
2 About Doors1, October 1993
3 Goal Doors1 CD-rom
4 Getting Started; Learning to Speak
5 General Structure 1: Speakers' Documentation
6 General Structure 2: The Auditorium
7 Home Area
8 Connecting the Speakers and the Auditorium
9 Oops & Home
Introduction
In November 1993, the Netherlands Design Institute and
Mediamatic organised Doors of Perception 1 (Doors 1),
a two-day conference in Amsterdam on the design
challenges of interactive media. The conference was
part of the ongoing Doors of Perception programme
of the Netherlands Design Institute.
When afterwards we were discussing the possibility
of a publication, it seemed only natural to produce the
digital `proceedings' of the conference.
This was the first full-blown CD-Rom project by the
Design Institute as well as by Mediamatic. We've
made a lot of discoveries during the six-month production
period and we're writing this text to share some of those
insights with you.
It's the story of how we did it and why.
We'll start by briefly reflecting on the contents of the Doors 1
conference, the issues it dealt with and the questions that
have been posed. We'll then outline our goals in making this
multimedia CD.
The second part of the article will consist of more general
remarks about how we got started, the editorial and production
team and its needs and problems.
Finally, we'll discuss some of the design considerations that led
to this final product.
About Doors 1, October 1993
Organised by The Netherlands Design Institute and Mediamatic, Doors 1 was a groundbreaking conference at which leading thinkers from the fields of graphic and industrial design, architecture, information technology, philosophy, computer science, business and media assembled to consider the cultural and economic challenges of interactivity, the role of design in turning information into knowledge, the challenges of `smart objects' and `smart space' as computing and communications increasingly permeate the environment, the need for an ethical and cultural response to the $70-billion-a-year `digital gold rush' which is sweeping telecommunications, computing and consumer electronics, and the role of design within new forms of industrial organisation, such as the `virtual corporation'.
Issues of interactivity are not only important in the physical world. Telecommunications and cable companies continue to invest tens of billions of dollars each year in `Superhighways of the mind' -- high-speed information networks that will deliver exponentially-expanding volumes of digital text, data, sound and images. As the volume of information available to us increases, sifting out the good from the bad, the desirable from the useless -- in short, turning raw information into useful knowledge -- becomes ever more difficult. It also becomes a question of design.
The questions that dictated the contents of the conference included: What is `interactivity' and how can it be designed? How, for that matter, can interactivity design be taught? What methodologies and skills are needed for what is, by definition, a multi-disciplinary activity? The interactivity designer has a whole arsenal of new tools, including sound, light and space; but how do interactive or digital tools differ from physical tools?
Goal: Doors 1 CD-Rom
The primary goal of this CD-Rom was to capture the conference in its most attractive and valuable aspects. Apart from fulfilling the most mundane expectations one can have of conference proceedings, namely an overview of who spoke and what they said, we decided that we wanted to capture something of the atmosphere of the event, the issues, the opposing points of view. The latter are especially hard to grasp in a conventional proceedings format. Each presentation remains imprisoned in its own rhetorical context. But as our proceedings were going to be interactive, surely we had to do hypermedia.
Our second goal, given the nature of the subject, was to come up with some innovative uses of the medium. We talked extensively about how we had perceived the event and how we might fit all these perceptions into some coherent structure. Along the way we had to abort many of our early conceptions; many ideas emerging from the first meetings proved to be either ineffective, unrealistic or impossible to merge into a consistent structure.
It was - and is - tempting to use every little trick the medium allows, but in the end designing multimedia is not about the fun of adding time, space and sound alone. It is about connections between all these tools, the effects generated by the connections, and the consequences which every single decision and solution can have for content and style and vice versa. It is actually quite easy to assemble an intimidating pile of heavy-handed effects: what is difficult is to select the best possibilities from an endless array of options, to make decisions which show a sensible and attractive consistency, and to keep design simple and lucid.
In this project it was of the utmost importance to find the right balance of editorial, ergonomic and aesthetic ingredients.
Another knife-edge we had to balance on was the one between accessibility and challenge. If you make your interaction too accessible, you risk losing the users' attention. We decided we should not insult our users' intelligence, and therefore to make our interface a bit more complex and powerful, at the same time sacrificing some of its accessibility for novice users. We assumed that people interested in the subject, and in multimedia generally, would spend more than five minutes trying to figure out how the CD worked.
This conflict between direct comprehensibility and a powerful interface is something that will plague interaction designers for decades to come. Nowadays the average person in Western Europe spends the equivalent of three years of full-time training to become a savvy book user. We learn the alphabet, grammar, analytical reading, and a lot of handy conventions like page numbers, table of contents, index, footnotes, et cetera. After this period of training we are able to pick up a book and almost unconsciously get a very good idea of its structure and the quality and quantity of its contents. We figure out quickly how much time we're going to spend with it to get from it what we want. Usually it's not difficult to decide whether to buy or borrow it. We can do all this not only because of our training but also because the book is a very mature medium. All books work in roughly the same optimised way. And our culture and way of thinking are themselves determined by the way books work. This has been true for more than five hundred years.
For interactive multimedia, this development is just starting. We're trained to understand the world as a linear, diachronically coherent collection of causes and effects. Our media provide us with random access to their content, but we always expect to be somewhere on a line. We know where we are, and if we don't, we have tools to quickly figure it out and restore us to safety again.
The non-linear information structure of hypermedia blows the good old order to bits. Time becomes reversible and cause and effect can suddenly switch places. We want that, because we understand that some aspects of our world are more like that. And hypermedia are a very powerful way of dealing with this. But we're just beginning. We still have to develop our tools and grow confident in working with the new media. In a sense, developing interactive multimedia is like making books for children, because we all are children. We are just beginning our schooling. We are actually children designing books for children, because we ourselves are beginners too.
At the moment we are scrambling to invent the interactive equivalents of page numbers, paragraphs, chapters, indexes, tables of contents and footnotes. We are just beginning to develop the aesthetics and rhetoric of hypermedia. We are forced to experiment as designers and as users. These are very exciting times, and they will remain exciting until far into the next century when the best approaches will have emerged and be recognised as such. And when our descendants will learn how to deal with them in elementary school.
So at the moment we cannot afford to design for obviousness. That would limit us to existing conventions. On the other hand, there is no point in pushing development beyond comprehensibility. We must compromise, especially when we are making something that has to be useful for people interested in the content of our products. This places CD-based interactive media somewhere between computer games and location-based applications such as kiosks. Games invariably have a high challenge component in their design. The player wants to master the game, but the time required to master it is actually the whole life span of the game. Once mastered, it becomes uninteresting. Location-based interaction, such as point-of-sale information systems and possibly interactive TV, on the other hand, has to deliver immediately. The user doesn't have time to learn any new skills. Or the advertiser doesn't want to wait for the user to acquire them. In designing this CD we assumed that we could expect the user to spend at least some time familiarising him- or herself with it.
Getting Started: Learning to Speak
The lack of conventions also meant that we had to work much harder during the development of the product. The people who worked on this CD all had experience in some media production field. From within our respective disciplines, we had all been professionally involved in making books, magazines, film and the like. Here, no one really knew what to expect from anyone else.
It is insufficient for a development team to work with merely a strong intuitive notion of the implications of a certain project. The team must fully grasp the entire issue in all its dimensions. It must have concrete and abstract capacities, with each individual member able to exchange, communicate and interrelate all of his or her insights and skills, as well as to understand those of members from other disciplines. This way of working led to a blurring, even a disappearance, of the distinctions between traditional disciplines.
In total, nine people have been involved in this project, four of them full-time for six months, others for shorter periods or on a part-time basis. The main tasks were: information design, editing; image research; screen design; sound design; interaction design; programming; image processing; and production co-ordination. The team members came from various disciplines including art history, journalism, film, computer game design, graphic design, media theory and literary theory.
At the start of production, communication was difficult. The individual team members all started out from their specific disciplines and spoke their own languages. It took a while before we adjusted our individual professional languages to a vocabulary that was comprehensible to all team members. For an `image person', for example, it is difficult to attend meetings about editorial aspects, whereas the editorially talented person is confronted with a visual language he has yet to learn. And think of the trouble a programmers' language can give non-programmers! Through striving to be teachers and students at the same time, and by devoting a lot of energy and patience to the problem, we finally developed a multidisciplinary language that enabled us to communicate about the project. We feel this is an important achievement.
We ended up with a much closer collaboration and exchange of ideas than we were used to in productions for other media. The designers had considerable editorial input and some of the best design ideas came from the editors. Especially at the beginning of the process, we had lengthy meetings with everyone present, discussing our options in depth.
The input of our partner, The Netherlands Design Institute, especially that of its director John Thackara, has been very useful. The `Vormgevingsinstituut' proved to be the best client we could wish for. Not only did we enjoy John's everlasting confidence and support, but we also profited from his inspiring ideas and constructive criticism.
General Structure 1: Speakers' Documentation
We ended up dividing the contents of this CDrom into three main categories. One of these is the conventional proceedings. This section contains the raw material that came out of the conference; the texts of the lectures, biographical notes about the speakers, some additional papers. We organised these on a speaker-by-speaker basis, basically providing access to the material in alphabetical order.
In this section we didn't really try to exploit the possibilities of the electronic book. There is no way for the user to make notes in the text, there are no cross-references, one cannot even copy a paragraph and paste it into a word processor. (We did provide the full text in MS Word format elsewhere on the CD, though.) This is not because we didn't think all this would have been useful; we would have loved to do it. But given our limited time and budget it was impossible. There were also design considerations that made us choose not to use Voyager's excellent Expanded Book Toolkit for this part of the CD. Although this authoring tool would have enabled us to implement a much better handling of the text at low cost, it would have severely limited us in the development of the rest of the interface. We wanted to go further beyond the conventions of the book than the Expanded Book Toolkit would have ever allowed.
When a user clicks on a speaker's face in the Home area, a full screen portrait comes up. This portrait is not just a visual `cover screen' for information about and from that speaker: clicking the mouth icon at the bottom of this screen enables one to hear a `musical portrait' of each speaker. Although we couldn't resist the temptation to play around a bit with our honourable speakers' voices and physiognomies, our primary intention here was to provide clear and memorable `images' of all the speakers. Images work better than names in helping users to recognise speakers when they appear elsewhere on the CD.
General Structure 2: The Auditorium
The second part of the CD is the result of our analysis of the material. We began looking for a way to systematically interrelate all the information we had. Soon we discovered that the material itself was not systematic at all. It would have been possible to force all the content into some structure, but this would have led to a very disappointing product owing to the utterly messy character of the collection of talks. We were not the authors of the text. We were confronted instead with the task of presenting the work of very diverse authors, from very diverse disciplines, using a wide variety of methods to present their work.
So after a short period of desperation, we decided to stop trying to be faithful to the material and to the internal structure of each and every presentation. We just took bits and pieces and used them for our own purposes. After very thoroughly reviewing all the videotapes of the conference again, we made a selection of statements that we could juxtapose with others, thus taking opinions and insights offered by the speakers completely out of their original context. With these loose bits we created a discussion that never `really' took place at the conference. Of course, at first, this idea seemed very irresponsible to us, as decent, academically trained people. But after a while it seemed only natural.
The structure we developed for organising these discrete samples of the conference is, as we have said, that of a discussion. It is a group of people having a conversation that is very interactive: someone says something, another adds to it or perhaps strongly disagrees. Often someone brings up an idea that is slightly beside the point and others pick up on it, thus leading the discussion in new directions.
The discussion model turned out to be very powerful in establishing connections within this basically incoherent collection of utterances we were dealing with. After testing a few variations we developed a maze of disagreements, agreements and more indifferent relations -- the three basic kinds of relationship in any discussion. The maze is constructed according to a very simple set of rules: each node branches out in exactly these three directions. After a rather intensive period of fitting all the pieces together, and again discarding everything that we couldn't fit into the now further evolved concept, we had a system that could be explored again and again in an unlimited number of ways, always generating a new discourse on the fly.
But we had to tweak the concept a bit further. It proved, for instance, rather tricky to avoid short loops in the structure, where one would end up at the same node after a few branches, and at the same time to make it possible to explore a certain discussion topic in depth without being catapulted into new realms right away. There were no rules, just trial and error. We carefully mapped out the structure on paper (computer screens are too small for this; the flowcharts filled a whole wall in our studio), and gradually achieved the right balance. One trick we used was to divide the maze into eight semi-permeable areas centring on specific topics. Limiting the number of branches that connected these areas enabled us to forge relatively long loops without adding too much to the confusion. This also allowed us to create multiple entrances to the discussion which could be labelled with keywords.
Parallel to developing the information structure, we were faced with the challenge of designing an interface that would enable the user to browse in a pleasant and meaningful way. We've included an early version on this disc (Auditorium Sketch). It consists of a simple screen with three buttons. After listening to and/or reading a statement, the user clicks on one of these buttons: No, Don't Know or Yes. Although the interface did exactly what it was supposed to, which was to enable the user to navigate the maze, it proved unsatisfying. Being confronted with such clear choices stopped users in their tracks. They were forced to choose! Also, after having decided where to go, they often remarked that they didn't feel that what they got was exactly a disagreement or an agreement. After expressing an explicit opinion themselves, they expected something that was a very explicit disagreement or agreement indeed. Their expectations were too high for our still rather messy collection of statements. The bottom line was that users didn't enjoy playing -- it felt too much like work.
The second prototype (Auditorium Prototype) is very similar to the final interface. We decided we had to present the user with a wider array of less explicit choices. To quickly test this option, which actually demanded a different information structure, we removed the three buttons of the original prototype and replaced them with an icon that the user could slide from left to right on the screen. The horizontal position of the slider would then connect to a richer array of choices that would not feel like the first prototype's `voting' model.
The proportional slider evoked very positive reactions from our testers. They quite liked to play with it. They were no longer forced to choose between things, but only to express their feelings or general attitude. This also made it much easier for them to accept what the system gave them.
We now had to adjust the information structure to provide the necessary multitude of choices. We implemented various solutions, adding more branches and introducing random factors into the mechanism, so that the actual probability of going in a certain direction would change according to the position of the slider.
When we began testing new versions, we discovered to our amazement that users couldn't tell the difference between them. Apparently, relieving them of the chore of continuously making explicit choices was enough to satisfy them. Our attempts to give them more control went unnoticed. In retrospect (and after two weeks of programming) we realised that this conforms exactly to the discussion model we were using in the first place. In real-life discussions, people have an enormous choice of expressions: they can say things, they can grimace, they can make approving or disapproving noises, they can express disinterest or contempt. In short, they feel they have many ways of controlling the situation. But their actual power over what their conversation partner is going to say next is extremely limited. Most people don't have that many different things to say anyway, they merely follow a line of thought, and it is often painfully hard to change the direction a discussion is taking. Human beings are used to dealing with the limitations of other human beings, rather than with the rigid logic of a clear information structure. They (well, most of them) like that. They have other words for what we have here called limitations: words like character, opinion or culture.
Upon entering any of the Auditoriums entrances, you are welcomed with an intruductory screen like this:
One obvious problem in the whole structure still had to be dealt with: the
Auditorium and the speakers' documentation were each very self-contained. The
only way to go from one to the other was via the Home area. There needed to be
a way to check out a speaker immediately when he or she said something
interesting in a discussion. The obvious question, `Who is this person?' had to
be immediately answerable, without requiring a lot of systematic research. So
we added a `This Speaker' button to the Auditorium interface, thus enabling the
user to bypass the Home area and immediately jump to a speaker's
documentation.