ArticlePDF Available

Videogames: the new GIS?

January 1456

Authors:

Ifan D H Shepherd

Middlesex University

Iestyn Bleasdale-Shepherd

Videogames and GIS have more in common than might be expected. Indeed, it is suggested that videogame technology may not only be considered as a kind of GIS, but that in several important respects its world modelling capabilities out-perform those of most GIS. This chapter examines some of the key differences between videogames and GIS, explores a number of perhaps-surprising similarities between their technologies, and considers which ideas might profitably be borrowed from videogames to improve GIS functionality and usability.

Content uploaded by Ifan D H Shepherd

Content may be subject to copyright.

1

Videogames: the new GIS?

Ifan D H Shepherd

Middlesex University, The Burroughs, London, NW4 4BT, United Kingdom

Email: I.Shepherd@mdx.ac.uk

Iestyn D Bleasdale-Shepherd

Valve Corporation, Seattle, USA

Abstract

Videogames and GIS have more in common than might be expected. Indeed, it is

suggested that videogame technology may not only be considered as a kind of

GIS, but that in several important respects its world modelling capabilities out-

perform those of most GIS. This chapter examines some of the key differences

between videogames and GIS, explores a number of perhaps-surprising

similarities between their technologies, and considers which ideas might profitably

be borrowed from videogames to improve GIS functionality and usability.

Keywords: Videogames, computer games, geographical information systems,

virtual geographical environments, virtual environments, virtual worlds, spatial data

visualization, user interaction.

“Comparisons are odious.” (Gilbert of Hay, 1456)

I. INTRODUCTION

The positioning of one technology or medium relative to others is a common academic

pastime. For example, numerous studies have compared GIS with other spatial and

non-spatial digital technologies, including: remote sensing [1; 2], mapping [3; 2],

databases [4], and CAD [4; 5]. For their part, videogames have been compared and

contrasted with novels, drama, comics and film [6]. Until comparatively recently,

however, the subject matter and technologies of most videogames have not been

similar enough to warrant detailed comparison with GIS. However, this is no longer the

case. Not only do many games present vibrant and realistic worlds on screen (Figure

1), but the spatiality of games is increasingly recognised as being a key property of

these distinctive digital environments [7; 8

Not every kind of videogame is germane to our analysis. There are numerous genres

and sub-genres of this form of interactive multimedia [

]. For this reason, a detailed head-to-head

comparison with GIS is long overdue. In undertaking such an analysis, we advance

two broad propositions: that the two technologies and their uses are far more similar

than might be expected; and that GIS have a great deal to learn from videogame

technology and design. Although the central focus of this chapter is on videogames

and what they have to offer, they will also be used as a lens through which to better

understand GIS, and to identify areas for further development.

9; 10], but the main focus here

is on games which have recognisable geographical content. Among these are most

action games (and especially first-person and third-person shooters), many role-

2

playing and adventure games, several kinds of simulation game (including driving and

racing games), and strategy games which involve the evolution of settlements, cities

and civilisations. There are many videogames with a less recognisable geographical

content, including: puzzle games (especially board games, quizzes and classic titles

such as PacMan and Tetris), music games (such as Guitar Hero and Jam Sessions),

and many children’s games. Although a large number of sports games (e.g. tennis,

soccer, snooker, American football, and Olympic events) have little intrinsic

geographical content, others (such as golf and car racing games) are of considerable

interest to GIS, because of their often fastidious and innovative approaches to micro-

terrain modelling and townscape representation. These, and other less geographically

relevant videogames, will be discussed where aspects of their design, role and usage

are relevant in a GIS context.

Figure 1. Realistically rendered fictional landscape of first-person shooter game

(Half-Life 2; courtesy of Valve Software).

Although a broad view is taken of what constitutes a videogame, it was decided to

exclude virtual world software whose primary goal is to enable users (typically in

online environments) to act out alternative lives in relation to one another, albeit in

specially constructed spatial environments. (These include virtual social worlds such

as Second Life and Habbo.) Although some of the functionality of such virtual

environments is also found in multi-player role-playing games, such as World Of

Warcraft, the primary objective of these sites is social networking rather than game

playing. It is, however, recognised that this distinction is not watertight, and that at the

time of writing some social networking sites are beginning to incorporate games-like

activities as a means of attracting and retaining subscribers. It should also be noted

that we use the term GIS throughout as a shorthand for the relatively close-knit family

of software that enables analysts to manage, display, analyse and report spatial

3

information in an integrated environment. This term therefore should be understood to

include full-service GIS, desktop mapping software, spatial data visualization

programs, and related software. The focus will be largely on software that models

aspects of the real world for users driven by relatively serious analytical intent.

Throughout this chapter, the acronyms VG and GIS will be used respectively to refer

to videogames and geographical information systems. Assiduous readers will note

that while numerous videogame titles are mentioned in the text, few GIS are identified

by name. This is partly because of the videogame focus of this chapter, but it also

reflects the fact that there are relatively few GIS and, unlike individual games, most

full-featured GIS contain a relatively wide range of functionalities characteristic of this

kind of spatial analysis and display software. In contrast, there are thousands of

individual VGs, and more than a dozen VG genres, and there is consequently a far

greater diversity between individual products than among GIS. Where distinctive

features are found in specific GIS, individual software will be identified by name.

II. SIMILARITIES BETWEEN VIDEOGAMES AND GIS

A common definition of GIS is that it is a system (part hardware, part software, and

part human) that is used to solve problems with a spatial dimension, by providing tools

for the acquisition, storage, integration, update, retrieval, manipulation, analysis and

display of spatial information. In this section, key GIS characteristics and

functionalities will be critically examined in relation to VGs, with the following three

aims in mind: to evaluate how far VGs provide the capabilities of GIS; to determine

limitations in the functionality of VGs in terms of spatial problem solving; and to

indicate ways in which GIS might benefit from adopting VG functionality.

A. Spatial representation

Videogames may be classified on the basis of two aspects of their spatiality: the

extent to which they attempt to model the real world, and the kind of representational

system they use in portraying elements within their worlds. (In terms of the second of

these dimensions, only visual representation is considered here, for reasons that will

become clear in section IIIC.) For the purposes of the current discussion, these two

aspects will be referred to as reality and realism respectively. Figure 2 maps out this

reality:realism space, and indicates where three related spatial technologies (VGs,

GIS, and social networking virtual worlds) are positioned in relation to these twin

characteristics. It is suggested that VGs occupy a far greater territory in this space

than either GIS or social networking virtual worlds. Within the area of the diagram

occupied by VGs, a selection of videogames has been located. (For readers less

familiar with contemporary VGs, a lookup table is provided in the Notes.) These

locations are best thought of as modal locations, because a number of VGs present

several contrasting virtual environments to the player.

Visual realism

The rich visual realism which many contemporary VGs present in their virtual worlds is

immediately apparent to any spatial analyst who looks over their children’s shoulders.

Landscapes extend across extensive areas, terrain is draped with plausible vegetation

patterns and the infrastructure of human habitation, building interiors are designed

with considerable attention to architectural and engineering detail, and many of these

4

virtual environments are enlivened with the dynamic trappings of modern life, including

vehicles and people in motion. Some of this visual realism is undeniably a showcase

for VG talent and technology, and especially for graphic artists and graphics hardware.

However, a great deal of it reflects the way in which many modern videogames place

the player at ground level, and therefore represent the world as seen through the eyes

of the pedestrian or the driver, rather than from the air as in most GIS. However, much

of the visual realism is wrapped around a world model that is largely or entirely

fictional.

Figure 2. The reality-realism space occupied by videogames, GIS and social

networking virtual worlds, with acronyms of selected videogames

from the text.

Representing the real world

The environments encountered by players in VGs run the gamut, from authentic

reproduction of real places using real spatial data to entirely fictional places using

synthetic data. The majority of VG environments, and especially those built for role-

playing games, are imagined worlds, created from the imagination of the games

designer. At this lower end of the reality spectrum (as depicted by the vertical axis in

Figure 2), are to be found innumerable fantasy worlds (e.g. Final Fantasy, Jak &

Daxter, the God Of War series and Little Big Planet) and several impossible ones (e.g.

5

Echochrome, with its Escher-like 3D geometry, and Braid, with its distortions of time).

There are also some casual and passive games which visualise user data to create an

abstract 3D scene (e.g. Packet Garden, which uses 3D glyphs to construct virtual

landscapes from multivariate data describing the player’s online surfing habits).

Also at a low level on the reality scale are worlds which may appear to be based on

some form of geography or cosmology (as in role-playing games such as Fable II), but

where the reality is only skin deep. The spherical planetoids in Super Mario Galaxy,

for example, bear little or no resemblance to real world exemplars; and the 3D globes

that may be explored in Civilization: Revolution, Katamari Damacy, Spore, and

Ratchet & Clank: Going Commando lack either the dimensions or the realism of the

earth as seen from space. However, numerous VGs at this relatively low point on the

reality scale attempt to mimic the visual appearance of the real world. These include a

number of first-person games (e.g. Crysis), and many racing games (e.g. Burnout:

Paradise), while many spatial strategy games (e.g. various versions of Civilization,

Sim City and The Settlers) also provide recognisably human landscapes and elements

of real-world behaviour. In creating the level of visual fidelity evident in these games,

designers are increasingly aided by software that procedurally generates terrain and

buildings, often using architectural templates taken from real cities.

Many games with this reality:realism mix are set in historic locations, but designers

typically model the environments with a mix of artistic licence and architectural and

geographical authenticity. Examples include: Assassin’s Creed (set in the cities of

Acre, Damascus and Jerusalem at the time of the Third Crusade in the late 12th

century), and Hellgate: London (set in a futuristic version of London). Although most

videogames based on past wars (the military shooter genre) appear to recreate the

look and feel of a battlefield, they tend to use a particular conflict merely as an

‘inspiration’ or hook; the gameplay itself often bears only passing resemblance to the

battles fought. Examples include Medal of Honor: Allied Assault, based loosely on

World War II, Call of Duty 4: Modern Warfare, based on modern battlezones, and the

squad shooter Conflict: Desert Storm, based loosely on Desert Storm. Some titles

(e.g. Rainbow Six Vegas 2 and Call of Duty 4: Modern Warfare) take considerable

pains to visualise weaponry, vehicles, soldiers clothing and insignia in true detail, but

their representation of geography and historical events is much less authentic. Among

VGs with a greater purchase on reality are Call of Duty: World at War, which aims for

both geographical and historical authenticity (including the use of audio recordings of

World War II weaponry), and Brothers in Arms: Hell’s Highway, which not only models

actual theatres of conflict (the Dutch countryside during in World War II), but also

enables the player to participate in reasonably authentic forms of fighting. But for

every title that attempts to be true to the battle that inspired the game, there are many

others which make only a token effort to replicate the geographical environment of the

war in which their virtual fighting is staged.

Much the same mixture of relatively low reality and high visual realism may be found

in several VGs set in contemporary locations. In some examples (e.g. Fracture, set in

the San Fransisco Bay Area), the cityscape is ‘re-imagined’, using a mixture of familiar

and unfamiliar elements. In other games, locales are described as ‘inspirations’ rather

than replicas. Examples include the skateboard game Proving Ground, which is based

loosely on Philadelphia, Baltimore and Washington DC, and the car racer Project

Gotham Racing, which has been set variously -- and in most cases somewhat

6

symbolically -- in New York, London, Tokyo, Las Vegas, Macau, Shanghai, St

Petersburg and Quebec.

However, some VGs are more closely based on real world locations, altering the real

geography only where it is dictated by the needs of the gameplay. An example of this

approach is the car racer Getaway, whose designers took considerable efforts in

collecting detailed streetscape information (including building textures captured

through extensive field photography). Although the street layout is simplified and

stylised, it bears more than a passing resemblance to the major thoroughfares in

those parts of central London in which the game is set. Similarly, Grand Theft Auto IV

involved considerable research into New York, including the taking of tens of

thousands of photographs [11]. (See [12] for further examples of VGs set in replicas of

real cities.) This mixture of real-world and synthetic data is also found in largely

fictional VGs (e.g. WRC: Rally Evolved and Forza Motorsport 2), which use real-world

data in the same way that some GIS worlds incorporate imaginary (or at least

synthetically derived) data [13; 14].

Figure 3. World strategy game showing buffers around high latitude radar

stations on cylindrical map projection (Defcon; courtesy of

Introversion).

There is only a small number of VGs whose virtual environments are based on real

geography and real spatial data, and therefore occupy the upper end of the vertical

axis in Figure 2. One example is the car racer Test Drive Unlimited, which models the

terrain and route network of the Hawaiian island of Oahu by integrating multiple

sources of geospatial data for the island. (This permits the display of a realistic in-car

satnav.) A second example is Microsoft Flight Simulator, which drapes aerial photo

7

data onto digital terrain models, and uses accurate 3D digital models of actual

airplanes and detailed airfield and airport layouts, to provide a realistic simulation of

flying that unfolds in real time. (It should be noted that much of the functionality of

these two games, including the virtual satnav, is available in Google Earth via user

mashups [15].) A third example of a game based o real world geography is the multi-

player world strategy game, Defcon, which charts player progress on a cylindrically

projected world map (Figure 3).

Although these three examples are located towards the top of the vertical reality scale

in Figure 2, they are spread across the horizontal realism scale. In terms of its real-

world visual appearance, Microsoft Flight Simulator is closest to the naturalistic end of

the horizontal scale, although for most of the time the player has a relatively high-

altitude bird’s eye view. A little way to the left of this is Test Drive Unlimited, whose

visual representation of island scenery is frequently rather stylised. Finally, Defon is

over at the symbolic end of the horizontal scale, with gameplay being organised

around an atlas-like world map. It should be noted, however, that VGs evolve through

time, and many titles shift their position and/or modify the area of the reality:realism

space they occupy as new versions appear on the market. In this context, it is

interesting to note how Google Maps, originally developed to represent the real world,

is increasingly being used to display virtual worlds (e.g. GTA IV’s Liberty City at

grandtheftauto.ign.com/maps/1/Liberty-City-Map and Second Life at: slurl.com).

Fitness for purpose

For VGs and GISs alike, the level of ground truth provided by the software is only

useful to the extent that the player or the spatial analyst requires it. Indeed, the

reaction from some quarters of the gaming community to increased realism in VGs

has been extremely negative. One commentator [16] rails against the obsessive

pursuit of realism by some games producers, suggesting that the essence of some

classic games (such as Sonic the Hedgehog or Pac-Man) would be destroyed by

using realistic rendering and physics. For Biffo, “one of the big appeals of games is

that they are a window on to fantasy … in the more general sense of being

transported to somewhere that isn’t real”. In a statement that serves to maintain clear

water between videogames and GIS technology, he goes on to argue that: “If you’re

going to create a CGI world that’s only one step away from our real world, then why

not just make a live-action movie?” -- or, one might add, use a GIS? What GIS

analysts know only too well is that unrealism, in the form of map symbols and

visualization glyphs, can often serve to deliver greater understanding than

photorealistic displays.

A considerable tension has been introduced into VGs by the desire to create truthful

representations of the real world while at the same time designing virtual

environments which provide a compelling game experience. This tension is central to

the game designer’s craft, as is evidenced by the critical view of a highly influential

games designer [17]:

“A game creates a subjective and deliberately simplified representation of

emotional reality. A game is not an objectively accurate representation of

reality; objective accuracy is only necessary to the extent required to support

the player's fantasy. The player's fantasy is the key agent in making the game

real.”

8

Using modern VG technology it is now just as easy to model the geography of an

actual racing circuit or a battlefield as it is to recreate an actual race or battle.

However, the faithful recreation of real events runs the risk of suppressing the freedom

associated with player decision-making behaviour that is the central feature of

interactive games. Instead of interaction, the player would confront a digital version of

the kind of event reported on a sports TV channel or staged by a battle recreation

society. As one war simulation game reviewer put it [18

The foregoing discussion reveals as illusory the presumption that VG environments

are always synthetic while GIS worlds are always real. Indeed, GIS have until very

recently provided low-fidelity visual realism (in the form of cartographic symbolism),

minimal fidelity in terms of non-visual data representation [

], the challenge for developers

is to tread the fine line between testament and entertainment. This important dilemma

is largely absent in GIS, whose primary goal is the representation of geographical

reality.

19; 20], and only medium

fidelity in terms of dynamic modelling (a reflection of the static map inheritance of

GIS). Moreover, the central problem of spatial data uncertainty further underscores the

fact that even though GIS attempt to model the real world as accurately as possible,

their representations are frequently significantly flawed [21; 22

Various algorithms are also similar in VGs and GIS. These include: spatial search

(using sophisticated spatial indexing), object buffering (used for object collision

detection), and path planning algorithms (used to transport the user from one location

to another). (The artificial intelligence developed in VGs for automatic camera control

in third-person games is perhaps unique to VGs, and will be discussed further below.)

]. It is therefore a crude

over-simplification to suggest that while videogames have been high on realism but

low on reality, GIS have been low on realism but high on reality. Each in its own way

adopts blends of the real and the imagined, the naturalistic and the symbolic, as

needs dictate.

Spatial modelling technologies

The technologies used by VGs and GIS in building their respective digital worlds are

increasingly common to both domains, though with perhaps different emphases.

Vector data models are widely used, though in VGs they are almost exclusively

applied to 2.5D and 3D objects. Digital terrain models are also commonplace, with

both VGs and GIS using multiple resolution terrain data. The grid-cell data model is

largely absent from VGs, though a variation of it appears in the grid-based terrain

models used in a number of strategy games (e.g. Sim City). On the other hand, image

and texture data are widely used, and recent innovations in texture mapping from the

VG and CGI communities are now routinely used by GIS to drape aerial photos and

satellite images over terrain models. Additional realism is provided in VGs through

surface rendering that uses a range of sophisticated computer graphics techniques,

including: bump mapping, environment mapping, radiosity, dynamic lighting, and dirt

mapping (which is put to good effect in the car rally game, Colin McRae: DIRT). Other

technologies used in both camps include: the use of hierarchical data structures (and

especially quadtrees), spatial data integration (especially of various kinds of surface

data), techniques which minimise processor loads in complex scenes (e.g. tiling and

level of detail or LOD), and layering (which in VGs is mainly used to model dynamic

accretions on surfaces).

9

However, this is where VGs and GIS begin to part company. For their part, VGs have

focused more on innovative and flexible methods of user interaction and navigation in

3D environments, and in developing dynamic models of environmental and human

processes. Several specialist data structures are used to support fast surface

rendering (e.g. indexed vertex arrays), dynamic lighting (e.g. normal maps), rapid

search (e.g. BSP trees and nested voxel grids) and navigation (e.g. traversability

graphs) in 3D environments, and related structures for rapid collision detection (e.g.

convex hulls). In contrast, GIS have placed more emphasis on developing tools for

spatial analysis, and perhaps most notably in terms of 2D spatial overlay. In general,

VGs tend to adopt spatial models and data structures that ensure optimal real-time

rendering and interaction, realistic sound and haptic responses, and continuous data

updating during gameplay. (In Splinter Cell: Conviction, for example, the dynamic

navigation mesh which enables realistic navigation behaviour is updated as objects in

the world move or change.) In contrast, GIS have tended to adopt spatial models and

data structures which are primarily optimised to support analytical operations on data

that represent relatively static worlds.

B. Attribute data

One of the more distinctive features of GIS technology is its integration of attribute and

locational data [23; 24

In both VGs and GIS, attribute data are crucial, not only for data visualization and

object selection, but also for more advanced forms of simulation, spatial analysis and

reporting. Indeed, there is frequently little difference between VGs and GISs in either

the type or quantity of non-graphic data they require. (For example, some 80,000 lines

of spoken dialogue were specially written for Grand Theft Auto IV, which used 740

different voices.) However, one significant feature of VGs is that a great deal of game

information is not attached to objects as standardised descriptive information. Unlike

GIS, much of the information stored in VGs supports elements of gameplay not

normally found in GIS. Thus, for example, the text in GTA IV are delivered in an

intelligent fashion based on in-game events and player actions. Half-Life 2: Episode 1,

for example, uses a fairly generic event-driven contextual system in which dialogue is

situation, event, and action specific. This makes the world and its characters react to

the player in a (seemingly) more intelligent fashion. Some lines of dialogue are always

]. On the surface, VGs appear to be all visual representation

and action, with little obvious evidence of the attribute data associated with real-world

features and objects. This perception, however, is misleading. Although VG software

may run to millions of lines of code, most games also contain very large volumes of

‘content’ or ‘assets’. Like GIS, VGs also store large amounts of attribute data,

describing elements of the virtual environment, including discrete objects and

characters. In a typical contemporary videogame, object assets may include: vertex

data (multiple attributes -- e.g. position, normal(s), texcoord(s)), texture data (multiple

layers -- e.g. albedo, normal, reflection mask), collision mesh, (multiple versions of the

prior data for LODs), animations and related audio files and particle effects, and

physical data about the object’s surface and volume. Although such data may not be

organised in the form of a relational database, for reasons of simplicity and

performance, many data tables adopt the simple row-column form of the standard

attribute data matrix. However, because of the highly dynamic nature of VG

environments, attribute data in VGs are more frequently stored in other data

structures, such as trees, sparse voxel grids and hash tables.

10

heard – usually associated with events which are guaranteed to occur – but some will

only be heard if certain things happen.

A second difference resides in the uses that can be made of attribute data. In contrast

to GIS, much VG data are only accessible to the games software, and not directly to

the player. In general, GIS analysts actively extract data, while in videogame players

are passively fed data. In both cases, algorithms are at work, and user decision

making and interaction is required, but the degree of control of data-extraction and

display is skewed towards analysts in GIS, and to the software in VGs. Thus, while in

VGs data are mainly ‘pushed’ to the player, in GIS, data are mainly ‘pulled’ by the

analyst.

However, there are some notable exceptions to this general rule. Titles across several

game genres (e.g. role-playing adventure games, strategy games, racing games and

sports games) not only maintain large attribute databases related to their virtual

environments, objects and characters (and particularly for player avatars), but they

also display some of these data on screen during gameplay, and update attribute data

for the player’s avatar and its behaviour as the game proceeds. Stored attribute data

is displayed in various ways during gameplay. In some car racing games (e.g. Forza

Motorsport 2), detailed readouts of car settings and behaviour can be displayed, and

the player is able to adjust these during the game to modify the driving behaviour of

their vehicle. (In Sega Splash! Golf, for example, golfer statistics can be viewed in a

player room.) Such data may provide players with feedback and motivation (as in

many car racing and role-playing adventure games), or may aid dynamic decision-

making (as in many strategy games).

In many adventure, strategy and simulation games, relevant text is provided on screen

as a readout or mouseover display (akin to Windows tooltips), while many shooters

provide head-up displays with summary information about the current state of the

game. In Civilization: Revolution, visual modifiers (brief displays of summary

information) are attached to objects on screen, while a street-name indicator appears

at the top of the Burnout Paradise screen, reminiscent of GIS practice. Another genre

of VG which routinely accumulates in-play data is the massively multi-player online

game (MMOG). Such data are made available to game community websites which

feed game statistics back to players so they can learn from others’ gameplay

performance. (A notable example is Bungie.net, which reports on online, multiplayer

deathmatches in Halo.) These data are also used by game designers to help steer

new feature development, and plan game marketing activities. There are lessons to be

learnt here by GIS, where the actual problem-solving behaviour of expert analysts

might be of considerable value to less experienced users, or peers working in the

same field. In addition, the accumulation of data on trainee actions during training

courses could be used to make informed decisions on further learning activities.

C. Spatial data visualization

Few (if any) VGs adopt the kind of symbolic map representation that has benefited

spatial data analysts for the best part of two hundred years, and are thus able to

provide little guidance on effective 2D data visualization. (One of the few exceptions is

Defcon, the multi-player world strategy game, which displays its symbolism on a

cylindrically projected world map -- see Figure 3.) However, because most

contemporary VGs display realistically modelled scenes in 4D, they provide a suitable

11

benchmark against which data visualization in virtual worlds modelled by GIS can be

measured. Although few VGs would appear to adopt the strict rule-based approach to

data visualization propounded by Bertin [25; 26; 27; 28; 29

Unlike most GIS, VGs also make use of time as a visual variable. A common example

is found in shooter games, where coloured bars of dynamically varying length

represent the gradual decline or rapid replenishment of a player’s health (Figure 4).

Another example is the pulsation of an object’s size or colour, where the frequency of

the variation indicates the current threat level -- e.g. when an object is likely to explode

], it is nevertheless

instructive to understand how VGs adopt the graphical sign system, albeit at an

informal level, and to identify innovative applications in VGs that might be useful in a

GIS context.

Perhaps the simplest example of the operation of data visualization rules is the

use of categorical visual variables such as shape, colour and texture to

distinguish between qualitatively different groups of characters, objects or events.

Most videogames pay careful attention to the visual design of entities so as to

allow at-a-glance disambiguation between the many entity types present. The

use of color-coding to relate groups of entities is a very common practice in

videogame design. In the team-based online game Team Fortress 2, for

example, players on opposing teams are differentiated through the use of simple

colour schemes (e.g. one team wears red clothing, the other blue), while in

Schizoid, enemies of a particular kind are given distinct colours (see

Figure 8). Numerous combat games also distinguish friends from enemies by the

style of clothing they adopt, many role-playing games (RPGs) distinguish male

and female characters b y their clothing styles, and in some shooter games the

lethality level of monsters is identifiable from their 3D shapes. In Black & White II,

the avatar changes its form based on behaviour, so that it in turn looks good or

evil, fat or skinny, weak or strong. Colour is also widely used for signalling

purposes. In Mirror’s Edge, for example, features of the environment (such as

ledges and machinery) that might be useful to the player in traversing the

cityscape are depicted in primary colours, so they stand out from the mainly white

and blue of the virtual world, and while safe paths are clearly illuminated,

dangerous paths are shown in shadow. In the Half-Life series, interactive objects

are coloured red, and lighting is commonly used to attract attention to important

objects or locations.

Videogames also use gradational variations in visual variables to reflect continuous

variations in numerical variables (e.g. counts or measurements). Normally, this applies

to individual objects or characters in the game. Size, for example, is often related to

the strength of enemies or the value of items in a game (as in Blinx: The Time

Sweeper, in which gold ingot size implies value), and variations in lightness or hues

along a colour ramp is sometimes used to convey some internal state value (e.g.

brightness of a head-up display (HUD) element relates to the health of the player’s

avatar, the health of an enemy is reflected in a ‘health bar’, typically floating in the

world above the enemy’s head, which changes in colour from green (healthy) to yellow

(damaged) to red (nearly destroyed)). Games which use a perspective projection also

provide implicit scalar visual feedback, by relating an object's size on screen to its

distance from the viewpoint (in addition to the obvious relation between on-screen

position and virtual 3D location).

12

(e.g. in Left4Dead, pipe bombs beep with increasing frequency as they near

detonation, and the player’s view becomes increasingly dark and blurred as their

health runs out when they are in need of rescue by another player).

Figure 4. Urban night-time scene, showing dynamic health indicator bars

(Left4Dead; courtesy of Valve Software)

Several VG practices may be of interest for data visualization in a geographical

context. Some games eschew the normal per-symbol approach to data visualization,

and adopt instead what might be termed ambient visualization. In Brothers in Arms:

Hell’s Highway, for example, a colour intensity gradient is used to darken and redden

the entire screen as the risk of the player being shot increases. A variation of this

technique is used in Half-Life 2, where threats to the player from the side or rear are

indicated by the appearance of edge-of-screen red bars. With the left bar appearing if

the threat is from the left, the right bar if from the right, and both bars if from directly

behind the player? Also, the redness of these bars is proportional to damage, and

fades over time after the damage occurs. In the case of repeated damage from a

given direction, the colour will maintain a strong intensity due to the slow fade.

Another distinctive use of data visualization is to provide on-screen visual prompts

derived from several relevant variables to assist the player in making spatial decisions,

or undertaking complex actions. Examples include: the 3D trajectory line symbol in

Tiger Woods PGA Tour 08; to assist the player in making a shot; and the

quarterback’s ‘vision’ cone in Madden NFL 2006, whose length and width is

proportional to that player’s level of expertise, is used to determine which receivers

can be thrown the ball. The location of certain visual prompts in the virtual

13

environment can often be of considerable importance to effective decision making by

the player. A good example is the visual feedback provided by the balance bars

included in early editions of the Tony Hawks skateboard games. Because these bars

were placed next to the skater, it was found that found that they competed for the

player’s visual attention. Consequently, they were replaced in Proving Ground by a

screen-wide coloured arc, which shifted the visual prompt to the player’s peripheral

vision. (Similar thinking was used in the design of the edge-of-screen coloured bars in

Half-Life 2, as described earlier.)

A further lesson for spatial data visualization concerns how well 2D visual symbol

coding translates to 3D. Although the approaches explored by Kraak [30; 31] do

not appear to have been implemented in VGs, careful graphical design has

established some working principles for effective 3D visualization. An example is

found in the team-based, multi-player game Team Fortress 2, where the player

takes on one of 9 distinct roles. Because it is important to be able instantly to

identify another player’s team and role, the avatar for each role has a unique

silhouette, easily distinguishable from all the others, which retains its

distinctiveness when viewed from various directions and under varying lighting

conditions. A related technique is used to show objects even when occluded, as

in the ‘shadow Mario’ used in Mario Sunshine/Galaxy, whose silhouette at least is

always visible, even when the character is occluded. In general, automatic

cameras and environmental design are concerned with mapping 3D worlds to

effective 2D views at any given moment.

Figure 5. Atmospheric evening scene involving darkness and local lighting

(Left4Dead; courtesy of Valve Software).

14

A final lesson for GIS concerns the way in which data visualization can often serve

multiple roles in VGs. It is frequently the case that visual symbols not only represent

information, but at the same time add emotional depth, invoke engagement and

develop a sense of immersion for the player (see Figure 5). This suggests potentially

fruitful lines of collaborative research between VG and GIS, involving such topics as

denotative and connotative responses to data representations, the semiotic

boundaries between monosemy and polysemy, and possible interactions between

cognitive and affective responses.

III. WHAT VIDEOGAMES CAN CONTRIBUTE TO GIS

In this section, we consider areas in which VGs have surpassed GIS in their treatment

of elements which may be relevant to spatial data explorers and analysts. Three areas

will be discussed in which VGs can contribute in a significant way to GIS: dynamic

process modelling, user interaction, and multi-sensory data representation. Space

prevents discussion of several other significant aspects of VGs that may be relevant to

GIS.

A. Dynamic process modelling

Videogames are change-rich environments. From their earliest appearance,

videogames have provided dynamic experiences, with events continually unfolding on

screen. Changes to the virtual environment are in part the result of dynamic

simulations running within the software. However, they also result from user

interactions with the virtual environment, including shifts in player position and

viewpoint, but also user modifications to the virtual environment. Indeed, change tends

to be continuous rather than episodic in most VGs, and permeates most aspects of

the virtual environment, including parts that may currently be unseen by the player.

The management of change therefore represents one of the key challenges to the

games designer and player alike, and these challenges constitute what Poole [32

In earlier games, the impact of change within the virtual environment was often

temporary, with killed creatures fading away shortly after their death, and bullet holes

disappearing or being generalised on vehicle bodies and building facades soon after

they had been created. Nowadays, however, bodies will accumulate on the ground,

]

refers to as the kinetic narratives which are such a distinctive feature of videogames.

Not all change evident on screen is significant to the goal or purpose of the game; a

great deal of dynamic behaviour is modelled for purposes of ambient realism. Thus,

for example, offices contain working computers and fax machines, homes have TVs

which play ads, vehicles have satnavs that emit spoken instructions, and in-car radios

emit music and DJ banter. Some games mimic real environments by providing usable

telephone booths in the street (as in Proving Ground) or mini-games that can be

played in shopping mall arcades (as in Sega Splash! Golf). This dynamic modelling is

taken to new levels in representing change initiated by player actions. Thus, for

example, footprints appear when walking on wet sand; spray appears on a camera

lens during a powerboat trip; newspapers flutter in the vortex created by a passing car;

and plaster flies off a wall when hit by bullets. Some of this dynamic realism is more

significant in terms of gameplay, as when a box breaks apart when hit by a crowbar, to

reveal hidden ammunition.

15

and bullet holes will stay exactly where they first appeared. VGs thus not only

represent events; they also record the accumulated effects of events, so that the

evolving game environment becomes a dynamic palimpsest. What is often missing,

taking a GIS perspective, are facilities to interrogate and analyse this change.

However, in Sim City 4, the player can look at pollution or property-price colour maps

as the city evolves, and in De Blob, the player progressively covers more and more of

a 3D environment with various colours of paint.

For their part, GIS initially drew their modelling approaches from static 2D maps, with

digital spatial models typically representing a synchronic snapshot -- the classic cross-

sectional view. Over the past couple of decades, however, GIS technology has moved

progressively from mimicking static 2D maps to providing digital representations of 4D

data [33; 34]. Only comparatively recently have GIS begun to change, with the

introduction of 3D viewing, brushing, and animation [35; 36]. Increasingly important is

the visual representation of diachronic change, typically undertaken by creating

sequences of temporally spaced snapshots (e.g. the spread of population across the

UK from decennial census data; the diffusion across America of WalMart stores; or the

tracks of hurricanes in the Caribbean). Objects have also begun to move and change

on screen by coupling the software to real-time data feeds (e.g. from GPS or ground-

based tracking technologies) or links to simulation model output.

Although most GIS can handle historical data, typically in the form of time-spaced

snapshots (e.g. remotely sensed image sequences for a given area, or decennial

population censuses for a particular country), the essential characteristic of dynamic

representation in GIS is that it is mainly a data-driven process, where the data are the

outcome of prior surveying or monitoring activity, or as a result of real-time data

capture. Most do not display the output of dynamic process models. Thus, while most

GIS are adept at handling spatio-temporal data, few model dynamic spatial processes.

This is not to ignore the considerable success within the geoscientific community, in

developing dynamic environmental and social models. These range from the

geological and hydrological models of the Kansas Geological Survey during the 1970s

[37], through to the more recent geosimulation models of the 1990s [38]. There have

been numerous attempts since the 1990s to link GIS to environmental process models

(e.g. [39; 40; 41

Videogames, unlike GIS, do not tend to store a great deal of time-varying data, other

than music or recorded sounds. (Exceptions include character or object animation,

which is often canned rather than simulated in real-time, which is discussed further

below). Unlike GIS, most modern videogames, across several genres, embody

numerous dynamic process models which not only enable various elements of the

virtual environment to appear to behave realistically, but also permit the environment

to respond in realistic ways to player interactions. Many car racing games (e.g. Forza

Motorsport 2) combine a high degree of visual realism with a high degree of process

realism. These simulation-heavy not only replicate circuits with considerable

geographical and architectural fidelity, and highly realistic vehicles with every detail of

the bodywork precisely modelled, but the behaviour of the vehicles is accurately

replicated in the player’s driving experience. For example, specific vehicular sounds

are replayed during the game, and dynamic vehicle handling characteristics are also

]) mainly through loose and tight coupling rather than embedding.

Nevertheless, these models still occupy a parallel technological space to mainstream

GIS, especially in the geosciences.

16

faithfully modelled based on a wide range of actual performance variables, to give the

player the experience of real vehicle driving. Another example of an embedded

process model is found in Hydrophobia, in which a fluid dynamics model,

encapsulated in a specialised software engine, manages the dynamic behaviour of

water bodies and its effect on objects. This model not only handles realistic water

movement in complex spaces, and modifies the acoustics of the enclosed spaces it

occupies, but it also models interactions between the water and the game

protagonists.

The difference between VGs and GIS in this respect is crucial. While the outcome of a

game’s dynamic processes may appear to be realistic, and indeed may be generated

by physically accurate models, they are synthetic outcomes, and will typically vary

each time they are produced, usually in context-dependent ways. GIS also display the

outcomes of real-world processes but, in contrast to VG outputs, this realism derives

from the fact that what is being modelled usually consists of recorded occurrences.

Here we find a contrast not only between the synthetic and the real, but also between

the what-might-be and the what-was, between the possible and the actual, and

between the fictional and the factual.

Dynamic realism in VGs is closely related to notions of real-world time, both

chronological and perceptual. The representation of cyclic time, and changes

associated with it, is relatively common. Several games track the circadian cycle, for

example by moving the sun across the sky (e.g. the Jak & Daxter series), by revealing

changes in urban air pollution (as in Midnight Club: Los Angeles), or by modelling

inhabitants’ daily routines (e.g. retiring to bed at night in The Elder Scrolls III:

Oblivion). Others unfold their story through the seasons of the year (e.g. Animal

Crossing and Harvest Moon), with some (e.g. the Harvest Moon series) modelling

vegetation, crop and landscape changes through the seasons, while some games

(e.g. Grand Theft Auto 3 and The Elder Scrolls III: Morrowmind) also introduce varied

weather during gameplay. Several epic adventure and role-playing games (e.g. Day

Of The Tentacle and Eternal Darkness: Sanity’s Requiem) simulate change through

far longer periods of time.

Process distortion

Not all games behave strictly in tune with chronological time or Euclidean space. For

example, some games set in extensive virtual worlds (e.g. The Elder Scrolls III:

Morrowmind and Legend of Zelda: Ocarina of Time) permit players to traverse long-

distances between widely separated locations almost instantly, while others permit

time to be frozen (e.g. Timeshift and Blinx: The Time Sweeper), sped up (e.g. Frontier:

First Encounters and Sim City 4) or slowed down (e.g. Proving Ground, Burnout and

Max Payne). (Braid, being a game entirely based around time manipulation, allows

you to do all of these things and more.) In Mirror’s Edge, the ‘Runner Vision’ tool slows

down time during gameplay to assist the player in executing complex moves, and is

similar in visual style to ‘bullet time’ adopted in Matrix-inspired games. In a GIS

context, maps have long been distortable through map projection or topological

restructuring (as in cartograms). As for temporal distortion, modern data visualization

software permits analysts to vary the speed and duration of events in spatial

animations and in dynamic brushing, while spatial simulation software has always

enabled models to be run in other than real time. The key difference between VGs and

GISs, in this respect, lies in the purpose of temporal and spatial distortion. In the

17

former it enhances various elements of gameplay, while in the latter it supports

effective analytical interpretation.

Even when VGs provide an accurate simulation of the world, they sometimes permit

the player to select the degree of realism they wish to see operating. This is

particularly common in car racing games. Thus, for players who enjoy the fidelity of

precise simulation, vehicle handling realism might be set at a high level, while those

who prefer a more thrill-seeking driving experience might set the vehicle’s handling

system at a lower and far more forgiving level. Although visual, aural and behavioural

realism are the primary goals of this often costly modelling effort, other liberties are

also taken to enhance playability at the expense of temporal realism. Midnight Club:

Los Angeles, for example, permits its high-performance cars to be raced at speeds in

excess of 200mph around the streets of Los Angeles. Another interesting deviation

from behavioural realism occurs in those squad-based combat games (e.g. Brothers in

Arms: Hell’s Highway), in which the game storyline requires that if key squad

members are wounded in battle, they can be revived to rejoin the action. In a similar

fashion, the AI which directs non-player characters in numerous games genres may

be varied between a level of optimal decision-making efficiency (where the human

player has little chance of winning) to a much more flawed level where it makes

human-like errors (which gives the player some chance of winning). These and many

other examples suggest that while VGs may be increasingly adept at creating

accurate models of real-world processes, the fidelity of such modelling will always

need to serve the maximisation of player satisfaction, in terms of the balance to be

struck between challenge and reward.

Figure 6. Countryside scene showing farmhouse being torn apart by alien

creature (Half-Life 2; courtesy of Valve Software).

18

Dynamic world modification

Unlike GIS editing software, which is typically used to build and modify spatial data

models prior to analysis, most VGs facilitate the dynamic modification of their game

environments during gameplay, based on player actions and decisions. This requires

sophisticated software not only to model multi-way interactions between player,

environment and objects, but also to update spatial and attribute data models on the

fly. Dynamic landscape modification takes several forms. Buildings and other items of

infrastructure are blown up in many games (Figure 6), while in Fracture, players can

use various kinds of grenade and firearms to modify the land surface itself, in order to

gain tactical battle advantage [42

Many strategy and role-playing games involve large numbers of humans, with which

]. More constructive changes include the moving of

objects (e.g. Elebits), altering the look and feel of the existing environment by pulling

on the ground surface to change its shape (as in Civilization and Age of Wonders), by

planting or cutting grass (as in Viva Pinata), or by extending the environment by

adding land and infrastructure (as in SimCity). The virtual worlds of many videogames

are becoming almost as malleable as the real world.

Terraforming, deformation, destruction and construction are not restricted to

landscape elements. In car racing and shooter games, for example, it is common for

discrete objects such as opponents, furniture, vehicles and roadside objects (such as

lampposts or fire hydrants) to be damaged or destroyed. (The earlier but innovative

Red Faction game had a geomod, or geographical modification, feature which enabled

players to incrementally destroy infrastructure, while the more recent Crysis and

Mercenaries 2: World in Flames have more extreme destruction capabilities.) Some

games also permit damage to be undone. In Burnout Paradise, for example, drivers of

damaged vehicles may visit a virtual repair shop to have them instantly repaired, while

in Grand Theft Auto, layers of dirt accumulated on vehicles may be removed by

passing through a carwash.

Human behaviour and social processes

Human behaviour is also simulated with increasing realism in VGs. In games which

involve competition between player and software, most titles incorporate some form of

artificial intelligence (AI) to operate non-player characters (e.g. monsters or zombies)

and competing drivers’ vehicles in a realistic fashion. In many shooter games, for

example, combatants may be given varying degrees of intelligence, which may be

selected by the player to match their skill level. In car racing games, competing cars

are also driven with increasingly sophisticated software. In at least one such game

(Forza Motorsport 2), the player may select several AI drivers to drive for them.

Squad-based behaviour is increasingly evident in combat games, and is increasingly

modelled using sophisticated AI. In Brothers in Arms: Hell’s Highway, and the World

War I title, To End All Wars, the squad behaviour mimics multi-level military decision-

making behaviour, down to squad and individual levels. Human behavioural modelling

is taken further in the tactical shooter BlackSite, in which a squad morale system is

used to influence the battlefield behaviour of individual members of the player’s team.

(The player indicates what squad members are expected to do, but the current squad

morale level affects how they do it.) Even more remarkable is the growing practice of

developing and communicating the emotions of game characters, as in the trait and

‘moodlet’ system implemented in The Sims 3.

19

the player can interact and/or control. At a relatively simple level, some games (e.g.

Assassin’s Creed) provide realistic simulation of crowd behaviour. At a more

sophisticated level, are the strategic games, often set in cities (e.g. Settlers: Rise of an

Empire), which have evolved through several iterations to present high-fidelity urban

environments in which citizens are seen acting out their daily lives in transparent

detail. AI is particularly important in such games (e.g. The Sims 3), which often involve

large numbers of inhabitants, because it would be virtually impossible for the player to

micro-manage the behaviour of all of the individuals inhabiting the game. Some RPGs

mimic society in other ways. The Elder Scrolls III: Morrowmind, for example, models a

character-class system, in which attributes and behaviours are established for game

characters (usually including the player) in a way that resonates with socially stratified

societies in the real world. Although some of these class systems are caricatures of

the real world, they do represent a form of social simulation that is likely to achieve

greater realism as in-game AI is further refined. At a higher social level, some games

attempt to model cultural differences, but this is currently done rather poorly. The

crude national stereotyping in such games as Battalion Wars 2, for example, fares

poorly when compared with the heavily researched national cultural classifications of

Hamden-Turner and Trompenaars [43], Hofstede [44] and others.

However, some commentators (e.g. [45

While the effectiveness of a GIS may largely be measured in terms of the power and

flexibility of its toolset, “a computer game, unlike most other computer applications,

lives or dies by the effectiveness of its user interface design” [8]. Because several

thousand games are published each year, and many introduce innovative interface

ideas, the games scene represents a huge experimental laboratory for interface

design. This provides a major resource that may be raided by GIS designers in

]) suggest that AI can become too clever for its

own good, in that it can make a game too difficult to play against. Dumbing down the

AI also has its limitations, however, because of the danger of mismatches appearing

between the visual realism of game characters and the relative stupidity of their

behaviour. As a result, in games such as Unreal Tournament 3 and Burnout, the

designers attempt to create AI that behaves in ways that humans would behave, and

makes the same mistakes that humans would make.

There are several lessons here for GIS. One might be in terms of monitoring the

behaviour of spatial analysts during and across sessions to provide intelligent context-

sensitive help (as in the case of Left4Dead described earlier). It is also worth noting

that since the AI would be used entirely on behalf of the protagonist (i.e. the analyst) in

the case of the GIS, there would be few of the problems associated with over-

intelligent antagonists. In GIS, the AI would always be on the player’s side; the

problem to be solved would play the abiding role of combatant. A second lesson might

be drawn from those racing games (e.g. Forza Motorsport 2) which provide both a

highly realistic simulation mode and a more forgiving ‘arcade’ driving mode (also

described earlier), with the latter permitting the player to accumulate considerable

physical damage to their vehicles without incapacitating them. The principle that is

useful in a GIS context is that the degree of simulation realism must always be fit for

purpose. Thus, for example, while a highly realistic dynamic spatial model might be

necessary for environmental planning purposes, a ‘dumbed down’ version might be

more appropriate in an educational context.

B. User interfaces

20

developing more effective methods for spatial exploration, search and interrogation

[46; 47] In this section, a number of VG interface ideas will be described, and their

GIS potential indicated. (See [48] for a fuller discussion.) The discussion will focus

primarily on software rather than on game hardware devices.

Figure 7. Highly stylised landscape of a first-person game with player’s avatar

temporarily on display (Team Fortress 2; courtesy of Valve Software).

Viewpoint

One of the most important aspects of the videogame user interface is its viewpoint --

i.e. the location of the player in relation to the virtual environment. Three viewpoints

are especially evident: the first-person viewpoint, in which the player moves through a

virtual environment using the eyes of the largely unseen protagonist (see Figure 1);

the third-person viewpoint, in which the player is distanced from the game’s

protagonist, which is represented by an avatar which they control (see Figure 7); and

the god viewpoint, in which the player looks down on, or obliquely across, the field of

play, directing events from a location that affords high situational awareness (see

Figure 8). (It should be noted that while the player’s control of characters in The Sims

have led to it being called a ‘god-game’, it does not adopt what is commonly referred

to as the god viewpoint.) There are variations on these main viewpoints, such as the

‘over-the-shoulder’ approach of some third-person shooters, in which the player is

partly immersed in the scene by following closely behind the seen protagonist, and the

‘through-the-rear-window’ viewpoint in some car racing games. The viewpoint serves

multiple functions in VGs, not only providing a degree of immersion (as in the first-

person viewpoint), or situational awareness (as in the other two viewpoints), but also

providing opportunities for user identification with the seen protagonist (as in the third-

person viewpoint).

21

Most games adopt a single viewpoint throughout. However, in recent years, some

popular games (e.g. Metroid) have switched from one viewpoint to another in

successive releases, and in other games, it is possible to switch between viewpoints

during gameplay. In the first-person Brothers in Arms and Ghost Recon, the player

can momentarily switch to a third-person camera to provide a tactical overview when

the combatant has taken cover. The RPG game The Elder Scrolls III: Morrowmind, for

example, offers both first- and third-person viewpoints, with the latter being useful

when the player’s avatar wishes to display its winnings. In some games, the viewpoint

is switched automatically, either to reveal important events (as in Civilization:

Revolution), to taunt opponents with amusing character animations or for replay

purposes (as in Team Fortress 2, Figure 8), or to reveal the player’s avatar when it is

injured (as in Left4Dead). The temporary switch between viewpoints to make use of

their particular advantages highlights the benefits of making both available in the same

game.

Figure 8. Urban infrastructure and irregular terrain as modelled in a grid-based

strategy game (Sim City IV; courtesy of EA Games).

In contrast, most GIS traditionally adopt the map-based, god viewpoint. Recent

innovations in Web mapping, however, are introducing alternative viewpoints more

akin to the first-person and third-person viewpoints found in videogames. (The

elevated camera position in the terrain-following viewpoint in Google Earth, however,

22

still maintains most of the elements of the god viewpoint.) As GIS introduce more 3D

scenes into the analyst’s working life, the need for additional and alternative

viewpoints and interaction methods is likely to increase markedly (

49

). One of the

striking things about most GIS is that little attempt is made to use viewpoints to

develop a sense of immersion, or to encourage user identification with a protagonist

(the analyst does not normally appears as an avatar in GIS scenes). By and large, the

analyst stays outside the scene, keeping his or her distance, and essentially playing

god.

Viewpoint choice has major implications for GIS design. Detailed studies have

been undertaken in a military context (e.g. [50; 51; 52; 53

A common feature of VGs is that specific interaction methods are designed to fit

particular player actions and operations. For example, methods to control various

]) which reveal that the

egocentric and exocentric viewpoints are suited to different tasks and benefit

users with different levels of experience. Since some of the tasks undertaken in

these military studies (e.g. judgements of relative position, mobility assessment

and line-of-sight visibility) are similar to those undertaken in GIS, it may be

important for GIS software to enable users to switch as needed between the two

types of viewpoint. A required first step might be to consider inserting an avatar

into the scene, and using a third-person viewpoint rather than the conventional

god-like viewpoint.

Camera position and movement

Camera control is an important means by which players are able visually to explore

the game environment, and is critical for rapid decision making. In some games (e.g.

the first-person shooter, and some car racing games), the camera equates to the

player’s eyes, and is locked to the player’s movement through the game environment,

which the player controls with mouse, keyboard or games controller. However, in

games where both the camera and the protagonist’s avatar have to be controlled

simultaneously (e.g. those with a third-person viewpoint), this poses a heavy

interaction load on players. To help resolve this problem, many such games take over

camera control, leaving players to control the movement of the avatar. There is

therefore a balance to be struck between player and software camera control, with

some games (e.g. Resident Evil 4) offering higher levels of player control, and others

(e.g. Super Mario Galaxy or God of War) offering lower levels of player control.

In some games (e.g. Ratchet & Clank), the player may opt either to control the camera

or to cede control to the program’s AI. A further distinction is made between, on the

one hand, character-relative motion, in which the avatar moves relative to their current

state in the game world (the classic example is the early Resident Evil series of

games), and on the other hand camera-relative motion, in which the avatar moves

relative to the viewpoint’s current state in the game world (the recent Resident Evil 4

finally switched to this approach). At root, this is a choice of how the user controls map

to in-game state change, with the primary purpose of automatic camera control being

to reduce the player’s control burden, especially in fast-moving and stressful

situations. However, in some games camera movement is also used for dramatic

effect. In Gears of War, for example, the camera shakes as the player runs, which

adds to the sense of immersion.

Fitness for purpose

23

locomotory actions by players are provided in a wide variety of games, and especially

those adopting a third-person viewpoint. However, because of the god-like perspective

adopted by the majority of GIS, their potential role in assisting spatial exploration and

analysis has been largely unexploited. (For example, the grip and climb systems

available in several action games, including Ico and Assassin’s Creed, could be

adapted for the exploration of mountainous terrains in GIS.

Interface diversity is a significant feature of videogames, with the same action often

controlled in different ways in different games. This can be illustrated by the process of

targeting characters and objects, which supports a variety of purposes, including

shooting, path planning, information extraction, selection, and general ‘trigger’

interactions. In Metroid Prime 3: Corruption, for example, a complex trigger facility

allows players to lock onto an enemy, and then either fix their perspective on the

target or look straight ahead. Players can also move the crosshair around the target

extremely rapidly after they have locked onto it, which makes it extremely easy to

handle multiple small targets within a clearly defined area. Finally, individual games

may provide several options for weapon-targeting, as illustrated by Grand Theft Auto

IV whose alternatives include: free aim; snap from target to target; target and modify;

and lock-to-objects (i.e. the camera’s anchor point locks to the object), all of which

operate on both static and dynamic objects. Many of these innovative interface

methods have potential applications in GIS.

Free and constrained exploration

Most early videogames constrained the player to movement in 2D space. The advent

of full 3D in the 1980s (e.g. Elite) and 1990s (e.g. Quake) provided players with

increasing freedom to explore and navigate in three dimensions. In parallel with this

increased freedom emerged numerous ways of mapping player manipulations of their

interface devices (mouse, keyboard, handheld controller, etc.) to a growing repertoire

of exploratory movements in space (walking, running, climbing, jumping, flying, etc.).

In time, extensive experimentation across thousands of game titles led to the

emergence of a handful of classical interface mechanisms whereby spatial exploration

and action could be undertaken. Among these were the first-person running and

jumping (Quake), the third-person running and jumping (Jak & Daxter), the third-

person per-scene viewpoints (Alone In The Dark), and the god’s eye view (Sim City).

Although there has been a progressive move in VGs towards increasing freedom of

spatial exploration, this freedom has often been deliberately constrained by games

designers, chiefly because of the challenge to players of controlling free movement in

three spatial dimensions. One form of constraint is found in those games (e.g. Jak &

Daxter, Ratchet & Clank and World of Warcraft) which, although they present a highly

realistic 3D virtual environment on screen, constrain the movement of the player’s

avatar mainly to the degrees of freedom of the ground surface. Another technique

adopted in shooter games is to adopt an ‘on rails’ style during intense fighting

episodes (as in Resident Evil: The Umbrella Chronicles), in which the camera follows

a fixed path through a battle zone, and the player merely controls the direction it faces

in order to shoot at large numbers of objects and assailants en route. In a similar

manner, the circuit racer game largely prescribes where the player is able to drive

during gameplay. Although they may have extensive rooms, the castle (as in

Wolfenstein 3D) and the underground laboratory (as in Half-Life) offer similarly

constrained spaces. The ‘constrained navigation’ approaches for moving across

24

complex 3D terrains using 2DOF of movement [54

Many games (e.g. those involving driving, unarmed combat, skateboarding and

parkour) require the player to control complex on-screen behaviour, either of vehicles

or their personal avatars. In most such games, these behaviours (e.g. the trick

repertoire in successive editions of the Tony Hawk’s skateboarding game, or the

skyline leaping and running behaviour of the hero in Assassin’s Creed) are achieved

by far less complex control manipulations, and in some cases only a single button

press is required [

] is of particular relevance in this

context.

The recent emergence of non-linear, free-roaming games (championed by GTA III,

and taken up by such titles as Far Cry 2, Burnout Paradise and Tony Hawk’s

American Wasteland) is beginning to challenge the hegemony of titles which keep the

player within confined spaces or moving along constrained courses. This spatial

opening up of games has led to players needing on-screen orientation and navigation

aids. Although games have tended to avoid displaying on-screen aids, in order to

sustain a sense of immersion, the emergence of ‘open world’ VGs set in extensive

environments has led to the re-appearance of devices such as the inset map or

minimap (e.g. GTA San Andreas, which has both an inset minimap and a fullscreen

map view). Metroid Prime goes further than most games with its simplified, rotatable

and zoomable 3D wireframe map. GIS have much to learn from videogames in this

respect, especially as and when they adopt third-person user modes involving an

analyst avatar.

Simplification

A notable feature of VG interfaces is the way in which often complex on-screen

behaviour is achieved by relatively simple interface actions on the part of the player.

This process of interface simplification is enabled by the games software, which

undertakes the necessary behavioural elaboration. In Super Mario Galaxy, for

example, the protagonist’s avatar is not directly controlled by the player (nor, indeed,

is the camera, as discussed earlier). Moreover, not only are Mario’s jump trajectories -

- especially between planetoids -- software generated, but the use of modified gravity

makes walking across planet surfaces or up walls without falling off as simple as

walking across the ground. Indeed, this would be an entirely different game if players

were required to take fine-grained behavioural control of their on-screen avatar, with

no distortion of real-world physical laws to help them.

55]. Much of this complex on-screen behaviour is pre-animated, and

in some cases (e.g. when the player wiggles their joystick around to control their

avatar), several animations are cross-blended and phased in and out to produce

plausible walking animation. In Splinter Cell: Conviction, not only are the player’s

relatively simple control actions (e.g. pressing a button) translated into more complex

on-screen behaviour (e.g. throwing an opponent across a room), but the software also

intelligently analyses the current context, and elaborates the behavioural response to

the player’s control action in a way that is most appropriate for a given situation. A

context-dependent approach is taken in several games, so that the action triggered by

using a particular control depends on the current state of play. In Mirror’s Edge, for

example, instead of providing a ‘grab’ button, the software uses edge-detection

routines to determine whether the player’s avatar is close enough to a building ledge

to invoke a gripping action. Similar techniques are used in Fable II, which generates

moves for its melee fights that reflect their immediate context. It is in ways like these

25

that the design of virtual environments in VGs is able to guide and encourage player

behaviour, in ways that are largely absent from GIS.

Although the advantages of interface simplification and behavioural elaboration are

relatively obvious, there are also dangers to be avoided. One is the player frustration

caused when the software takes autonomous decisions about the game context,

especially when there are several potential actions available at a given game point,

and the software chooses one that does not match the player’s intention. Also, many

commentators and players part company with games designers when elaborate cut

scenes or heavy AI are used to animate spectacular action sequences (as found in the

introductory sequences of many VGs, and in the destruction sequences in Half-Life:

Episode 2). Because player interaction with the game is normally suspended while cut

scenes are playing, the player’s role is reduced to little more than that of a spectator.

C. Multi-sensory representation

In an earlier section, data visualization was shown to be an important ingredient of

effective gameplay. However, not only is it an informal craft in most VGs, but the rule-

based creation of visual symbols from data is intermixed with the design of graphical

symbolism which serves other, less representational roles. Although the advantages

of presenting spatial information to more than the visual human sense are well

established [19; 20], most GIS still privilege the sense of sight. In plotting the future

development of data representation, GIS designers could therefore do far worse than

examine the track record of VGs, many of which have mastered the art of multi-

sensory information presentation [56]. We will begin by examining how sound and

haptics are used to convey information to players, and then consider the effective

combination of multi-sensory information.

Sound

Sound, and especially music, is used mainly for emotional effect in videogames, as is

exemplified by the mood established by the soundscapes of such games as Defcon

and Ico. However, there are cases where music is also used to represent information.

For example, it can indicate different states of play, including the presence of danger

and a lack of remaining time, and can also provide notification of certain events. In

Left4Dead, for example, one music track begins just before a ‘horde’ (of zombies)

attacks, another starts up as a ‘tank’ (a super-powerful zombie) approaches (and

continues playing until it is dead), and yet another gradually increases in volume and

sound layers as the player approaches the hiding place of a ‘witch’ (another extremely

dangerous zombie). In one scene of Super Mario Galaxy, the tempo of the music

increases and decreases in step with the player’s movement of a star ball. Such

redundant representation, in which sounds reinforce visuals, is found in many games.

In Ratchet & Clank, for example, sound encodings are used to convey the nature of

the contrasting surfaces on which the protagonists are walking. In Crysis, however,

the game deviates from the rather strict rules of data visualization, in that weapons

which sound similar in the real world are given distinctive sounds in the game because

of the importance of players distinguishing aurally between them. A similar form of

aural exaggeration, this time for emotional effect, is used in the Thief games, in which

the volume of the protagonist’s footsteps on certain surfaces is enhanced to alert them

to the fact that they might be heard by others.

26

It is very common for classes of entities to be identified by means of audio 'tags' in

VGs. In Half-Life 2, for example, each type of enemy has a distinctive aural signature,

and will emit easily recognised sounds when distinctive events occur. (These sounds

are of crucial importance when those events are off screen or obfuscated by other on-

screen events.) Similarly, audio tags are used to provide aural confirmation and

identification as the player collects types of item (e.g. medical supplies or

ammunition), or defeats types of enemy. Sound can also be used to reflect measured

data, through variations in pitch or frequency (though not amplitude, which is generally

used to encode distance-from-viewpoint). For example, puzzles in the Zelda series of

games give timing information to the player in this way.

A further role for sound in VGs is in providing locational information. Variously named

spatial audio, localised sound, spatialised sound, positional audio or 3D sound, the

ability to alert players aurally to the location of characters and events in 3D space is

an important element of the soundscape of most modern VGs. 3D spatialisation of

audio implicitly provides scalar feedback by changing audio volume on the basis of a

sound source’s distance from the player or viewpoint. Most game designers have

extensive soundscape-building tools, which means they are able to add sound to a

virtual environment by identifying the location of sound emitters interactively and

drawing boxes around spaces affected by individual sounds.

Haptics

Due to the widespread availability of interface devices (including joysticks and

game controllers) that emit vibrations or output force feedback, many games are

able to harness the haptic senses (the vibrotactile, which is mainly located in the

skin, and the kinaesthetic, which is mainly located in the joints) to transmit

information to the player. (The kinaesthetic sense has the unique property of both

receiving and issuing force or stress, and is thus uniquely able to engage in

active exploration of environments.) Many games (e.g. Ratchet & Clank: Tools of

Destruction, Motorstorm, and Heavenly Sword) adopt rumble -- i.e. vibrotactile

feedback -- to identify events occurring in the virtual environment, with the 'profile'

of a rumble (that is, how its strength varies over time) commonly used to

differentiate between different kinds of event. Controller rumble is also used to

generate a sensation which indicates that the player's avatar is in contact with

certain surfaces in the virtual environment. Scalar feedback can also be provided

by the controller, as when the amplitude of a sound or the frequency of an aural

pulse is used to convey proximity to otherwise secret (e.g. underground) items.

Force-feedback devices are not normally used to provide categorical information

to players, but are commonly used to emit the values of continuously varying

parameters. In some flight simulator games, for example, force-feedback

joysticks are used to mirror the forces felt through the control column in real

aircraft, and force-feedback steering wheels available for some driving games

convey the resistance of the car's wheels to turning based on physical simulation

of the vehicle in contact with the road.

Combining sensory outputs

Multi-sensory output has two broad roles: sensory substitution, in which one sensory

modality is replaced by another, and sensory combination, in which one sensory

output is used to reinforce another through redundant data coding, or to encode

27

multiple data variables [56]. A great deal of sensory fusion plays a redundant encoding

role. In Zelda: Twilight Princess, for example, the Wii controller combines both sound

and tactile feedback (by emitting a 'twang' sound and rumbling) when the player fires a

bow and arrow. Similarly, in Super Mario Galaxy, the Wii controller rumbles when the

player’s avatar lands from a great height. Audio cues can help players focus on a

critical visual element, for example, by teaching the player to associate a distinctive

audio cue with a given enemy action (and its related animation, such as raising an arm

high prior to striking the player). In several games, spatial distance and audio volume

are often used together to help separate the camera from the player avatar, or

conversely to make them feel more tightly bound to them. An interesting (but rather

rare) effect found in some third-person games (in which the player controls an avatar),

is where the visual and audio information is decoupled, in that audio effects are

spatialised as if heard from the avatar’s location (i.e. as if a virtual microphone was

positioned there) rather than from the camera position.

The multi-sensory output of multivariate information is perhaps less frequently used

than redundant encoding of individual variables. However, in some car racing games

(e.g. Forza Motorspot 2), the driving experience is fully multi-sensory in a multivariate

sense, with a combination of visual, aural and tactile (e.g. rumble) feedback being

continually provided. The non-visual information is highly nuanced, conveying a

variety of engine, exhaust and tyre noises with both a high degree of realism and

emotional impact.Some other games also use multi-sensory output to reflect

multivariate information. For example, Pikmin has a hundred or so little creatures

running around following the player’s avatar, and colour is used to represent the type

of pikmin and sound (in addition to animation) is used to represent their current state

(i.e. whether they are working, running, endangered, hurt, etc). It should be borne in

mind that the integration of multiple sensory outputs in VGs combines with tightly

integrated storylines and compelling game mechanics to transcend mere information

presentation, enabling the player to participate in an immersive experience. While this

might give VGs the capacity to support marketing activities, it might not mesh in an

obvious way with the typical requirements of the spatial analyst.

IV. CONCLUSIONS

In this chapter, we have examined two kinds of spatially intelligent software: the

videogame and the GIS, and have compared several of their key characteristics and

capabilities. One of the major conclusions drawn from our analysis is that some of the

differences between videogames and GISs are more apparent than real, and that the

former have a surprising number of similarities with the latter. A second conclusion is

that VGs constitute a public laboratory in which numerous innovative ideas have been

road-tested with large numbers of highly demanding users. The results of these

experiments are freely accessible to GIS designers, and can serve to short-circuit the

improvement of existing features as well as the introduction of additional functionality.

Videogame contributions

We have evaluated VGs against key GIS capabilities to reveal their strengths and

weaknesses, and have identified several areas in which VGs could make potentially

significant contributions to tasks normally undertaken with GIS. Of course, this is not

to suggest that particular videogames could be used off the shelf as fully-fledged GIS,

28

but it is to suggest that some widely exhibited functionalities in VG technology could

contribute in significant ways to the future development of GIS.

It has been argued that in three specific areas (dynamic process modelling, interfaces

and multi-sensory representation), VGs exhibit more developed capabilities than GIS,

and have much to offer the spatial analyst. Firstly, VGs may be characterised as

dynamic, process-based simulation systems, in which the interactions between

players and modelled processes provide much of the software’s functionality and

appeal. In contrast, GIS are essentially spatial data handling environments, which for

the most part display spatio-temporal data rather than model spatio-temporal

processes. In this context, VGs arguably have more in common with dynamic,

process-based geospatial simulations than with GIS. Nevertheless, given numerous

past attempts to enable dynamic modelling in a GIS framework, VG achievements in

this field are well worth examining.

Secondly, it has been shown how VG interfaces and interaction styles are highly

diverse, and may be selected according to the needs of individual players and the

tasks they are required to perform. The way in which VG interfaces are designed for

exploring 3D virtual environments offers GIS designers many ideas for providing

analysts with additional interactive approaches for exploring their own geographical

worlds. Finally, we have shown that while GIS are generally more advanced in rule-

based data visualization, the leadership roles are reversed when the representation of

data to other sensory modalities is concerned. We have demonstrated ways in which

VGs might help to rectify a longstanding gap in GIS functionality.

Both VGs and GIS may be broadly interpreted as being concerned with solving

problems in a spatial context. While it is readily apparent that GIS tools for spatial

modelling, display and analysis provide support for environmental and social problem

solving and decision making, this role may not be so evident with VGs. Nevertheless,

games present players with considerable perceptual, motor and cognitive challenges

in spatial environments, and innumerable tools with which to solve spatial problems, at

several levels of challenge and complexity. VG approaches and tools may therefore

be relevant for undertaking tasks equivalent to those found in a GIS context. For

example, given the highly developed capability of VGs in assisting players to navigate,

explore and interrogate virtual environments, many of the innovative techniques

developed for 3D navigation and interaction might also be useful in digital GIS worlds.

In this context, it is instructive to note Google’s current positioning of its own GIS-

related functionality in term of: ‘Explore, Search and Discover’. However, VGs offer

considerably more than quasi-3D map browsers in application areas such as virtual

tourism and virtual heritage exposition [57

Charting the future

A not insignificant question raised by this chapter concerns the means by which VGs

might contribute to GIS in the future. Several possible scenarios may be outlined

whereby spatial analysts might be able to acquire more game-like tools. The first

would involve analysts modifying selected VGs to undertake GIS-related activities.

There are already indications of this happening in several other application areas,

] which require the use of highly realistic 4D

technology. The fact that an increasing number of today’s youth generation is already

well versed in using VG technology should give GIS developers considerable food for

thought.

29

including: scientific visualization [58], marketing [59], education [60], scientific

research [61], military training [62; 63], and environmental planning [64]. A second

route might see developers using VG design toolkits (e.g. VG middleware) for rapidly

prototyping and deploying games technology for GIS applications. This approach is

perhaps most attractive for 3D data visualization applications [65]. A third scenario

sees games developers broadening their target markets to include spatial exploration

applications. There are several indications (e.g. in serious games, passive games and

some casual games) that the industry might have an appetite for this. A fourth

scenario might be based on interoperable collaboration between GIS and VGs,

creating powerful software federations that capitalise on the individual strengths of

each technology [5]. A final scenario would be for the developers of GIS and other

spatial software to adopt ideas read-tested in VGs in their own software, to deliver

added value to users. An example of this approach may be seen in the adoption of

MMOG game mechanics in business software designed for trans-national

organisations [66

The authors wish to thank the organisers and attendees at the Virtual Geographical

Environments conference, held in Hong Kong in January 2008, at which some of the

].

It remains an open question as to whether ‘entertaining GIS’ will become as popular or

as relevant as ‘serious games’. Equally uncertain is whether there will be increasing

convergence of videogame and GIS technologies in the future. It remains to be seen,

for example, whether continuing innovations in videogame technology will contribute

to major improvements in GIS, whether videogames will chip away at the GIS

consumer base with products that are increasingly able to model worlds for insight as

well as for entertainment, or whether a visionary product might appear that

encapsulates the best that both technologies currently have to offer. Amongst this

exciting uncertainty, one conclusion emerges with clear-cut outlines: those of us who

use technology to understand and manage the world we inhabit are facing a future of

ever-increasing possibilities, with technological options becoming increasingly

available from complementary sources. Without succumbing entirely to the hubris of a

former political leader, it may truly be said that never in the history of spatial science

have we had it so good.

NOTES

The names of the videogames included as acronyms in Figure 2 are:

AC (Assassin’s Creed), BiAHH (Brothers in Arms: Hell’s Highway), BP (Burnout

Paradise), BR (Braid), CIV (Civilization), CoD4 (Call Of Duty 4), CoDWaW (Call of

Duty: World at War), CR (Crysis), DC (Defcon), EC (Echochrome), FR (Fracture), F2

(Fable II), FF (Final Fantasy), FM2 (Forza Motorsport 2), GA (Getaway), GoW (God Of

War), GTAVC (Grand Theft Auto: Vice City), HGL (Hellgate: London), J&D (Jak &

Daxter), KD (Katamari Damacy), LBP (Little Big Planet), MFS (Microsoft Flight

Simulator), PM (Pacman), RCGC (Ratchet & Clank: Going Commando), SC4 (Sim

City 4), SMG (Super Mario Galaxy), SP (Spore), STH (Sonic The Hedgehog), TS (The

Settlers), and WoW (World of Warcraft).

ACKNOWLEDGEMENTS

30

ideas discussed in this chapter were first presented. Thanks are also due to EA

Games, Introversion Software Ltd. and Valve Software for permission to reproduce

images from their games. The authors have made every effort to secure the

necessary permissions to use copyright material. If there has been any oversight,

copyright holders are asked to contact the publishers.

REFERENCES

1

Marble, D. F. & Peuquet, D. J. (1983) Geographic information systems and remote

sensing, in: R. N. Colwell (ed.) Manual of Remote Sensing, second edition, Vol. 1,

pp.923-958 (ASPRS/ACSM, Falls Church, VA).

2

Fisher, P. F. & Lindenberg, R. (1989) On distinctions among cartography, remote

sensing, and geographic information systems, Photogrammetric Engineering and

Remote Sensing, 55(10), pp.1431-1434.

3

Duecker, K. J. (1987) Geographic information systems and computer-aided mapping,

Journal, American Planning Association, 53, pp.383-390.

4

Cowen, D. J. (1988) GIS versus CAD versus DBMS: what are the differences?

Photogrammetric Engineering and Remote Sensing, 54, pp.1551-1555.

5

Shepherd, I. D. H. (1990) Build your own GIS?, Land and Minerals Surveying,

8(4), pp.176-183.

6

Picard, M. (2008) Video Games and Their Relationship with Other Media, in:

Mark J.P. Wolf (ed.) The Video Game Explosion: A History from Pong to

PlayStation and Beyond, pp. 293-300 (Greenwood Press, Wesport, Conn.).

7

Murray, J. (1997) Hamlet on the Holodeck: the future of narrative in cyberspace

(Cambridge, MA: The MIT Press).

8

Aarseth, E. (2007) Allegories of space: the question of spatiality in computer games,

in: F. von Borries, S. P. Walz & M. Böttger, Space, Time, Play: Computer games,

architecture and urbanism: the next level (eds) pp.44-55 (BirkHäuser).

9

Aarseth, E. (2007) Allegories of space: the question of spatiality in computer games,

in: F. von Borries, S. P. Walz & M. Böttger, Space, Time, Play: Computer games,

architecture and urbanism: the next level (eds) pp.44-55 (BirkHäuser).

10

Wolf, M. J. P. (2002) The Medium of the Video Game (University of Texas Press).

11

Anon (2008b) The making of Grand Theft Auto, Edge, April, pp.61-75.

12

Anon (2006) Never a city so real? Available at:

http://ask.metafilter.com/46273/Never-a-city-so-real (accessed 18 October 2008).

13

Parish, Y. I. H. & Müller, P. (2001) Procedural modelling of cities, Proceedings

ACM SIGGRAPH, pp.301-308 (Los Angeles, CA: ACM Press).

31

14

Hudson-Smith, A. & Evans, S. (2003) Virtual cities: from CAD to GIS, in: P. A.

Longley & M. Batty (eds) Advanced Spatial Analysis: the CASA Book of GIS,

pp.41-60 (Redlands, CA: ESRI Press).

15

Taylor, F. (2008) Paragliding simulator with Google Earth plugin,

gearthblog.com/blog/archives/3d_models/ (accessed 23 October 2008).

16

Biffo, Mr. (2007) Down with physics, Edge, November, p.124.

17

Crawford, C. (1984) The Art of Computer Game Design (Berkeley, CA: McGraw-

Hill).

18

Anon (2008a) Conflicting stories. Fun or real: what do war games want to be?,

Edge, February, p.42.

19

Shepherd, I. D. H. (1994) Meeting the Design Challenge of Multi-Sensory GIS',

Proceedings AGI '94, pp.12.1.1-12.1.9 (London: Association for Geographical

Information).

20

Shepherd, I. D. H. (1995) Multi-sensory GIS: mapping out the research frontier,

in: T. Waugh (Ed.) Proceedings of the 6th International Symposium on Spatial

Data Handling, pp.356-390 (London: Taylor & Francis).

21

Goodchild, M.F. & Gopal, S. (eds) (1989) Accuracy of Spatial Databases (London:

Taylor & Francis).

22

MacEachren, A. M. (1992) Visualizing uncertain information, Cartographic

Perspectives, 13, pp.10-19.

23

Flowerdew, R. (1997) Spatial Data integration, in: Maguire, D. J., Goodchild, M. F. &

Rhind, D. (eds) Geographic Information Systems: principles and applications, Vol. 1,

pp.375-387 (Harlow, UK: Longman Scientific and Technical).

24

Shepherd, I. D. H. (1997) Information integration, in: Maguire, D. J., Goodchild, M.

F. & Rhind, D. (eds) Geographic Information Systems: principles and applications,

Vol. I, pp.337-360 (Harlow, UK: Longman Scientific and Technical).

25

Bertin, J. (1973) Semiologie graphique: les diagrammes, les reseaux, les cartes,

2nd edition, Paris-La Haye: Mouton. (First French edition, 1967; first English

edition, 1983, tr. W. J. Berg, Semiology of Graphics: diagrams, networks, maps).

26

Bertin, J. (1978) Theory of communication and theory of 'The Graphic',

International Yearbook of Cartography, 18, pp.118-126.

27

Bertin, J. (1979) Visual perception and cartographic transcription, World

Cartography, 15, pp.17-27.

28

Bertin, J. (1981) Graphics and graphic information-processing, Translated by: W.

J. Berg & P. Scott, Berlin; New York: Walter de Gruyter. (Original French edition:

La graphique et le traitement graphique de l'information, 1977, Paris:

Flammarion).

32

29

Bertin, J. (1983) A new look at cartography, in: D. R. F. Taylor (ed.) Graphic

communication and design in contemporary cartography, Chichester: John Wiley,

pp.69-86.

30

Kraak, M-J. (1988) Computer-assisted Cartographical Three-dimensional Imaging

Techniques, Delft University Press, Delft.

31

Kraak, M-J. (1989) Computer-assisted cartographical 3D imaging techniques, in

Three-dimensional applications in geographical information systems (ed. J.

Raper) Taylor & Francis, London, pp.99-114.

32

Poole, S. (2000) Trigger Happy: the inner life of videogames (London: Fourth

Estate).

33

Peuquet, D.J. (2001) Making Space for Time: Issues in Space-Time

Representation, Geoinformatica, 5(1), pp.11-32.

34

Peuquet, D. J. (2002) Representations of Space and Time, (New York, NY:

Guilford).

35

Dorling, D. (1992) Stretching space and splicing time: from cartographic animation to

interactive visualization, Cartography & GIS, 19(4), pp.215-227.

36

Roberts, J. C. & Wright, M. A. E. (2006) Towards ubiquitous brushing for

information visualization, Information Visualization, 5(7), pp.151-156.

37

Merriam D F. Computer Applications in Geology–Two Decades of Progress.

Proceedings of the Geological Association, 1980, 91: 53-58.

38

Benenson, I. & Torrens, P. (2004) Geosimulation: Automata-based modeling of

urban phenomena (John Wiley & Sons).

39

Goodchild, M.F, Parks, B. O. & Steyaert, L. T. (1993) Environmental Modelling with

GIS (New York, NY: Oxford University Press).

40

Goodchild, M.F, Steyaert, L. T., Parks, B. O., Johnston, C., Maidment, D., Crane, M.

& Glendinning, S. (eds) (1996) GIS and Environmental Modelling: progress and

research issues (Fort Collins, CO: GIS World Books).

41

Steyaert, L. T. & Goodchild, M. F. (1994) Integrating geographic information

systems and environmental simulation models: a status review. In: Michener W K,

Brunt J W, and Stafford S G (eds). Environmental Information Management and

Analysis: Ecosystem to Global Scales, pp.333-356 (Bristol, PA: Taylor & Francis).

42

Smith, R. (2007) Toeing the disassembly line, Edge, Christmas, p.122.

43

Hamden-Turner, C. & Trompenaars, F. (1997) Riding the Waves of Culture:

understanding diversity in global business (McGraw-Hill).

44

Hofstede, G. (2003) Culture’s Consequences: comparing values, behaviours,

institutions and organizations across nations (Sage Publications).

33

45

Anon (2008c) Artificial idiocy, Edge, June, pp.74-79.

46

Germanchis, T., Pettit, C. & Cartwright, W. (2005) Using Computer Gaming

Technology to Explore Human Wayfinding and Navigation Abilities Within a Built

Environment, Proceedings of the 21st International Cartographic Conference.

(Coruña: International Cartographic Association).

47

Cartwright, W. E. (2006) Exploring Games and Gameplay as a Means of Accessing

and Using Geographical Information, Human IT, 8(3), pp.28–67.

48

Shepherd, I. D. H. & Bleasdale-Shepherd, I. D. (2008a) Towards effective

Interaction in 3D data visualizations: what can we learn from videogames

technology?, Paper presented at the International Conference on Virtual

Geographic Worlds, Hong Kong, 7-8 January 2008.

49

Shepherd, I. D. H. (2008) Travails in the third dimension: a critical review of 3D

computer graphics in geographic data visualization, in: M. Dobson & M. Turner

(eds) Geographic Visualization: Concepts, Tools and Applications, pp.199-222

(Chichester: John Wiley).

50

Wickens, C. D. (1999) Frame of reference for navigation, in: D. Gopher &. A.

Koriat (eds) Attention and Performance, (Vol. 16), pp.113-143 (Orlando, FL:

Academic Press).

51

Banks, R. & Wickens, C. D. (1997) Commanders’ display of terrain information:

manipulations of display dimensionality and frame of reference to support

battlefield visualization, Technical Report ARL-97-12/FED-LAB-97-2, US Army

Research Laboratory, Aberdeen, MD.

52

McCormick, E. P., Wickens, C. D., Banks, R. & Yeh, M. (1998) Frame of

reference effects on scientific visualization subtasks, Human Factors, 40(3),

pp.443-451.

53

Thomas, L. C., Wickens, C. D. & Merlo, J. (1999) Immersion and battlefield

visualization: frame of reference effects on navigation tasks and cognitive

tunnelling, Technical Report ARL-99-3/FED-LAB-99-2, US Army Research

Laboratory, Aberdeen, MD.

54

Hanson, A. and Wernert, E. (1997) Constrained 3D navigation with 2D

controllers, IEEE Visualization, 175-182.

55

Anon (2008d) Leap of Faith, Edge, August, pp.54-61.

56

Shepherd, I. D. H. & Bleasdale-Shepherd, I. D. (2008b) Multi-sensory Data

Realization in Virtual Worlds: abstraction or pragmatism, Paper presented at the

International Conference on Virtual Geographic Worlds, Hong Kong, 7-8

January 2008.

57

Champion, E. (2003) Applying Game Design Theory to Virtual Heritage

Environments, Proceedings of the First International Conference on Computer

34

Graphics and Interactive Techniques in Australasia and South East Asia, Melbourne,

Australia. pp.273-274.

58

Rhyne, T-M. (2002) Computer games and scientific visualization,

Communications of the ACM, 45(7), pp.40-44.

59

Miliano, V. (1999) Unreality: application of a 3D game engine to enhance the

design, visualization and presentation of commercial real estate, VSMM99.

60

Woods, S. (2004) Loading the dice: the challenge of Serious videogames,

Game Studies: The International Journal of Videogame Research, 4(1), also

available at: http://www.gamestudies.org/0401/woods/ (accessed 19 November

2007).

61

Lewis, M. & Jacobson, J. (2002) Game engines in scientific research,

Communications of the ACM, 45(1), pp.31.

62

Riddell, R. (1997) Doom Goes to War, Wired, April, pp.114–118.

63

Raybourn, E. M. & Bos, N. (2005) Design and evaluation challenges of serious

games, Proceedings CHI ‘05, pp.2049-2050 (New York: ACM Press).

64

Guy, B., Bidwell, N. J. & Musumeci, P. (2005) GamePlan: serious gaming for

place making, Proceedings of the Second Australian Conference on Interactive

Entertainment, pp.252-255.

65

Shepherd, I. D. H. (2002) It's only a Game: using interactive graphics middleware to

visualise historical data, Society of Cartographers Bulletin, 36(2), pp.51-55.

66

Kim, A. J. (2007) Putting the fun in functional: applying game mechanics to

functional software, available at: http://www.slideshare.net/amyjokim/putting-the-fun-

in-functional (accessed 23 September 2008).

CHEAP AND IMMERSIVE VIRTUAL REALITY: APPLICATION IN CARTOGRAPHY

Article

Full-text available

Sep 2018

In this paper, we analysed and tested the possibilities for the use of low-cost VR (Virtual Reality) headsets, with a focus on cartographic visualization. Low-cost devices were compared and classified into three categories (low-end, mid-range and high-end). We also created a pilot virtual environment, called “Carthoreality”, and conducted simple pilot user testing using this virtual environment and the three low-end headsets. Our pilot test shows a few drawbacks to these devices, including weight of the headset and penetrating light. Some problems in terms of user aspects were also identified, such as nausea while wearing headsets or disorientation after removing them, which occurred for all users.

Defacing the Map: Cartographic Vandalism in the Digital Commons

Article

Full-text available

Apr 2014
CARTOGR J

Andrea Ballatore

This article addresses the emergent phenomenon of carto-vandalism, the intentional defacement of collaborative cartographic digital artefacts in the context of volunteered geographic information. Through a qualitative analysis of reported incidents in WikiMapia and OpenStreetMap, a typology of this kind of vandalism is outlined, including play, ideological, fantasy, artistic, and industrial carto-vandalism, as well as carto-spam. Two families of counter-strategies deployed in amateur mapping communities are discussed. First, the contributors organise forms of policing, based on volunteered community involvement, patrolling the maps and reporting incidents. Second, the detection of carto-vandalism can be supported by automated tools, based either on explicit rules or on machine learning.

A Cost-effective Workflow for Depicting Landscapes in Immersive Virtual Environments

Chapter

Full-text available

Nov 2017

Nathaniel Henry

Due to GIS software conventions and cost restrictions, many geogames today visualize their gaming environments from a top-down perspective and tend towards symbolic rather than naturalistic representations of game phenomena. This chapter introduces a workflow for representing real-world geographies as three-dimensional environments, allowing users to navigate from a first-person perspective and view ground objects as they might appear in real life. The workflow combines existing data collection and 3D reconstruction methods with a novel technique for importing geographic data into a game engine, offering geogame designers a low-cost and time-efficient method for representing real terrain features within a 3D virtual environment. By expanding the representational possibilities available to geogames, this technique facilitates player immersion in geogames, ultimately enhancing the player experience and learning outcomes.

Rendering technique of multi-layered domain boundaries and its application to fluid flow in porous media visualizations

Article

Full-text available

Nov 2014

Current visualization techniques for computational fluid dynamics applications are sophisticated and work well in simple geometries. For complex geometries such as pore spaces, multiple domain boundaries obstruct the view and make the studying of fluid flow fields difficult. To overcome these deficiencies, we use two-sided materials to render the domain boundaries. Using this technique, it is possible to place the camera inside the domain and have a non-obstructed view of the surrounding flow field without losing spatial reference to the domain boundaries. As a result, a larger part of fluid flow visualization is visible. Two-sided material rendering was successfully applied to display still images with Blender Cycles renderer, in a virtual reality environment, and several implementation techniques were explored for using the Visualization Toolkit.

Navigating sociotechnical spaces: Comparing computer games and sat navs as digital spatial media

Article

Full-text available

Aug 2012

Chris Chesher

Digital media increasingly mediate everyday spatial and navigational practices. From in-car satellite navigation (sat navs) to computer games, overpowered gadgets are combining multiple sources of abstract information to give users spatial guidance and experiences of movement. For example, open world computer games such as Grand Theft Auto IV render rich fictional spaces, and include intricate maps and indicators that allow players to navigate large gamespaces. Sat navs such as the TomTom Navigator follow similar practices of automated navigation in helping to guide cars through actual spaces. Their calculated routes display on personalized maps, including live data and visualizations that complement, or even override, what the driver sees through the windscreen. Games and sat navs are harbingers of historical shifts in technosocial space, suggesting that Henri Lefebvre’s (1991) influential critical analysis of space deserves to be revised. Digital spatial media open up abstract relationships to space, but not from the distance that Lefebvre associates with ‘conceived’ spaces. Instead, they work in ‘lived space’, which is becoming dominant. They calculate space in real time, and open up new political and aesthetic questions. The article examines three characteristics of navigation with digital spatial media: (1) they reify routes as persuasive data and procedures; (2) their maps become subjective and privatized; and (3) they offer an array of spatial information that become incorporated into the user’s ‘perceived’ space. These examples show that critical understandings of social space need increasingly to incorporate readings of digitally mediated spatiality.

Serious Games for Geographical Field Skills: An Exploration

Conference Paper

Apr 2010

This paper explores the synergies between geography, GIScience and serious gaming. From a skills-based rather than issues-based direction, the work highlights opportunities for using games engines for the development of fieldwork preparation tools. Our emphasis lies with exploring the contention that games can foster associative learning. Specifically, we evaluate the role that presence may play in realizing this goal.

The design-by-adaptation approach to universal access: Learning from videogame technology

Article

Full-text available

Aug 2011

This paper proposes an alternative approach to the design of universally accessible interfaces to that provided by formal design frameworks applied ab initio to the development of new software. This approach, design-by-adaptation, involves the transfer of interface technology and/or design principles from one application domain to another, in situations where the recipient domain is similar to the host domain in terms of modelled systems, tasks and users. Using the example of interaction in 3D virtual environments, the paper explores how principles underlying the design of videogame interfaces may be applied to a broad family of visualization and analysis software which handles geographical data (virtual geographic environments, or VGEs). One of the motivations behind the current study is that VGE technology lags some way behind videogame technology in the modelling of 3D environments, and has a less-developed track record in providing the variety of interaction methods needed to undertake varied tasks in 3D virtual worlds by users with varied levels of experience. The current analysis extracted a set of interaction principles from videogames which were used to devise a set of 3D task interfaces that have been implemented in a prototype VGE for formal evaluation.

Multisensory Data Representation in Virtual Worlds: abstraction or pragmatism?

Article

Full-text available

Contemporary data visualisation and GIS software continue to make almost exclusive use of the visual sensory modality in representing data to the human analyst. The potential roles of other sensory modalities in virtual geographical environments (VGEs), and especially those designed to support interactive data exploration, are identified. A trans-sensory model of sensory variables is proposed, built on Bertin's original graphical sign system, as a basis for the fusion of multiple sensory representations of data in multi-sensory data representation systems (MSDRSs). This is followed by a detailed evaluation of the (largely intuitive) adoption of sensory variables in contemporary videogame technology, in order to identify lessons which may be used to embed this model in future VGEs that aim to provide facilities for multi-sensory data representation.

Computer-assisted cartographic three-dimensional imaging techniques

Article

Jan 1986

Menno-Jan Kraak

In this paper the CARTDIM research project is presented. It attempts to trans fer the general interest in 3D computer graphics to cartography. With knowled ge of computer graphics and of cartographic theory a methodology for 3D car tography will be developed. To accomplish this, four types of 3D-maps will be produced and tested with mapusers. Comparison of the prototypes and their 2D counterparts will lead, via feedback to conclusions about design, the viability of the production process and to general conclusions about the prospects of 3D-cartography.

Integrating geographic information systems and environmental simulation models: a status review

Article

Jan 1994

The status of GIS integration is reviewed, with emphasis on dynamic, physically based modelling approaches that integrate the effects of different environmental processes across multiple scales. The first steps have been taken toward more tightly linked approaches, including embedded modelling within the GIS. Integrated multidisciplinary modelling has created new opportunities for GIS, but has also exposed the weaknesses of the current generation of GIS technology, particularly in areas of functionality and data modelling. -Authors

GIS versus CAD versus DBMS: what are the differences?

Article

Jan 1988
PHOTOGRAMM ENG REM S

Cowen D.J

GIS is best defined as a dicsion support system involving the integration of spatially referenced data in a problem solving environment. The most important part of this definition is the emphasis on integration. In other words, GIS provides the tools, particularly polygon overlay, that we have always needed to truly synthesize disparate sources of spatial information. Earlier forms of automated geography that simply retrieved, manipulated, or displayed predefined geographical features lacked the ability to combine maps with remotely sensed data and other forms of spatial data. The capability provides the technical foundation for the discipline of geography to fulfill its promise as the bridge linking various sciences, physical as well as social. -from Author

Geographic information systems and remote sensing

Chapter

Feb 2011

A Geographic Information System (GIS) is a collection of hardware and software for the storage, retrieval, mapping and analysis of geographic data (Table 4.1). The spatial data held in a GIS refer to a particular place on earth, represented in a coordinate system such as latitude and longitude (i.e. the data are geo-referenced). GIS data are of two general types: vector, with points, lines (arcs) and polygons as fundamental components; and raster, where the data are held as a grid of pixels. One of the most common sources of raster data for a GIS is satellite remote sensing. We may define remote sensing in a variety of ways but it refers to the process of acquiring information about the Earth's surface without being in contact with it (e.g. from aircraft or satellites). Remote sensing involves recording reflected or emitted radiation and processing, analysing and applying the output to a particular problem.

Attention and performance VIII

Article

Sep 1971

A. F. Sanders

Advanced spatial analysis: the CASA book of GIS

Article

Jan 2003

Book description: Describing the latest developments in GIS applications at the Centre for Advanced Spatial Analysis (CASA) at the University College, London, this book demonstrates how CASA is advancing spatial decision systems and spatial analysis, which are essential to solving problems and better understanding how people live. How these systems and analyses are drawn from archaeology, architecture, cartography, computer science, environmental science, geography, planning, remote sensing, geomatic engineering, and transport studies is explained. Highlighted are projects such as "Digital Egypt," which describes virtual reality reconstructions for Egyptian archaeological finds, and "Virtual cities," which explores the concepts and nature of virtual cities, from early CAD models to the newly emerging data-rich cities that merge GIS with three-dimensional visualization.

Representations of Space and Time / D.J. Peuquet.

Article

Donna J. Peuquet

Article

Pre-service teachers’ perspectives on using scenario-based virtual worlds in science education

December 2011 · Computers & Education

Shannon Kennedy-Clark

This paper presents the findings of a study on the current knowledge and attitudes of pre-service teachers on the use of scenario-based multi-user virtual environments in science education. The 28 participants involved in the study were introduced to Virtual Singapura, a multi-user virtual environment, and completed an open-ended questionnaire. Data from the questionnaire indicated that gender ... [Show full abstract] and current computer game use were likely to affect the perceived benefits of using virtual worlds in a classroom setting. Behavior management was seen as being a constraining factor on a pre-service teacher’s willingness to use a virtual world in the future. Overall, the results of the study indicate that pre-service teachers as a result of their use of Virtual Singapura are both aware of virtual worlds and have a reasonable understanding of both their potential advantages and disadvantages within a classroom setting.

Article

Your virtual avatar has all your best moves

August 2007 · The New Scientist

Matthew Busse

Technology that captures a person's full body and movements in real time and injects them into a virtual environment could revolutionise computer games and virtual worlds

Article

Full-text available

From Chinese Gardens to Virtual Environments. A Gateway to Cyberspace

October 2003

Chinese gardens provide situated portals through which poets or artists can enter their imaginary worlds. Similarly, a visual interface (virtual environment) of a cyberspace provides entrances to this potentially infinite space. Derived from design principles of Chinese gardens, we propose a design method for creating virtual environments. We use this method to design and visualise several ... [Show full abstract] cyberspaces, including web sites, virtual Chinese gardens and 3D computer games. We conduct empirical studies and find virtual environments, created by the proposed method, provide users with experience correlated to the expected result.

Article

Full-text available

Spatial data processing for the purpose of video games

January 2016 · Polish Cartographical Review

Advanced terrain models are currently commonly used in many computer games. Professional GIS technologies, existing bases of spatial data and cartographic methodology are more widely used in their development. This allows for achieving a realistic model of the world. On the other hand, the so-called game engines have very high capability of spatial data visualization. Preparing terrain models for ... [Show full abstract] the purpose of computer games requires knowledge and experience of GIS specialists and cartographers, although it is also accessible for non-professionals. The authors point out commonness and variety of use of terrain models in computer games and the existence of a series of ready, advanced tools and procedures of terrain model elaboration. Finally the authors describe the experiment of performing the process of data modeling for “Condor” - a flight simulator game.