Building the Virtual City:

Public Participation through e-Democracy

Andrew Hudson-Smith, Stephen Evans, and Michael Batty

In this paper, we outline how we have d e v e l o p e d a series o f technologies that

enable planning information to be disseminated to affected citizens so that profes-
sionals and politicians can engage with these stakeholders in realizing more effec-
tive plans. Our main theme is based on the generic idea o f the "virtual city" which
is conceived in terms o f the geography and geometry o f the real city. This is a
digital representation using a variety o f software and multimedia, made interac-
tively available over the web. We begin with a brief c o m m e n t on different types o f
virtual city and then summarize the key p r o b l e m s o f using such virtualities in
public participation, more recently considered as part o f the e-democracy m o v e -
ment. We outline our previous attempts to engage in such online participation in
east London for these have been an essential prerequisite to the d e v e l o p m e n t o f

Andrew Hudson-Smith is systems administrator and senior research fellow at the Centre for Ad-
vanced Spatial Analysis (CASA) at University College London (UCL). He has pioneered various
multimedia, virtual reality methods 3D GIS, and CAD for problems of public participation in plan-
ning. He is currently the director of the Greater London Authority's Virtual London project within
CASA. Amongst his recent contributions, he has published (with S. Evans, M. Batty, and S. Batty)
"Community Participation in Urban Regeneration Using Intemet Technologies," in London: A Sus-
tainable World City and "30 Days in ActiveWorlds: Community, Design, and Terrorism in a Virtual
World," in The Social Life of Avatars, Presence and Interaction in Shared Virtual Environments. He
may be reached at <[email protected]>.eHH
Stephen Evans is also a senior research fellow at CASA. His background is in GIS and he has worked
on several projects including LEO--London Environment Online, PROPOLIS--the development of
GIS interfaces to urban land use transportation models, and currently on the Virtual London and
Camden Panoramas projects. Amongst his recent work, he has published (with P. Steadman) "Inter-
facing Land-Use Transport Models with GIS: The Inverness Model," in Advanced Spatial Analysis:
The CASA Book of GIS and (with A. Hudson-Smith and M. Batty) "Homes in Hackney Point to the
Future," Planning, 12 July 2002. He may be reached at <stephen.evans @ucl.ac.uk>.
Michael Batty is professor of Spatial Analysis and Planning and director of CASA. He has a joint
appointment between the Department of Geography and the Bartlett School of Architecture and
Planning. His books range from Urban Modelling to Fractal Cities. He is editor of Environment and
Planning B. The work of his group can be seen at <http://www.casa.ucl.ac.uk> and in the recent book
Advanced SpatialAnalysis. He was elected a Fellow of the British Academy in 2001 and was awarded
the CBE for "services to geography" in 2004. He may be reached at <[email protected]>.
Knowledge, Technology, & Policy, Spring 2005, Vol. 18, No. 1, pp. 62-85.
Building the Virtual City 63

"Virtual London," the application reported here which is currently being fash-
ioned for widespread dissemination of planning information by the Greater Lon-
don Authority (GLA). We then argue that virtual cities should go well beyond the
traditional conceptions of 3D GIS and CAD into virtual worlds and online design.
But we also urge caution in pushing the digital message too far, showing how more
conventional tangible media is always necessary in rooting such models in more
realistic and familiar representations.

Defining Virtuality

The expression "virtual" was used long before the development of the digi-
tal computer whose origins lie in the development of electricity in the late
nineteenth century. One hundred years later, by the mid-1970s when the com-
puter revolution was quite well-advanced, dictionary definitions still did not
refer to any computational meaning in the term "virtual" (OED, 1974). Only
within the last 20 years has the term become coincident with those realities
which are communicated graphically using desktop, network, and various
immersive media powered by digital computation of various kinds. Yet for
cities, the idea of the virtual is a long-standing characteristic feature, present
since we first began to c o n s c i o u s l y m a n i p u l a t e our e n v i r o n m e n t for all
planning involves "imaginings" e n c o d e d as plans and designs. City plans
are virtual statements, ranging from rather staid projections of the present
to futures that we consciously intend to implement, to futures that are purely
subconscious reflections of our utopian desires. In literature too, the city has
been an icon for the virtual (Calvino, 1974) while our obsession with scaled-
down replicas finds wide expression in the various scale models of buildings
such as those that can be purchased as souvenirs from visits to most large
cities.
With computers and computer graphics, our current concern for the virtual
city has taken a new direction. The media, in which we are able to encode
both existing conditions and future plans, provides an immediacy, an accu-
racy, and a clarity of demonstration that enables us to understand and dis-
seminate this understanding in ways that are very different from those we
have traditionally used. This movement to the virtual world is being fast em-
braced by many professional activities. It has its seeds in the origins of digital
computers that, as soon as they left the laboratory 50 years ago, began to be
used to represent text and graphics alongside their traditional, numerical con-
tent. It then took 30 years for miniaturization to reach the point where com-
puter memories could be uniquely associated with graphics and text in a routine
way, and it took the micro-revolution to make such graphics widely available,
first through games. Since then, the convergence of telecommunications with
computers has evolved a medium, currently the net or web, which makes
these graphics widely and routinely available in a standardized one-to-many
manner. We are only beginning to realize the potential which this revolution is
unlocking for disseminating information about cities. This paper is about one
small corner of this world which we are developing to communicate ideas
about the future of the heart of a world city, London.
64 Knowledge, Technology, & Policy / Spring 2005

Computer graphics goes right back to the beginning when designers began
to represent geometries using vector graphics for a variety of systems of which
the building and small districts of the city were key examples. By the late
1970s, fly-throughs of city centers such as Chicago were being demonstrated,
largely to visualize the impact of new, usually high buildings on the surround-
ing environment. These were visualizations which required mainframe and
mini-computers and remained largely inaccessible to all but their designers.
As such they were more part of scientific visualization which was being ad-
vanced on all fronts in the 1980s using workstations and supercomputers.
Paralleling these developments were rudimentary software packages which
really began with the micro-computer, AutoCAD being the example par excel-
lence. Such software began in almost a toy-like gaming format 20 years ago
but it quickly evolved into industrial strength digital design tools. By the late
1980s, graphical user interfaces (GUIs) were becoming popular, and by the
mid-1990s had become routine. At the same time, the advent of such net-
work-based GUIs from MOSAIC/Netscape to Explorer has moved graphic
communication to the de facto standard for all computer interaction.
There are now many digital virtualities involving the city and these com-
bine differences in technology with differences of philosophy (Batty and
Hudson-Smith, 2001). Although we will not review these here, it is worth
noting significant differences between the key approaches for in any distinct
application, there are usually elements of more than one approach included.
First, the web is now widely used to disseminate information about particular
cities in terms of their services, and increasingly, these web-based cities pro-
vide the interface to government. This is something a little more than simply
replacing paper documentation with its digital equivalent, for services can be
delivered this way in much the same way that economic transactions can take
place over the net. The second virtuality involves more professional usage,
being based on graphic representation of the 2D map and its 3D geometric
representation. The geography and geometry of the city are now routinely
represented in GIS (geographic information systems) and CAD (computer-
aided design) software with various multimedia spinning off from such usage
and finding its way into the first style of virtuality, the web. What was profes-
sional speciality usage yesterday is increasingly the routine usage of today.
The third style of virtuality involves the media for representation. The web
and the desktop are the main ways of dissemination with both being essential
to the delivery of services and information in one-to-many and many-to-many
contexts. More esoteric media, however, based on various kinds of virtual
reality from the headset to theatre to CAVE to virtual world, from soft copy to
hard copy, from intangible media to tangible, are finding their way into pro-
fessional and even routine contexts. Network and wireless interactions are
providing entirely new possibilities for the way users interact with the virtual
and these are increasingly being embedded into our wider environment from
how we are changing our behaviors to the ways in which we are gaining
physical access to these technologies.
This brings us to a fourth style of virtuality. This involves the very penetra-
tion of this infrastructure, the software, the hardware, the orgware, and even
Building the Virtual City 65

the people-ware, into the city itself. Cities are being wired in much deeper
ways than we envisaged a generation ago (Dutton, Blumler, and Kraemer,
1987) as all the first three varieties of virtuality become reflected in the very
production of urban space (Thrift and French, 2002). Hard environments are
being interwoven with soft, generating simulacra, realities within realities,
models of models, on all levels (Baudrillard, 1994). In a sense, this of course
is the kind of cyberspace envisaged by writers such as Philip K. Dick, William
Gibson, and Neal Stephenson where software and hardware are fusing with
the physical fabric of the city, indeed with every artifact we engage with (Gra-
ham, 2004). Indeed a new geography is being fashioned around the way such
digital technologies are being used in the production of space and, although
we will not have time to dwell any further on these here, it is worth noting that
our very own technical efforts in designing and using virtual cities are part
and parcel of these new realities, virtual or otherwise. Nevertheless we will
speculate a little on such deep recursion in terms of the virtual and the real as
our argument progresses.
As we have noted, our focus on the model in this project which we euphe-
mistically refer to as "Virtual London," is inspired by both central and local
government's desire to engage the citizenry in more effective and immediate
forms of public participation. We begin by briefly looking at current develop-
ments in e-democracy, and we illustrate these by reviewing some antecedents.
These provide the technologies and techniques as well as the social and politi-
cal context which have enabled us to fashion the various media that we are
using in constructing the virtual city. We outline the structure of the interface
and the various components which we use to disseminate information about
the city, and then digress to show how different media can be used to commu-
nicate variants of this information in digital and more tangible form. We con-
clude by presenting our intended uses of the models, focusing on the dilemma
of participation posed by both traditional and new media in engaging stake-
holders in thinking about their future environment.

Public Participation and e-Democracy

Virtual cities first became significant in the mid-1980s in two very different
forms. First there was the idea that cities might be wired to enable citizen
participation in a variety of home activities, particularly shopping and other
forms of entertainment. This was paralleled by economic development based
on high-tech networks which would enable cities to gain a new competitive
advantage through the use of high tech for various forms of economic link-
age. Second, there was professional usage through the development of com-
puter mapping and computer graphics which had begun in the 1960s using
remote media such as line printers, vector plotters, and primitive display tubes.
Public participation was in some ways implicit in these developments but the
notion that citizens could actually participate in using such technologies to
enable them to understand and actively interact with those producing plans
was some way off. This required the convergence of telecommunications and
computing to reach a level where there was some standardization and univer-
66 Knowledge, Technology, & Policy / Spring 2005

sality of interaction. This was not achieved until the late 1990s when comput-
ers were acquired for the first time simply as devices to unlock information
flowing across networks. It is this focus that is all-important to the kinds of
participation emphasized here.
In fact there had been various attempts at involving the community in de-
sign using computers dating back to the 1960s. At MIT, Nicholas Negroponte's
Architecture Machine group (Negroponte, 1973) fashioned many c o m m u -
nity-based experiments using computer-linked devices and this group was
amongst the first to actively involve the community in projects in Boston's
South End where community activists could react to designs and produce
suggestions using online teletypes. In fact these early experiments have only
occasionally been repeated as technologies and access have improved, with
much of the action still centering on MIT, in particular the Media Lab. Only in
the last decade has it become possible to generalize these early experiments,
and much emergent online participation is still based on delivering informa-
tion somewhat passively, rather than engaging in two-way dialogue.
The range of participatory styles and approaches which involve engaging
citizens in policy and planning issues is extremely wide, from schemes to
raise community awareness all the way through to formal quasi-legal involve-
ment in the decision-making process. In the United Kingdom, there has been
a statutory right to appeal against planning decisions since the mid-twentieth
century and formal participation has been a requirement for local government
plan preparation since the publication of the Skeffington Report (MHLG, 1969).
It is virtually impossible to classify this range of participation but a useful and
long-standing model was proposed by Arnstein (1969) whose "ladder" maps
participatory processes onto "8 rungs." These rungs range from non-partici-
pation involving manipulation of the public by the plan, planner, or politician
at the bottom of the ladder, to tokenism which reflects the one-way traffic of
providing information under the pretext of consultation midway up the lad-
der, to different degrees of direct citizen power and control towards the top.
The further up the ladder, the greater the engagement of the citizen in the
process of producing a plan. It is clear that engaging the public through com-
puters and networks is a little different from the conventional means of face-
to-face contact and dissemination through hard copy which were the principal
means of dissemination when Arnstein (1969) wrote her paper.
There are an increasing number of attempts at participation using the Internet
through various renditions of the virtual city concept. All we can do here is
note the salient features but these do guide us in what is feasible. Most at-
tempts are not well documented and exist on a hand-to-mouth project basis as
an adjunct to more mainstream work. In a sense, this has always been the case
with public participation but often engagement using digital media is a spin-
off from some other, perhaps more high profile and better-resourced project.
Two key issues stand out: first there is often no provision for maintenance of
the activity once the immediate project has ended. In the case of long-term
projects, often only enough resources to begin the exercise are available. Con-
tinuity is thus always in doubt and this tends to blight such projects. Second,
there is rarely any provision for follow-up to assess how effective the partici-
Building the Virtual City 67

pation has been, as this is rarely budgeted for and if considered at all, is usu-
ally considered an added luxury. Thus a clear evaluation of effectiveness is
lacking for most applications are supply-led, commissioned by agencies who
have some idea of what they want but have not tested this in any way with the
publics which they seek to influence.
This is an important limitation on practice, for there are very few clear
examples of where such new technologies have been thoroughly tested with a
view to evaluating their effectiveness for better participation. Often it is not
possible to compare the use of such technologies with anything that has gone
before, because there are no examples of traditional practice that have taken
place already. New technologies tend to generate their own applications, which
would not have happened had such technologies not been in existence. As we
have implied, there are always resource constraints on such applications which
tend to limit any active testing and evaluation. But the nature of the impact of
such technologies on and within the wider social milieu into which they intro-
duced is so complex that any objective or even considered evaluation is diffi-
cult, if not impossible.
One very clear point, which is emerging from the better-documented stud-
ies, is that the most effective approaches involve not only digital but many
traditional media. For example, the Virtual Slaithewaite project developed by
Kingston (2002) is a straightforward community plan-making exercise for a
small village which is comprehensible and manageable in terms of its spatial
extent. Nevertheless, even in this case where the issues concerning housing
design and change were straightforward, physical models of the site were
used to complement the digital and paper-map versions. Moreover the digital
interface to the map design which was made available to local residents was
the most straightforward available, being based on purpose-built rather than
off-the-shelf generic software. Similar examples of community design devel-
oped over a decade by Shiffer (2001) also mix conventional media and much
face-to-face contact with a modest degree of digital presentation. It is this mix
of traditional with digital that we consider essential if virtual cities are to be
used effectively in enabling a wider public to engage in community design
which leads to acceptable schemes. In many ways, the other feature of digital
participation and the virtual city is the dual use to which participatory systems
might be put. They often contain functionalities and foci which are as useful
to working planning practitioners as they are to those who are impacted by
planning proposals. In short, the line between professional and lay public use
is often blurred in these systems for as yet, there is much experiment and little
standardization.
These styles of e-participation have been given added impetus in the United
Kingdom by the development of explicit and centralized governmental initia-
tives in e-democracy. The idea of delivering services online is at the heart of
the contemporary e-government strategy. As part of this, the e-democracy
initiative aims to increase levels of participation by citizens in the democratic
process at a local and regional level through the use of advanced technologies
that depend intimately on notions of the virtual city and the dissemination of
information in virtual fashion over the net. Technologies are seen as key in
68 Knowledge, Technology, & Policy / Spring 2005

stimulating participation in democratic processes, an essential component in

creating an active sense of citizenship. The focus is on the role new technol-
ogy may play in connecting policymakers with ordinary people, and whether
new technology can help engage parts of the c o m m u n i t y who, for social,
economic, or cultural reasons, do not take part in traditional forms of public
consultation.
It is important to be realistic about how technology can be used to deliver
information to make the process of planning more democratic. At one level,
online delivery of information is simply a contemporary form of publication
which has characterized the role of government planning since it was institu-
tionalized in the late nineteenth century. The need to justify this in terms of
assessing demand for such services does not negate in any way the public
duty of providing information about what government intends. Moreover, the
kind of information that public interest groups and the public in general re-
quire is not likely to be the sort that relates to broader strategic issues about
urban form and environmental quality; it is more likely to be specific, case-
based information about particular proposals that have quite local impact.
Virtual cities can provide such localized information of a more superficial
kind, and it is often this that drives participation. This potential difference
between how government needs to publicize planning issues and what the
public is interested in tends to blur the content of the virtual city. Such virtuali-
ties are first and foremost a visualization of environments acting as a forum
for further discussion and participation rather than providing a detailed set of
tools that the public can use to generate their own plans. The latter, of course,
is an ideal which is often incorporated in virtual city designs. But for the most
part, virtual cities deliver fairly obvious digital content in visual terms which
can be used to show the impact of local change as well as the broader goals
that the public agency aspires to through its planning.
These are goals that have developed quite naturally from the diverse efforts
characterizing the development of online participation over the last decade
(Hudson-Smith, 2003). Virtual cities provide the obvious arsenal of techniques
and concepts which might enable such democracy to be rooted in electronic
media. Online voting systems may be the most obvious form of e-democracy
but information about what might be voted upon is much more central than
the mechanics of how such votes might be recorded. In this context, there is
no real intention to establish such an evaluative mechanism relating to plan-
ning proposals. The focus is much more on how such information might be
provided, and how the public-at-large might react to it both positively as well
as negatively, engaging not only in critique but in constructive debate and
bottom-up design. This is a long way from the original goals for building
virtual cities which were largely for professional purposes and for analysis,
for visualization associated with professional use rather than for a wider pub-
lic. All these goals are reflected in the various projects that we have developed
prior to our current applications to London. To illustrate how we have devel-
oped different approaches and medias, we will now recount our experiments
in digital urban design and visualization which provide the essential context
for applications in the rest of this paper.
Building the Virtual City 69

Antecedents to the Virtual London Project

Our developments of virtual cities are based on two key requirements: that
our renditions of urban geography and geometry need not be detailed to the
point where architectural drawings need be prepared from them and that the
material delivered through the virtual city should reach as wide an audience
as possible. These principles have dominated all our work so far and they
imply a relatively eclectic approach with respect to the way our digital rendi-
tions have been developed. Our emphasis has always been on low cost and
wide accessibility and, from our earliest work which began some seven years
ago, this has focused us on web-enabled virtuality. This does not mean that
we have not tested our models using desktop environments for much of the
rapid prototyping takes place in that medium, nor does it mean that we are un-
aware of more esoteric solutions which involve more immersive virtual realities.
But it does mean that the ultimate products we deliver are primarily available
across the web, largely because we consider the web to be the most widely
understood graphical user interface, more so than desktop software solutions.
2D maps and 3D iconic content which represents visualization of the built
and natural environment captured in the map lie at the heart o f virtual city
construction. This virtuality has been available since the 1970s but it was only
in late 1990s that the digital map could be linked to digital photographic scenes
from the actual city. This is best seen in the association of map content with
photorealistic panoramas which the user can associate directly with viewpoints
within the map, thus capturing the visual quality of any scene in a simple and
cost effective way. Linking panoramas to maps was the technique used in our
first foray into building the virtual city which one of us developed as a tour
through the heart of London (Hudson-Smith, 2003; Batty, Dodge, Doyle, and
Smith, 1998). We disseminated this tour of "Wired Whitehall" in a web-en-
abled environment from which we then began to add much more elaborate
2D and 3D visual content. Our "London Bridges" project, which coincided
with the opening of the Millennium Bridge in 2000 and was developed for the
Museum of London's bridges exhibit, took the panorama idea further adding
to this photorealistic 3D content through which users could navigate and fly.
The production of 3D geometric form can be handled in several ways. The
simplest, which we used in London Bridges, is to "sketch" the geometry onto
some photorealistic rendering of the set of objects, with multiple photographs
from many angles providing added realism and detail. This kind of software
has been used for generating 3D renditions of the past--from old photographs
or paintings of past scenes--and these provide quite evocative visualizations.
Traditional geometric constructions are usually based on computer-aided de-
sign packages such as AutoCAD or 3D Studio MAX which enable highly pre-
cise geometries to be developed but are data hungry for detailed built form. In
the last decade, GIS and Remote Sensing packages such as ArcGIS and ERDAS
have been extended to generate 3D content and the current proprietary plug-
ins for GIS--ArcScene, for example---enable detailed content to be built which
is entirely compatible to the analytical functionalities and querying capabili-
ties of state-of-the-art GIS.
70 Knowledge, Technology, & Policy / Spring 2005

We have explored all these techniques in building geometric content. Our

work with the Hackney Building Exploratory linked panoramas to maps, 3D
GIS, and wire frame rendered forms to photographic detail on desktop and in
web-based applications. These were targeted at raising environmental aware-
ness amongst school children being educated to explore their local environ-
ments. Throughout these projects, we have produced the basic content off-line
using a diverse range software and computer systems but with the intention
that has become firmer as we have progressed, of delivering the content as
snapshots, movies, and simplified fly-throughs which less experienced users
can interact with quickly and easily on the web. Many of these techniques
have come together in our projects with groups in Hackney where we have
been involved in grass-roots community design from the bottom up, develop-
ing 2D maps, 3D geometries of neighborhoods, and panoramas using fast
rendition software which has enabled us to deliver content across the web in
the fastest time possible (Batty and Hudson-Smith, 2001).
We have also begun to explore different digital environments in which this
information can be visualized. In parallel to the model construction, one of us
(Hudson-Smith, 2003) has explored the way virtual worlds can be adapted to
real urban environments. By porting the digital models into such worlds, these
worlds, which have largely emerged from the gaming community, can render
as if they are the "real world" of interest. In short, the virtual world can be the
virtual city but users can interact with each other over the network appearing
as avatars in the world in question. Many users from many remote locations
can thus interact and engage in design decisions. This focus has been devel-
oped as a vehicle for learning the problems of manipulating objects in such
worlds but as we demonstrate in our Virtual London project below, we are
adapting such media to ways in which the public-at-large can interact and
view planning information. In fact, what we show below is how such worlds
can be constructed recursively with the virtual city appearing as another world
within the world itself; porting the virtual city into a world which is rendered
as the same virtual city and placing the model in a context which appears as
though it is a material object in that world, in much the same way that it can be
manufactured as a real model in the real world.
Many of these techniques of model construction have come together in the
public participation project we are involved in for the Woodberry Down Re-
generation (Hudson-Smith, Evans, Batty, and Batty, 2003). This project in-
volves the regeneration of a series of public housing estates comprising some
2500 houses and flats in the London Borough of Hackney. This project is
typical of many grass-roots c o m m u n i t y d e v e l o p m e n t schemes which now
dominate the housing sector in Britain. Public housing developed a genera-
tion or more ago, is being regenerated to bring it up to contemporary stan-
dards while attempting to stabilize the rootless c o m m u n i t i e s which have
emerged within such areas. Public participation is central to this process with
a regeneration team located on the site, involved in everything from daily
problem solving for tenants to implementing a vision for the new community.
The Woodberry Down web site < h t t p : / / w w w . h a c k n e y . g o v . u k / w o o d b e r r y >
which has a more reliable mirror at <http://www.casa.ucl.ac.uk/woodberry>
Building the Virtual City 71

reflects the online participation that we have developed but this parallels and
complements a much wider process of participation based on traditional face-
to-face contact, community meetings, and communication using paper-based
information. The site is largely devoted to routine concerns over services and
housing conditions but woven throughout is visual information about the en-
vironment, based on panoramas loadable from zoomable maps. Part of the
site is reserved for community design and some experimental 3D manipula-
tion and primitive fly-through of the existing community and four options for
the future are available.
This was our first real attempt at bringing the diverse visual content of the
local environment together and representing it in a form accessible to a wide
public. It has set the scene for our current project in that many of the tech-
niques that we developed for Woodberry Down are being refined for Virtual
London. Moreover, we cut our teeth on key issues of public participation in
this project and we will return to these after we have described how Virtual
London is being constructed. One final point before we launch into a more
technical presentation: virtual city modeling has now developed to the point
where once a model is built, many different variants can be spun off from it.
In our Woodberry Down project, panoramas, zoomable maps, and 3D ma-
nipulable design options were all built from the same database. Such different
products can thus be tailored to different kinds of software use and GUIs can
be built to match the preferences of the intended user groups. Moreover the
variants can be adapted to the bandwidth available. As the online population
expands and as bandwidth increases, different products are required and this
is the basis on which Virtual London has been created.

Constructing Virtual London

Virtual London is a partnership between the Greater London Authority

(GLA), ourselves in CASA as the contractors, British Telecom, London Con-
nects and the Corporation of London, under a central government initiative
known loosely as "e-Democracy." The general goal of this initiative aims to
increase levels of participation by citizens in the democratic process at a local
and regional level, and to test the role of advanced technologies in achieving
this. The idea is to explore a range of innovative tools which will stimulate
participation in democratic processes, an essential component in creating a sense
of citizenship. It will examine the role new technology plays in connecting the
public-at-large to policymakers, and whether new technology can help engage
parts of the community who for social, economic, or cultural reasons do not take
part in traditional forms of public consultation. It will develop, experiment with,
and evaluate a suite of state-of-the-art e-democracy applications. Evaluation will
consider how far the new technology can help improve the quality of decision-
making by improving the quality and scope of public debate. Applications
developed in the project will be used in physical face-to-face sessions and
remotely, via the Internet (with broadband and narrowband versions) to de-
velop understanding of the different combinations of tools and techniques
that may best stimulate citizens to participate in democratic processes.
72 Knowledge, Technology, & Policy / Spring 2005

The team has very ambitious aims for the project and the initial phase that
we report here is but a beginning. The heart of this first stage is the construc-
tion of a high-resolution 3D interactive, photorealistic model of London neigh-
borhoods, beginning with the Corporation area--the City which is the financial
quarter, where excellent geometric and geographic data at site specific level is
already available. Moreover, the City has had the greatest functional and vi-
sual change of any area of the metropolis in that the rate of construction there
is greater than anywhere else in the United Kingdom with the turnover of
office staff and movement of firms amongst the highest anywhere. This model
will pioneer the techniques, which will be extended to other parts of London
in due course, but once the model is complete it will be used to develop a
variety of web-based environments in which real policy issues can be ex-
plored and debated. These will take the form of web-based games relating to
a range of live policy issues and will reflect the use of e-democracy tools
developed in structured consultation exercises with local authorities and the
GLA. The team will be able to evaluate the extent, depth, and quality of par-
ticipation of citizens generated by different combinations of tools in different
physical and Internet-based locations, and it is hoped it will lead to a partner-
ship of e-democracy practitioners in London.
As we have already implied, this project is unusual in that it is a "blue-
skies" attempt to push e-democracy forward using ideas developed over the
last decade with respect to virtual cities. This practical focus however is the
only way in which such a project can be developed for it needs to be in situ
even though there is considerable risk involved that potential participants will
find the technologies strange to use. It is thus designed to test a number of
assumptions about the potential of new technology to contribute to demo-
cratic renewal, to build collective knowledge about tools that work in terms of
e-democracy, to develop a set of software tools for use by other public au-
thorities, and to pioneer a method for building such models in other towns and
cities. There have been a number of 2D/3D geometric models of central Lon-
don, some based on work in groups at UCL as well as companies ranging
from architectural consultancies to games developers working on Playstation
2 and XBox consoles. Games such as Project Gotham on the XBox have brought
city modeling to a new level of detail with cities as diverse as Chicago, Flo-
rence, Glasgow, Moscow, and London getting the "console treatment." It would
be foolish to claim that such models compete with fully functional 3D GIS but
they do enter the publics' consciousness and raise the stakes in how such
models can be distributed and communicated. As such, the challenge of creat-
ing a virtual London, with the aim of effectively getting over to the public a
sense of location and place as well as containing embedded geographic infor-
mation, is enormous. Virtual city models, no matter however detailed, always
tend to disappoint in that they do not appear "quite right."Add to this the need
to distribute urban information via the Web delivering it effectively and quickly
for a range of low-end users, then the challenge becomes even more substan-
tial.
The first stage of our project is based on constructing a 2D map and ex-
truded 3D model that is optimized for Internet-based distribution. The map is
Building the Virtual City 73

based on detailed aerial photography onto which the geometric content in 3D

is built from detailed Ordnance Survey MasterMap vector data at scales in the
order of 1:250, using height data from several sources, largely LiDAR. The
model contains three levels of detail: basic block outlines, photogrammetricaUy-
derived sections, and photorealistic texture mapped areas. Each of these lev-
els is designed to explore the trade-off between geometric fidelity, level of
detail, and bandwidth requirements within city modeling over the web. At the
general level the model is being created in ESRI's ArcGIS (Version 9.1 with
the plug-in ArcScene) using LiDAR data with Ordnance Survey MasterMap
which allows us to extrude buildings to heights which represent the average
of the cluster of LiDAR points that define the skeletal shape of the associated
roof lines. The basic block model is shown in Figure 1. Deriving reliable
height data and roofing shapes from LiDAR is a research project in itself, so
this aspect of the model will continue to evolve as new techniques are devel-
oped and higher resolution LiDAR data is obtained. Such development routes
need to be flexible allowing new versions of the model to be easily imported
and exported to the relevant formats. As such the core model is composited in
3D Studio MAX, thus allowing instances to be loaded and unloaded as and
when new information is acquired, while ensuring models load into their geo-
graphically correct position.

FIGURE 1
The Basic Geometric Model Built in ArcScene
74 Knowledge, Technology, & Policy / Spring 2005

Photorealism is achieved either via texture mapping direct to the model or

from the ground up using packages to derive heights and roofing shapes from
o b l i q u e p h o t o g r a p h y . The l o w - e n d s o l u t i o n s such as Canoma f r o m
Metacreations, which we used in our previous work in Hackney, do allow for
quick and easy extraction of textures and models but are not suitable for the
detailed constructions required in Virtual London. We are therefore using
ImageModeler from RealViz which provides a mid- to high-range solutions,
and is being used to provide low polygon count models of key buildings from
oblique photography. The end result in such situations is often dependent on
the quality of the photography, which in our case is provided by Jason Hawkes
<http://www.jasonhawkes.com>. Some of this detail has been added to the
model, which we show in Figures 2 and 3, where we fly through the model at
different levels. This however is not the form in which we intend to distribute
the model. The workstations involved in such visualization and animation are
too specialist with respect to their graphics cards and memories than are ever
likely to be available to the casual user, and in any case, the goal of this
project is to make these kinds of models available across the net. As we im-
plied above, virtual city modeling has now moved to the point where we are
able to develop many different realizations from the more elaborate GIS data-
base, exploiting different kinds of software, meeting user requirements more
effectively, and utilizing available bandwidth.
In addition to the 3D model, a series of 180 x 360 degree panoramas are
being captured which are then mapped onto a globe allowing users to "step
inside" and view panoramas in a true x-y-z space. We developed these ini-
tially for the Woodberry Down Regeneration but there are numerous methods
for panoramic production from single-shot lenses and proprietary systems
such as IPIX. We are using a multi-row solution capturing three rows of 12
photographs merged into a single shot using RealViz Stitcher. These panora-
mas provide a quick and easy way to augment the full 3D model as well as

FIGURE 2
The Model with Ground Detail, Roof Lines, and Rendering
of Key Buildings Visualized in Viewpoint

~. ~ 9 9 84
Building the Virtual City 75

FIGURE 3
A Typical Fly-Through across the City with the "GLA HQ," the "Gherkin" and the
"Nat West" Tower (Tower 42) Clearly in View

providing a means for checking height data and roof morphology. Although
panoramas provide an essentially low-end method of gaining a sense of the
city, they are critical in Virtual London for they provide a unique view of the
city compared to a purely 3D approach which has dominated earlier versions
of such models for other large cities (Batty et al., 2001).
As noted, the 2D element of the work is already aimed at early Internet
distribution and it utilizes the Viewpoint product Zoomview which we show in
Figure 4. The 3D element is however non-optimized in its holding environ-
ment of 3D Studio MAX, resulting in an initial file in excess of 1GB for the
basic model and data. File sizes vary according to the format chosen for ex-
port and in our case we are using the native Viewpoint format of .MTX. View-
point has a number of advantages over other options such as VRML; firstly it
is considerably more compressed than an equivalent VRML export file, and
secondly it is XML-based allowing data integration and modification. Without
getting tied down by file size, it is important to ensure the data is suitably
compressed for modem-based distribution. With this in mind, the model is
further optimized in third-party software, while ensuring its overall integrity.
Currently this results in file sizes below 400K. It is still early days but a clear
76 Knowledge, Technology, & Policy ] Spring 2005

route has been defined in which we can optimize the model sufficiently to
distribute large "chunks" of London across the Internet.
In terms of making use of the model to view and comment on changes in
the built environment, the model is animated and distributed as "non-fixed"
for full public participation. Animations such as options to insert new sky-
scrapers are easily added at the production end in 3D Studio MAX and then
exported as a series of "buttons" that are attached to the model. Clicking on
these buttons results in the coded animations being triggered, allowing a clear
and easy way to compare various d e v e l o p m e n t options. These options are
then linked to a bulletin board that facilitates discussion of the issues involved.
For more open design options, the w h o l e model can be made interactive,
allow users to freely rearrange buildings and then "post" set views via the
website along with comments.
In addition to the 3D browser-based model, the aim is to port sections of
London into a multi-user environment, samples of which we show in Figure
5. This is seen as a key ability to effectively communicate changes in the city.
Users are then able to walk around an interlinked gallery as avatars (digital
representations of themselves) containing visual, audio, and 3D media. The
main draw of using a multi-user environment is its collaborative nature and
the ability to communicate with users via either voice or text regardless of
physical location. Traditionally cities are visualized at full size, that is, users
are able to walk around the streets or fly around the buildings using Virtual
Reality but as we note in the next section, this is far too high end and nowhere
in sight with respect to its distribution and access across the net. We are taking
a different approach by using the model within a gallery environment. This allows
various options to be loaded and unloaded depending on the user's choice. Our
experience of virtual worlds and the need to limit these to the exploration of plan-
ning issues and design options, rather than enabling them to be used as exten-
sive visual chat rooms, as in AlphaWorld for e x a m p l e , is b a s e d on our
experience with engaging users in these kinds of forum (Smith, 2002).
There are a number of multi-user environments available on the Internet
that allow users to construct their own worlds. Despite our long-standing ex-
perience of such worlds (Schroeder, Huxor, and Smith, 2001), this is still a
rapidly emerging field based around companies such as ActiveWorlds <http://
www.activeworlds.com>, There <http://www.there.com>, and Adobe <http://
www.adobe.com/atmosphere>. At the present time, we are developing these
worlds using A d o b e ' s Atmosphere environment due to its close integration
with the Viewpoint format, which is the basis for viewing the main model. In
addition to this, Atmosphere has the ability to be embedded seamlessly in an
Acrobat document which removes the user's need for separate plug-ins and
allows 2D information to be easily integrated with the 3D. The ability to im-
port models, complete with animations, opens up the ability to hold public
meetings in the virtual environment and discuss issues relating to changes in
the built environment. In addition, maps will be integrated in each scene by
placing Flash or SVG animations in a similar manner to the way a picture
might be hung in a gallery but with the ability to interact in the same way as in
a standard 2D web page.
Building the Virtual City 77

FIGURE 4
The Map Extent (a) and the Detail of Zooming into a Landmark Site (b)

i: @' 0 ~ ~ ~ ~ : ' ~ ~I' e : i ~ . ~ : m .:~:~ :::, :::~~ ............

......... lllllnl 111]11

78 Knowledge, Technology, & Policy / Spring 2005

FIGURE 5
Porting the Digital Model and Various Panoramas into a Virtual World
Building the Virtual City 79

New Media for Display and Dissemination

We will indicate how we intend to stitch all this visual material together in a
seamless web-enabled interface in the final section but, before we do this, we
will digress a little, sketching its implications for the media in which virtual
cities might be distributed and utilized. The medium we are working with is,
of course, the net that lets us communicate with potentially millions of users.
This restricts the level of detail that we can show but it massively enhances the
range of communication that is available. The virtual worlds' software based
on Atmosphere lets us make such a world from the virtual city model itself or
make a different w o r d - - a n exhibition s p a c e - - w h e r e we can place the model
as an exhibit. The key advantage is that whatever is in the virtual world is seen
from the vantage point of the user as an avatar with the users collectively or
individually manipulating the objects that comprise this world.
What we found when we produced our worldwide surveys of virtual cities
for the Corporation of London in 2000 (Batty et al., 2 0 0 l ) was that many
users and clients were uncomfortable with the new digital media. They wanted
some form of tangibility, some link to the material world which is the subject
of their interest. The Corporation of London, for example, postponed their
decision to acquire virtual city models despite demands by its various depart-
ments due to the fact that their material model was regarded as forming an
important medium for discussion and negotiation in ways they considered a
digital replacement could not. A halfway house is to use the digital model in a
context that makes it appear in material form. By producing a simulation of a
place where negotiation and discussion of planning information can occur
and by putting the digital model into the place as an artifact which can be
manipulated by those in that place, a digital rendition of the traditional forum,
or "simulacra" in Baudrillard's (1994) terms, can be emulated. We show such
a simulation in Figure 6 where we see the ArcScene model placed on a virtual
table with avatars around it, manipulating the "Gherkin"-shaped building, which
is one of the contemporary icons in the City of London. This is a fairly experi-
mental context but our past experience with AlphaWorld and our own labora-
tory version suggests that this way forward is worthwhile.
At the other extreme, it is now possible to produce traditional media from
digital. In sense, this is an extremely obvious use of digital technology as
contemporary photography now demonstrates. Map-making in fact is one of
the major uses of GIS where low cost printers can be used to enable perfect
maps to be produced in that one can experiment over and over again adding
more and more refinement. The ultimate product of course is still the paper
map. So with 3D digital media. Hard copy printers, which print the artifacts in
near traditional form, are now increasingly available. We are a long way from
producing anything that compares with the Corporation of L o n d o n ' s w o o d
and plastic model, for what can be printed from a digital version is small,
intricate, and expensive. In Figure 7, we show a section, which has been printed
from our Virtual London model, from the 3D file that is used in ArcScene and
in 3D Studio MAX. This is a tiny model section and it took two days to print.
But it does show what might be possible from such models and it e v o k e s
80 Knowledge, Technology, & Policy / Spring 2005

FIGURE 6
Constructing the Simulacra: The Digital Model as a "Material Artifact"
in a Virtual Exhibition Space

enormous interest when displayed alongside the digital equivalent on the desk-
top, the web, in a virtual world, and in its traditional material form.
The third kind of media is still digital but takes us into new developments of
immersive or augmented reality. Headset technologies have traditionally been
used for the user to immerse themselves in some virtual scene, or in an indus-
trial context to display some virtual version of an object that a user might be
working on in some kind of eye glass display. These are finding important
usage in precision engineering, in surgery, or in working in hostile environ-
ments where information needs to be supplied to the user continuously. There
are now, however, versions where many users can be immersed in the same
Building the Virtual City 81

Figure 7
Printing the Virtual City Using CAD/CAM Technology

environment. Instead of the user appearing as an avatar, the user views the
model in a headset which is coordinated with images on the headsets of other
users. Users can still see each other interacting with the model but their inter-
action is now coordinated. The ARTHUR project (Augmented Round Table for
arcHitecture and URban Planning) has developed new kinds of wireless head-
sets for this purpose, and as their key example is the City of London with one
of the key manipulations the positioning of the "Gherkin" building, we show
this example in Figure 8. This is yet another media in which these kinds of
virtual city can be used. We could add m o r e - - t h e CAVE, the VR Theatre,
various kinds of holographic display but the point is made. The media or the
use of several media to display planning information is all-important to the
process.

Using Virtual London

The biggest problem in building virtual cities is enabling potential u s e r s - -

professional users such as planners and architects, or wider public interest
groups and the public-at-large--to relate to the content in the most effective
way. Although there is a digital divide, in that there are groups that do not
have access to networked computers, the level of penetration in western soci-
82 Knowledge, Technology, & Policy / Spring 2005

FIGURE 8
The Arthur Interface: Viewing the Virtual City in a Multi-User Wireless
Headset Fully Immersive Environment

Source: <http://www.fit.fraunhofer.de/projekte/arthur/index_en.xml>)

eties is now very high, approaching in London for example the level of tele-
phone penetration some 40 years ago. Over 50 percent of households now
have access to the Internet and half of these have access to broadband (GLA,
2003). This is not to say that the unconnected 50 percent should be ignored,
but at any point, it is more likely the way planning information is distributed
over the net than access p e r se will determine the usage of this new medium.
In the longer term as more and more people get connected, different digital
models will be possible reflecting the preferences, interests, sensitivities and
ideologies of different users, and it is even possible that the models will be
constructed on-the-fly as users indicate their requirements when they log on,
in much the same way that desktop software used routinely is beginning to
configure itself to the user's keyboard behavior.
The way we intend to stitch this digital media together through the web will
be organized in three forms. An overall web interface will guide navigation
through the site to three different areas. These are based on a navigable map
which in itself represents a map of the web site as well as a map of the city, the
virtual city itself through conventional geometric interaction based on fly-
throughs, panning, zooming, and manipulation of objects in the virtual scene,
and the virtual room which we refer to as a "virtual planning observatory."
This incorporates all the features of the virtual world in that users are able to
interact with the virtual city as laid out in table form (see Figure 6) as well as
visualizing different scenes from the existing and future city. At a series of
key locations within each of these areas, a user will be able to navigate to one
of the other virtualities, thus enabling these three different perspectives on the
virtual city to be fully integrated. We will now deal with each of these in turn.
The map-based interface which we show in Figure 4, is z o o m a b l e and
pannable and the various landmarks which are positioned along the top repre-
sent a "film strip" that eventually can be pulled across the screen locating
Building the Virtual City 83

many key features of the city. Point and click on these and the user is taken to
the place in question as we show in Figure 4(b) where information is dis-
played about the place. This is a kind of virtual tourism and the usual hotlinks
to associated web sites are embedded in the interface. We also plan to let users
enter the virtual city or virtual room at points within this map (and within the
virtual room or virtual city), presenting a choice of media in which to interact
with planning proposals. In the first instance, our focus is on individual changes
to the city, in particular on the impact of high buildings, of preservation and
conservation, and of proposals for transport, and various forms of spatial taxa-
tion such as congestion charging. We also envisage that the interface would
act as a showplace for the City in attracting inward investment for the basic
3D GIS model will contain all kinds of site-related information that can be
displayed using the usual form of spatial querying. We intend for this func-
tionality to be somewhat separate from the main site for it indicates that ulti-
mately the virtual city would enable not only lay but also special interest groups
such as developers and business interests to use the system.
The navigable map will also enable various panoramas to be called up at
different locations, associated with entries into the virtual city which we in-
tend to be largely based on abstractions and simplifications from the full 3D
model. As in Woodberry Down, we will produce movies and enable pre-de-
termined options for new developments to be loaded and manipulated, and in
some cases we will let the user move key landmarks. All of this will be pre-
planned. Moreover, as the site is meant to be active in real time, this function
will require constant attention in that there needs to be continual interaction
between professional planners and policy makers and the web site designers
in enabling appropriate content to be put on the site. The virtual city interface
will also parallel the virtual room. So far, the web site is only navigable by a
single user but the virtual room enables any user who is logged on individu-
ally to enter a forum where they can talk with others, and engage various
forms of interaction such as discussion of design proposals. In Figures 6(a)
and (b), we show the virtual room with the city laid out on a table and an
avatar moving the position of the 'Gherkin'. This is typical of the interaction
that is possible in this part of the site. The virtual room is also full of related
media in that the user can point to areas of the model and load panoramas
which will display in the "allegorical" windows of the room.
As yet we have not explored how we script this web site for we are still
developing the media, but we will need to be very careful with respect to how
users interact with it and each other. It may be necessary for collections of
users to "book" the virtual room in advance if certain design possibilities are
to be explored collectively. Moreover, this introduces the idea that for this
kind of e-participation to be most effective, it may need to be paralleled with
more conventional forms of participation. We have already noted that provid-
ing users with the virtual city (Figures 1 to 3), the navigable map and its
panoramas (Figure 4), its tangible model equivalents (Figure 7), the virtual
room (Figures 5 and 6), and headset interaction with the model (Figure 8) all
provide different perspectives on virtuality which deliver diverse information
in different ways. We will focus on such issues in the next stage of develop-
84 Knowledge, Technology, & Policy [ Spring 2005

ment once we begin the process of integrating the different media but, suffice
it to say, that we need to develop different strategies for such stitching. It is
entirely possible that we will consider different types of web site that can be
called up by the user from a master site which determines a palette of web
interfaces most appropriate to the user, groups of users, problem or purpose
of the interaction in question.
All this is under intense development at the present time. What we have
shown here is how virtual city concepts which remained somewhat flat and
superficial when they were first proposed a generation or so ago, are coming
alive in many different ways. The focus is no longer on getting the geometry
right and building a single virtual city but on developing many different vari-
ants which can be tailored to various purposes and delivered to a diversity of
users in different ways using different media. The challenge is to devise ways
of making such delivery both effective and relevant: to develop virtual cities
in which the preferences and abilities of the users with respect to their inter-
ests in the problem in hand, can best be a u g m e n t e d through these virtual
realities.

Acknowledgments

The authors wish to thank Alex Bax and the G L A for support and advice
concerning this project. All opinions are solely those of the authors.

References

Arnstein, S.R. (1969). "A Ladder of Citizen Participation," Journal of the American Institute of
Planners, 35, 216-224.
Batty, M., Chapman, D., Evans, S., Haklay, M., Kueppers, S., Shiode, N., Smith, A., and Torrens, P.
(2001). "Visualizing the City: Communicating Urban Design to Planners and Decision-Makers,
in R. Brail and R. Klosterman (Eds.), Planning Support Systems: Integrating Geographic Infor-
mation Systems, Models, and Visualization Tools, ESRI Press, Redlands, CA, 405-443.
Batty, M., Dodge, M., Doyle, S., and Smith, A. (1998). Modelling Virtual Environments, in P.A.
Longley, S. Brooks, R. McDonnell, and B. Macmillan (Eds.), Geocomputation: A Primer, John
Wiley and Sons, Chichester, UK, 139-161.
Batty, M., and Hudson-Smith, A. (2001). Virtuality and Cities: Definitions, Geographies, Designs, in
P.E Fisher, and D.B. Unwin (Eds.), Virtual Reality in Geography, Taylor and Francis, London,
270-291.
Baudrillard, J. (1994). Simulacra and Simulation, University of Michigan Press, Ann Arbor, MI.
Calvino, I. (1974). Invisible Cities, Harcourt Brace and Company, San Diego, CA.
Dutton, W.H., Blumler, J.G., and Kraemer, K.L. (1987). Wired Cities: Shaping the Future of Commu-
nications, G. K. Hall, Boston, MA.
GLA. (2003). Londoners On-Line: An Analysis of Levels of Home lnternet Access from the London
Household Survey 2002, Greater London Authority, Queens Walk, London.
Graham, S. (2004). The Cybercities Reader, Routledge, London.
Hudson-Smith, A. (2003). Digitally Distributed Urban Environments: The Prospects for Online
Planning, unpublished Ph.D. thesis, University of London, London.
Hudson-Smith, A., Evans, S., Batty, M., and Batty, S. (2003) "Experiments in Web-Based GIS," in
P.A. Longley and M. Batty (Eds.), Advanced Spatial Analysis: The CASA Book of GIS, ESRI
Press, Redlands, CA, 369-390.
Building the Virtual City 85

Kingston, R. (2002). "Web-Based PPGIS in the United Kingdom," in W.J. Craig, T.M. Harris, and D.
Weiner (Eds.) Community Participation and Geographic Information Systems, Taylor and Francis,
London, 101-112.
MHLG. (1969). People and Planning, Report of the Committee on Public Participation in Planning
(The Skeffington Report), Ministry of Housing and Local Government (MHLG), HMSO, Lon-
don.
Negroponte, N. (1973). The Architecture Machine, MIT Press, Cambridge, MA.
OED. (1974). The Concise Oxford Dictionary, Oxford University Press, Oxford, UK.
Schroeder, R., Huxor, A., and Smith, A. (2001). "ActiveWorlds: Geography and Social Interaction in
Virtual Reality," Futures, 33, 569-587.
Shiffer, M.J. (2001). "Spatial Multimedia for Planning Support," in R. Brail and R. Klosterman
(Eds.), Planning Support Systems: Integrating Geographic Information Systems, Models, and
Visualization Tools, ESRI Press, Redlands, CA, 361-385.
Smith, A. (2002) 30 Days in ActiveWorlds: "Community, Design and Terrorism in a Virtual World,"
in R. Shroeder (Ed.), The Social Life of Avatars: Presence and Interaction in Shared Virtual
Environments, Springer, Berlin, 77-89.
Thrift, N. and French, S. (2002). The Automatic Production of Space, Transactions of the Institute of
British Geographers, NS 27, 309-335.