Real Learning in Virtual Worlds - CHAPTER 2: Literature Review

From RiskWiki

(Difference between revisions)
Jump to: navigation, search
m
Current revision (00:34, 28 October 2018) (edit) (undo)
m (2.8.4.2 ACSII Virtual Worlds)
 
Line 691: Line 691:
MUD1 had a significant impact on virtual world design and development that dominated the online game space until the mid 1990s therefore MUD1 is often marked as the beginning of the first generation in online virtual worlds (Bartle, 2003). MUD1 can still be played online today at british-legends.com (CompuServe, 2007).
MUD1 had a significant impact on virtual world design and development that dominated the online game space until the mid 1990s therefore MUD1 is often marked as the beginning of the first generation in online virtual worlds (Bartle, 2003). MUD1 can still be played online today at british-legends.com (CompuServe, 2007).
-
====2.8.4.2 ACSII Virtual Worlds====
+
====2.8.4.2 ASCII Virtual Worlds====
In the early 1980’s pseudo graphical interfaces were added to some MUDs in the form of ASCII virtual worlds. ACSII (American Standard Code for Information Interchange) is the most widely adopted character encoding on western computer systems. ASCII virtual worlds provided a pseudo-graphical display making use of shape symbols and character positioning escape sequences to create crude planar maps of the terrain (dungeon) environment. The maps enhance the description of the room provided by the text.
In the early 1980’s pseudo graphical interfaces were added to some MUDs in the form of ASCII virtual worlds. ACSII (American Standard Code for Information Interchange) is the most widely adopted character encoding on western computer systems. ASCII virtual worlds provided a pseudo-graphical display making use of shape symbols and character positioning escape sequences to create crude planar maps of the terrain (dungeon) environment. The maps enhance the description of the room provided by the text.

Current revision

Contents

CHAPTER 2: Virtual Worlds - Concepts, History, and Use in Education (Literature Review)

2.1 Introduction

Gartner (2007) predicts that as many as 80% of active internet users will have a ‘Second Life’ in a virtual world by the end of 2011. Depending on your definition of ‘virtual world’ this may seem a little ambitious. Certainly, the extent to which virtual worlds are seen to include massively multi-user online environments supporting collaborative exchange of information in shared virtual space, the prediction might prove reasonably safe. To the extent that this definition is constrained to massively multi-player online games then prediction may prove a little “braver”.


Today’s virtual worlds represent the convergence of multiple technology streams, with the latest examples of the genre representing the merger of internet, telecommunications, instant messaging, virtual reality, 2D & 3D graphics, a variety of 3D modelling technologies, spatial sound, distributed databases, spatial indexing, mapping, streaming data transmission, physics, scripting languages, object-oriented software, agent theory, artificial intelligence, networking, economic modelling, online trading systems, game theory and many, many more technologies.


While the developers of many virtual worlds are content within the game space, some virtual world developers, such as Linden Research (developers of Second Life) have ambitions to be the web platform of the future (Bulkley, 2007). To this end a number of the commercial developers of virtual worlds have joined forces with a number of major corporate consumers, systems integrators and US government bodies to explore common standards for inter-operability of virtual world platforms which is a necessary first step in moving the technologies from the isolated proprietary place they now inhabit to a world-wide shared web platform (Terdiman, 2007).


This chapter explores virtual worlds, reviews the literature considering alternative definitions, characteristics, history, key architectural features, research outcomes and applications in education. The chapter concludes with an examination of traditional education taxonomy and relates that to the virtual world context as a basis for structuring an approach to exploring education affordances offered by two approaches to education in virtual worlds.


2.2 Virtual Worlds

2.2.1 What is a Virtual World?

2.2.1.1 In Search of a Definition

“Virtual worlds are places where the imaginary meets the real”. (Bartle, 2003, p. 1)


Virtual, as defined in the Oxford Dictionary (1989) with respect to the computing context is: “… not physically existing as such but made by software to appear to do so from the point of view of the program or the user….” and defined in the virtual reality context to be “… a notional image or environment generated by computer software, with which a user can interact realistically as by using a helmet containing a screen, gloves fitted with sensors, etc.” (1997).


The term world is defined in the Oxford Dictionary (1989) as “the ‘realm’ within which one moves or lives”.


In simple terms, therefore, a ‘virtual world’ can be defined as a generated computer software realm in which a user moves, exists or lives in a manner that appears to be real to the user.


A common definition for the term ‘virtual world’ is passionately debated in the literature (see Combs, 2004; Jennings, 2007; Reynolds, 2008; Wilson, 2007). It is a term that is used to describe many types of software environments from a simple MUD (Multi User Dungeons, also referred to as Multi User Dimensions or Domains) (Bartle, 2003; Keegan, 1997; Slator et al., 2007) to a sophisticated fully immersive 3D virtual reality environment used in gaming, physical training simulators or social interaction spaces (MetaMersion; Patel, Bailenson, Jung, Diankov, & Bajcsy, 2006; Van Dam, Forsberg, Laidlaw, LaViola, & Simpson, 2000). The term virtual world can be used to describe a single user walk-through simulated environment (Dalgarno, 2004; Youngblut, 1998) or an environment such as a massive multiplayer online role playing game (MMORPG) like World of Warcarft (Bainbridge, 2007). The term virtual world is also interchanged with other terms such as - virtual environment, synthetic world, mirror world, metaverse, virtual universe, artificial world etc[2] (Grøstad, 2007).


Bartle (2003, p. 1) provides the following definition:

“Virtual worlds are implemented by a computer (or network of computers) that simulate an environment. Some -but not all- of the entities in this environment act under the direct control of individual people. Because several such people can affect the same environment simultaneously, the world is said to be shared or multi-user. The environment continues to exist and develop internally (at least to some degree) even when there are no people interacting with it; this means it is persistent.”


Therefore, using Bartle’s definition in conjunction with the Oxford Dictionary definition provided above a virtual world can be defined as:

A shared software environment (or realm) in which a person represented as a projected entity (such as an digitally projected image, text identity or other computationally representational object) moves, exists or lives in a manner that appears to be real to the person and capable of affecting that environment and, being affected by, in a manner that simultaneously effects the experiences of other entities within the environment and which generally remains persistent once the user has left the world.


The key components of this definition are:

  1. A shared environment in which a real-world participant shares a computationally generated artificial space with other real world participants and/or other computationally generated entities.
  2. The nature of the real-world participant’s projection into the computationally generated virtual space.
  3. The characteristics of the space, which establish a sense of realism to the participant.
  4. The manner and extent to which the real world participant is able to affect the shared space.
  5. The nature and form of persistence that the artificial space retains.


Throughout this section we will examine the current state of these components; the ideas and literature analysing contributing to the current expression of these concepts in the form of currently available virtual worlds. The realisation of virtual worlds in software has been (and continues to be) a rapidly evolving field continually consolidating mixed influences from a fiction, mechanical and electrical engineering, computer science, gaming theory, telecommunications, social science, commerce, religion and sociology. It is a field where advances are made as much in the act of amateur invention as in formal science, and a field in which the academic literature frequently lags the leading edge of the advances by a significant degree.


2.2.2 Recognising a Virtual World by its Features

While there is not as yet a single common set of universally accepted attributes, the literature offers a variety of feature based definitions that attempt to provide a basis for classifying whether a given application or environment is, or is not, a virtual world. Across these competing views there are some features that are most frequently repeated.


Coming from the perspective of virtual worlds as gaming platforms, Bartle (2003, pp. 3-4) proposes that a virtual world should adhere to the following conventions:

  • Physics: The world contains automated rules for the players that effect change in the world.
  • Character: The player is a part of in world experience that is represented by a character and with which they strongly identify.
  • Interactions: All interactions with the world are channelled thought the character.
  • Real-time: Interaction in the world take place in real-time.
  • Shared: The world is shared by others characters in common.
  • Persistence: The world is continuous and evolves to some degree regardless of whether or not the player is present in the word. While not present the player’s state in the game remains unchanged.


Bartle tends to use the term character, for what this thesis refers to as an avatar, and considers that the player (which will be identified as ‘the intelligence’ in this thesis) must strongly identify with that character. In the context of role playing games where the player assumes an identity not their own, this aspect of the feature list goes to recognise the effectiveness of the immersion and sense of presence the player experiences (concepts we will be exploring later), but outside of this space, where the player and the ‘character’ may be one and the same, this feature is less of a distinguishing criterion.


His use of the term Physics in the context of an application genre that may include 3D environments is perhaps a little confusing. In these spaces Physics most commonly refers to the physics engine that manages the simulation of an avatar and object dynamics in the space (such as gravity, acceleration, force, momentum and limb movement, etc). As used by Bartle, the term includes the ‘business rules’ and behaviours of the system – the rules governing all interaction, not just those simulating physical movement.


The nature of the shared space and interactive channel imply that the actions of one player affect the experience of another.


Edward Castronova (2001, pp. 5-6) proposes that a virtual world should have the following features:

  • Interactivity: Existing on one computer and can be accessed via a network (or the internet) by many simultaneous users. The actions of each user have influence on other users in the world.
  • Physicality: Users access the world by a computer, which provides a first person view of the world, the world is generally ruled by natural laws much like the real world with scarcity of resources.
  • Persistence: The world is continuous and evolves to some degree regardless of whether or not the player is present in the word. While not present the player’s state in the game remains unchanged.


Castronova’s feature requirements are essentially a subset of Bartle’s, although with the possible omission of the expectation that interaction is necessarily real time.


Sun Microsystems Inc (2008, p. 3) proposed the following common features of open virtual worlds (ie multi-user virtual worlds open to public access over the internet):

  • Shared space, allowing multiple users to participate simultaneously.
  • Users interact with one another and the environment.
  • Persistence.
  • Immediacy of the interactions.
  • Similarities to the real world rules.


We might, perhaps reject Sun’s expectation of any need to assimilate ‘real world rules’ as this would exclude many fantasy role playing games from being classed as virtual worlds, but outside from this aspect Sun’s list is essentially consistent with the views of Bartle and Castronova.


These three sources are essentially consistent with the body of the literature, making allowance for the additional attributes and some latitude in interpretation we can establish a minimum feature list that would be generally accepted:

  • The environment is shared;
  • Interaction are in real-time;
  • A person participates in the world through some form of representation with which they identify and are identified and that facilitates interaction and recognition (such as a character or avatar);
  • Interactivity in the world is channelled though the avatar;
  • Changes induced by a participant influence the experience of the space for other participants;
  • Rules govern the world and interactions are shared and commonly applied; and
  • The world is persistent.


2.3 The Avatar–The Nature of a Participant’s Projection into a Virtual World

While Bartle (2003) refers to a participant’s projection into a virtual world as a “Character”, the more widely accepted name today for a real world participant’s projection into a virtual world is an Avatar. This is the term this thesis will be adopting in this research.


The word avatar derives from avatara a Sanskrit word meaning “descent of a deity” or incarnation and utilised by the Vaishnavism religious tradition of Hinduism. The Hindi concept of an avatar is thought to originate as early as the second century B.C.E (Sheth 2002). One of the most recognised Hindu deities is Vishnu (Figure 1). In Hinduism, Vishnu, is said to have a standard list of ten avataras (collectively known as Dasavatara) with one of them said to be Buddha (Siddhārtha Gautama) the founder of Buddhism (Sheth 2002).



image:Vishnu_Hindu_Avatar_001.jpg Figure 1. Hindu Avatara

Left: Visnu (or Vishnu) Hindu deity the protector and preserver of the universe

Right: Ten avatars of Visnu (Dasavatara)

(Vivekananda Centre, 2008)


In computing terms, little has changed from the original Hindi meaning of avatar. As with Hindu avatara, the virtual world participant can be thought of as “descending” or “projected” from reality to become a computational representational in a virtual world. In virtual worlds, an avatar is generally (although not exclusively) a graphical representation of the user’s persona (Deuchar & Nodder, 2003) although it can also be a representation of a system or a function in some applications (Sheth, 2003), a simple name in the form of a text string (in some text based MUD’s) and is evolving to include virtualisations of other senses (such as aural and tactile) (S.-Y. Lee, Kim, Ahn, Lim, & Kim, 2005). The graphical representation of an Avatar was thought to originate from a networked multi-user virtual world game called Habitat in 1984 (Bye, 2008; Morningstar & Farmer, 1990). Early research seems to suggest that the use of digital avatars in virtual worlds provides the user with reduced inhibitions and dissolves social status, or reconstructs social status among users (Dede, 1995; Dickey, 2003; Rheingold, 1993).


The projected form is not necessarily a recognisable representation of the real world human form. In his or her projected form, for example, the avatar might be represented as an image of a human, an animal, an animated mechanical object, a simple name, or any form appropriate to the virtual world, and within the technical capabilities of that world’s object management systems. For example, in Eve (a space based virtual world) all avatars are space ships whereas in Second Life (a social based virtual world) an avatar can take any form (Figure 2) but regardless of appearance your avatar’s name remains the same.



image:SecondLife_Digital_Avatars_002.jpg Figure 2. Digital Avatars of Second Life (Levine, 2007)


In terms of today’s virtual worlds, and for the purposes of this research, an avatar should be thought of as a combination of a representation, an agent and an intelligence:

  1. The representation may be visual, aural, tactile or any other sense conveying the presence of the avatar to other avatars or agents in a virtual world.
  2. The agent is the library of capabilities of the avatar in a virtual world.
  3. The intelligence (or actor) provides the tactical and strategic control of the avatar, which could be artificial or natural (eg human).


In a virtual world the decisions of the intelligence are communicated to, and realised by, the agent. The consequence of the agent realising (enacting/implementing) the intelligence’s commands may result in a change in the state of both the agent and the representation, eg, in a 3D Graphical virtual world, a command to walk issued by the intelligence might result in the agent changing position and entering a movement or walking state and triggering the representation to display a walking animation (enter a walking animation state).

2.4 A Taxonomy of Virtual Worlds

2.4.1 Introduction

As might be expected, the literature contains extensive discussion of the appropriate taxa to be applied in classifying virtual worlds, and also an equal measure of disagreement among authors as to the appropriate criterion so to be applied. In spite of the range of discussions, most attempts are incomplete and therefore capable of classifying in a useable form only a portion of the genre. To be fair, this space is rapidly evolving and possibly as fast as it is classified a new entrant appears that change the paradigm, and old entrants are updated to include new capabilities.


2.4.2 A Taxon for Virtual Worlds

Outside of the education and virtual reality streams, possibly the largest single family of virtual worlds are those developed for games. While not actually claiming to propose a taxon, Bartle (2003, pp. 38-61), whose pedigree is essentially from the gaming stream, proposes a set of attributes that can be used to classify Virtual (game) Worlds. Not surprisingly, the attributes are most relevant to multi-user game focussed virtual worlds, but provide a workable superset of the current thought on the matter and with some adjustment can be extended to the more general examples of virtual worlds. He suggests that a virtual world can be categorised according to the following taxa:

  1. Appearance: To a ‘newbie’ (Bartle’s term for a new user of a virtual world application) the distinction is whether the virtual world is a ‘text based’ MUD, ASCI, graphical 2D or graphical 3D etc. To an ‘oldbie’ (as described by Bartle) this is only an interface issue and therefore not as important as the other listed categories.
  2. Genre: Is the world fantasy, cyberpunk, horror, social etc. The plot or the settings of the virtual world. This taxon is most helpful with purpose focussed virtual worlds. In the non-gaming or semi-gaming space occupied by some generalised social worlds, the virtual world is as much a platform on which other ‘sub-worlds’ can be based, and thus the genre of the virtual world can be all other genres. Examples of this might include PLATO and Second Life.
  3. Codebase: Although not as important for the user as it is hidden from them this is an important aspect to the designer of a virtual world. The codebase defines the technical makeup of the world - reusable content and controls, scripting language, database structure etc. This researcher suggests that the codebase is not a single taxon, but perhaps should be separated into multiple taxa. In its place one might propose the content management, asset management, game engine, environment application programming interface, AI, and scripting function library within the system as more relevant technical matters.
  4. Age: How long the virtual world lasts is an important aspect for the measure of success of the virtual world. Generally the longer you can keep a player (or user) interested the longer the virtual world survives which in turn attracts new users which adds to the player base of the virtual world.
  5. Player base: How large is the player (or user) base of the virtual world. This measure varies depending upon what you are counting for example, the number of registered users, the number of avatars (a user can have more than one character in a virtual world but in general not for simultaneous use), simultaneous users logged in, hours played per user, access over a period of time, number of active subscriptions, etc. In some worlds the meaningful measure of player base is in fact the number of owner occupied ‘acres’ of virtual land (as opposed to general users of the virtual world). The player base measures the current success of the virtual world, its popularity so to speak, which in turn lengthens the age of the virtual world. Given the number of ways a player base can be structured and measured a single measure is open to both misinterpretation and reporting manipulation, and for some measures (like subscribed users – where some subscriptions are costed and others free) may be completely erroneous when comparing one virtual world to the next.
  6. Degree to which they can be changed: Virtual worlds vary in the degree to which a user can change the content or add to the content of the virtual world. Virtual worlds such as World of Warcraft (and most game based virtual environments) allow no change by the player with all content created by the developers of the virtual world. Other virtual worlds such as Second Life, Active Worlds, TruePlay and PLATO rely on content created by the community. In the case of Second Life (for example) the entire virtual world is made from user created content by providing them with building tools, import and export capabilities, out-of-world interfaces and communications capabilities, an extensive library of API functions and a scripting language. The degree to which a virtual world’s content can be changed by the user adds to the technical codebase complexity and the user’s (and other user’s for multi-user virtual worlds) experience of and within the virtual world.
  7. Degree of persistence: Bartle defines persistence to be the degree to which a world’s state remains intact if you shutdown and restart the virtual world. He classifies persistence into ‘discrete’ or ‘continuous’ groups. At the extreme a discrete virtual world would regenerate - described a ‘Ground Hog’ world (named after the movie). Here all content and the location of the player would be reset to the start of play. In a continuous virtual world the content and locations are retained through a restart.
    Persistence also relates to what happens to the world when a user logs off, does the virtual world continue to evolve without the individual player – and if so can the player’s state be affected while off line? A virtual world generally displays some level of persistence and is generally a term used to distinguish if a ‘virtual world’ is really a ‘world’ or in fact just a simple ‘Ground Hog’ environment (see Gehorsam, 2003). The ultimate level of persistence being that akin to the real world which is constantly evolving and changing regardless of our existence within the World.


With some modification and generalisation most of the taxa can be applied in the general case of gaming and non-gaming virtual worlds. To be applied outside of the narrow RPG (Role Playing Game) grouping, the classification system would benefit from some subdivision of elements.


We have already noted codebase as one such category. Codebase is such a wide group that is could be applied to every functional capability of the virtual world not covered by another taxon, and thus is of limited help in establishing a consistent framework for classification. For example Castronova (2001) taxonomy recognises a grouping under marketplaces (implying commercial functionality) while both Kish (2007) and Cavazza (2007) recognise groupings covering Paraverses (although they use different terms). In Bartle’s taxa these might both be covered as distinguishing characteristics under codebase, yet the one relates to the ability to conduct real-world commercial transactions in the space, while the other addresses the merging of real-world content with virtual world content.


Persistence as framed by Bartle mixes up multiple discrete concepts – host state persistence, user state persistence, environmental evolution, and scenario persistence. This last item is generally typical of games (such as quest driven environments where on restarting a ‘quest’ the user can rely on the sequence of events being a repetition of the sequence that occurred previously – effectively a ground-hog space within a larger persistent environment), and absolutely essential for simulators and learning systems where a user taking a course should be able to rely on the lesson replaying in a consistent and predictable way each time (unless variation is an intended part of the training like in a military battlefield virtual world). In order to classify virtual worlds, recognising these attributes independently of each other would be more helpful than identifying the world as persistent or not persistent, nor are the sub-features linearly related – i.e. one form of persistence does not imply the inclusion of another form of persistence (Purbrick & Greenhalgh, 2002).


2.4.3 Applied Taxonomies

While Bartle proposes a reasonably extensive set of attributes (taxa) for classification, some authors have proposed simpler classification regimes, although all seem as yet to avoid claiming an actual taxonomy.


Kish (2007) recognised that with the appearance of the weakly defined ‘Web 2’ technologies, virtual worlds could be seen to encompass a wider range of social networking and world-imagining spaces. Kish’s classification groups virtual environments into the broad categories (Figure 3):


  1. MMORPGs: Massively Multiplayer Online Role Playing Games. A category which includes text and graphical gaming environments with the common theme of role playing and containing internally a hierarchical, level based player grading system to determine expertise and implied seniority, and generally plot or quest driven and goal oriented as their linking characteristic. Typical examples might include World of Warcraft, Entropia Universe, Everquest, MUDs, etc.
  2. Metaverses: Imagined public fantasy spaces, emphasising social interaction, creativity and lacking a single plot or purpose for participation. Generally exhibiting a devolved structure without a single levelling system or clear environment imposed hierarchic seniority system[3]. Typical examples might include Habitat, Second Life, Active Worlds, Furcadia, etc
  3. Paraverses: Spaces that intersect with the real world, incorporating content from the real world and thus could be described as virtual extensions of the real world. This group potentially includes many of the Web 2 spaces that contain sufficient functionality to create in the minds of their users a ‘real’ virtual community as strongly present to the participant as their real world existence.
  4. Intraverses: Spaces that are otherwise Metaverses or MMOLE’s but private or closed to the broader public. Virtual reality environments could be seen generally to fall into this category as well as private/corporate implementations of public virtual world spaces. Typical examples might include Qwaq, Sun System’s Wonderland, IBM’s Metaverse, etc.
  5. MMOLEs: Massively Multi-user Online Learning Environments. Possibly the oldest class of virtual worlds as it includes systems such as PLATO and is typified by educational environments supporting user social interaction. Primarily purpose (or although not goal) driven – such as learning, training, idea exchange, simulation, etc. This space includes the dedicated training / teaching environments of PLATO and planning / simulation management systems of SIMNET, Blackboard, Boston College’s Media Grid, etc.


image:Kish_Virtual_Geography_003.jpg Figure 3. Virtual Geography (Kish, 2007)


Cavazza (2007) proposes that a virtual world should be open (public) and contain taxa supporting strong and generalised capabilities in each of the dimensions (Figure 4):

  1. Social networking
  2. Gaming
  3. Entertainment
  4. Business


image:Cavazza_Virtual_Universes_Landscape_004.jpg Figure 4. Virtual Universes Landscape (Cavazza, 2007)

Consequently most of the virtual worlds identified by other authors are excluded from Cavazza definition of virtual worlds, but included under the broad category of ‘Virtual Universe’. To illustrate this idea Cavazza has classified a huge range of the existing virtual environments:

  1. Social
    • 2.5 & 3D Chats
    • Avatar Centric
    • Social Platforms
    • Branded Universe
    • Virtual Worlds
  2. Game
    • MOG
    • Sports
    • MMORPG
    • Avatar Centric
    • Social Platforms
    • Branded Universe
    • Adult Games
    • Virtual Worlds
  3. Entertainment
    • Virtual Sex
    • Virtual City Guides
    • 2.5 & 3D Chats
    • Avatar Centric
    • Branded Universe
    • Virtual World Generators
    • Virtual Worlds
  4. Business
    • Serious Games
    • Virtual Marketplaces
    • Adult Games
    • Virtual World Generators
    • Virtual Worlds


Cavazza’s definition and classification system is extensive, and possibly the most comprehensive to date. While Kish’s classification tends to focus on functionality, Cavazza’s emphasises purpose. Never-the-less, there is significant cross-over in their ideas. For example, both recognise the difference between games and social networking, and both accommodate the paraverses in a special category (Cavazza includes them in ‘Virtual City Guides’ among other groups). Cavazza’s analysis, however, lacks the accommodation of the education, training and simulation virtual spaces present in Kish’s categorisation, although, it might be argued that these are covered in multiple categories including ‘Virtual World Generators’ (eg PLATO, VastPark) and Serious Games (training simulators).

2.5 What’s in a Name? – Virtual Worlds versus Virtual Reality

Virtual Reality environments are generally a combination of user interface hardware (such as headsets and data gloves) and software. The availability of the (often costly or purpose built) user interface hardware has meant that the majority of these environments are either single user or very small scale multi-user environments (Jones & Hicks, 2004; Miller & Thorpe, 1995). A direct consequence of this is that Virtual Reality environments have tended to ignore the dimensions of user interaction, game play and collaboration in favour of the technology of immersion. This fact, possibly more than any other, has predisposed some authors to exclude virtual reality spaces from the domain of virtual worlds (Bartle, 2003; Yee, 2006).


While Bartle’s virtual world definition, contributes part of the definition we have adopted for virtual worlds in this research, the researcher departs from the entirety of Bartle’s embodiment of virtual worlds as expanded in that work. Bartle believes that a virtual world has a meaning divergent from that of virtual reality believing that “Virtual reality is primarily concerned with the mechanism by which human beings interact with computer simulations… [rather than] the nature of the simulations themselves” (2003, p. 3). To this extent Bartle’s definition specifically excludes Virtual Reality spaces from the definition of virtual worlds.


This researcher adopts a view consistent with some other writers in the field that excluding the body of work in virtual reality from the concept of a virtual world by writing virtual reality spaces out of the definition, places the emphasis narrowly on the social and gaming dimensions of these worlds, and away from the immersive experience thus excluding the vast body of research that predates or has been done in parallel to the development of gaming virtual worlds (Cosby, 1999; Heilig, 1955; Pimentel & Teixeira, 1994; Rheingold, 1992; Schroeder, 1997; Steuer, 1992; Sutherland, 1965; Walker, 1990; Woolley, 1994) and constrains the consideration of these environments in the education context to their collaborative and scripting capabilities.


Other authors have adopted definitions wider than that posited by Bartle of the virtual world concept, although in most cases constrained from some portion of the body of work that has contributed to the space. Dickey (2005, p. 439) implies an exclusion of 2D and non visual environments while providing: “Three-dimensional virtual worlds are a networked desktop virtual reality in which users move and interact in simulated 3D spaces.” Similarly, McLellan (2004) presents 10 classifications of virtual reality, a single virtual world being classified as ‘through the window’ where as a multi-user virtual world would be classified as ‘cyberspace’. Mazuryk and Gervautz (1996) make no distinction in the number of users in the virtual world but define a virtual world to be a ‘desktop VR (virtual reality)’ or a ‘Window on World (WoW)’ system. Biocca and Delaney (1995) defines a virtual world to be a ‘window system’ a computer generated three-dimensional virtual world viewed either by a computer screen or with the assistance of a head mounted display.


This researcher’s view is that all of these definitions are correct, but incomplete and that a definition that allows the participation of all of these examples is the most useful and appropriate in the education context. To appreciate the reasoning behind this argument we must look at some of the history of the development of the technologies and concepts that have contributed to the current family of virtual worlds and the problems and purposes these stepping-stones intended to resolve or achieve.


Authors adopting Bartle’s view have generally also adopted the view that virtual reality is essentially a hardware interfacing technology and hence the environments managed in this space are of no consequence. The misconception that virtual reality is a collection of hardware (data glove, head mounted displays etc) neglects the very meaning of virtual reality, which seeks to evoke a feeling of immersion and presence within the virtual space. In virtual reality research stream, using external hardware devices to enter a virtual world is only one method by which immersion and presence is achieved (Briggs, 1996; Steuer, 1992). No external device will ensure a user’s experience of immersion if the world they enter is an unconvincing generator of an alternative reality for the participant. Furthermore, if virtual reality is to be excluded from the scope of the definition of virtual worlds, then the existence of VR plug-and-play devices such as stereoscopic headsets, data gloves or haptic controls that are readily available to use with many mass market virtual worlds (that otherwise would fall within Bartle’s definition) for example, Vuzix iWear headset, Evolution Motion Glove of PS1, Wii Remote for Nintendo Wii, MS Force Feedback controller for Flight Simulator etc. would seem to contradict the proposed disconnect between the study of virtual worlds and virtual reality. Lastly the exclusion of virtual reality environments from the definition of virtual worlds ignores that fact that in the 3D virtual world space many of the technologies and concepts utilised were contributed by the virtual reality research stream (as will become clear from the history presented in the following sections).


In the education context, virtual reality technologies (as expressed for example in simulators) are a critical and essential contribution to the pantheon of virtual (training) worlds (Bailenson et al., 2007; Dede, 2004). In this researcher’s view, virtual reality environments are a subset of the virtual worlds, which are increasingly converging, if the space has not already converged in current virtual world examples such as America’s Army, Second Life, etc and massive multiplayer training environments like SIMNET (Lang, Maclntyre, & Zugaza, 2008; Lenoir, 2003; Zyda, 2005).


2.6 Dimensioning Virtual Worlds

2.6.1 The Degree of Virtuality

The degree to which a world is ‘virtual’ can be looked at as a sliding scale between physical and virtual. Milgram and Kishino (1994) presents a taxonomy for mixed reality visual displays called a ‘reality-virtuality continuum’ (Figure 5). On the left hand side of the scale is the ‘real environment’, which is equivalent to the real or tangible world, while on the extreme right is the ‘virtual environment’, which is equivalent to an artificially generated world. Between these two extremes is classified as ‘mixed reality’ (MR) made up of combination of both real and virtual matter.[4]


image:Reality_Virtuality_Continuum_005.jpg Figure 5. Reality-Virtuality Continuum: Representation Scale for Visual Display

(Milgram & Kishino, 1994)

Figure 6 illustrates an example of the use of the reality-virtuality continuum taken from the MagicBook Project (Billinghurst, Kato, & Poupyrev, 2001). On the left of the figure is a book that is real (ie. the real world environment); in the middle the same book but viewed though an Augmented Reality (AR) Display where figures appear like pop-up characters on top of the book (ie. mixed reality or augmented reality); while on the right the same book but viewed within a virtual environment where the “reader” becomes the characters within the book.



image:The_Magic_Project_006.jpg Figure 6. The MagicBook Project: An Example Of The Full Reality-Virtuality Continuum

While the MagicBook project was conceived around the integration of physical (tangible) real world objects with digital virtual world generated objects, when the real world objects are themselves digital or intangible – such as with course materials of photographic images, text, or other digital content the merging of the ‘Real World’ and the ‘Virtual World’ become less obvious. For example, real world authors Pamela Woodard and Wilbur Witt have published their works in the Second Life virtual world first or simultaneously with publication in the real world (Bell, 2006). Second Life virtual world can integrate conventional HTML web page content directly into the virtual environment (Release Candidate, 2008). Content developers and particularly trainers and presenters in Second Life routinely import textures and slides and stream sound and video from outside of the virtual world into the virtual space.


In the context of Milgram and Kishino’s reality-virtuality continuum, this research focuses on the right hand end of the scale i.e. using a desktop display of a virtual world in which all content is delivered virtually. In contrast to the MagicBook project this research considers (in the education context) the affordances from two virtualisation strategies – a direct reproduction of the real world delivery into the virtual (in part, by importing the non virtual world generated materials into the virtual world), and a transformation of the real world material into virtual material (in part, by recasting the non virtual world materials into virtually generated form).

2.6.2 The Degree of Immersion and Presence

2.6.2.1 Introduction

Virtual reality literature often separates a user’s experience of a virtual environment into physical and psychological components (Benford, Greenhalgh, Reynard, Brown, & Koleva, 1998; Biocca & Delaney, 1995; Sheridan, 1992; Mal Slater, 1999; Mal Slater & Wilbur, 1997; Steuer, 1992). The psychological components include the interaction (or connectedness) and belief where contribution of the participant or their willingness to believe in the reality of which they would otherwise know to be unreal and the physical aided by external mechanical and functional capabilities of the system.


In exploring the factors determining the effectiveness of Virtual Reality environments, Burdea and Coiffet (2003) determined that the aim of virtual reality is to achieve a trio of ‘Immersion, Interaction and Imagination’ (Figure 7. The Three I's of Virtual Reality), each of which holds equal significance to the user’s experience of virtual reality systems. A virtual reality system seeks fully to engage the user in the virtual space. They proposed that excluding any one of these features exposed a user to passive participation, and ultimately detracted from the perceived ‘reality’ of the experience.



image:Immersion_Interaction_Imagination_007.jpg Figure 7. The Three I's of Virtual Reality

Slater (1992) defined user involvement to be a combination of the human experience which in turn is dependent on the technology (Figure 8). Telepresence (or presence) is the human sensation of ‘being there’ in a virtual environment[5] and seen influenced in part by the technology in terms of vividness (richness, realism) and interactivity (response) of the environment.



image:Steuer_Variables_Influencing_Telepresence_008.jpg Figure 8. Technological Variables Influencing Telepresence (Steuer, 1992)

Slater and Wilbur (1999; 1997) revisited these concepts in later work, defining a user’s experience in terms of immersion and presence. Immersion is seen as an objective measure of ‘systems immersion’ technology such as field of view, quality of display etc and while presence is seen as a subjective measure, a psychological sensation of ‘being there’. From here on we will be using the terms immersion and presence as defined by Wilbur and Slater.

2.6.2.2 Immersion

Benford et al. (1998) propose classifications of artificiality and transportation for collaborative environments (Figure 9) that extends Milgram and Kishino’s reality-virtuality continuum. Artificiality (physical-synthetic) is equivalent to the reality-virtuality continuum. Transportation (local-remote) is the degree to which a participant becomes removed from their local space to operate in a remote space, which they define to be a similar to the concept to immersion. For example, CVEs (Collaborative Virtual Environments[6]) are placed on a scale of partial to remote transportation where a fully immersive CVE would be the ultimate level of transportation in a virtual reality system using devices such as HMD, data gloves, tactical and aural equipment that allowed for no outside distraction, the participant would be operating completely within virtual environment and be fully remote form their local environment[7]. Whereas, a desktop CVE is partially immersive as ones local surroundings form a part of the virtual environment eg field of view that allows for head turning away from the virtual space etc (Sheridan, 1992). In the context of Benford et al. transportation scale this research is conducted using desktop CVEs and is therefore only partially immersive according to their scale.


image:Artificiality_Transportation_as_SS_Metrics_009.jpg Figure 9. Shared Space Technology According to Artificiality and Transportation

2.6.2.3 Presence

Research in online gaming virtual worlds has tended to focus on the human experience (presence) of virtual worlds rather than the ‘systems immersion’ aspects, while studies of virtual reality environments have tended to consider both. This is possibly a function of the common standard interface for massively multiplayer game environments that has traditionally been the desktop computer equipped with a mouse and keyboard. Although various more advanced mass market input devices (head mounted displays and 3D mice, etc) have been available to the mass-market for many years, they are not yet widely utilised.


The degree of presence is often linked to the effectiveness of a virtual environments (Witmer & Singer, 1998) which due to its subjective nature is possibly the most difficult to comprehend and therefore measure (Mal Slater & Usoh, 1993). Hence, this area has been a widely researched with various explanations as to what constitutes presence in a virtual environment (Schuemie, Straaten, Krijn, & Mast, 2001). The sense of ‘being there’ in the environment is subjective as Slater and Usoh (1993; 1994) describe presence is similar to a person’s ‘willingness to suspend disbelief’, a concept derived from British poet and literary critic Samuel Coleridge (1772-1834) in his autobiography (1817) where he describes the phenomena of when a person becomes so engaged in a narrative that they are willing to believe an event is true if even for only a brief moment. Although suspension of disbelief is often linked today with mediums such as film and literature, virtual worlds (especially Role Playing Game (RPG) worlds) provide many of the same traits in which the user can be thought of as an actor within the virtual world that forms a part of the storyline.


A number of presence classification strategies have been proposed by various authors. We will consider:

  1. Schroeder - focussing on the importance of social interaction
  2. Bartle – focussing on the degree of commitment in the environment


Schroeder (2006) presents presence in a continuum of shared virtual environments (SVE) within a three-dimensional model (Figure 10). Presence (x), copresence (y) and connected presence (z) can be described respectively as ‘being there’, ‘being there together’ and ‘being connected together’. Connected presence can be thought as the extent to which a relationship is mediated when presence and copresence exist. Mapping is done on a comparison with a physical face-to-face relationship (0,0,0) and an entirely immersive environment such as a networked Cave (1,1,1). For example, face-to-face is (0,0,0) there is no presence (and thus no copresence) as no meeting is taking place in a virtual environment whereas in the case of a networked Cave (1,1,1) the entire relationship (and environment) is virtual where affordances are such for high connected presence.



image:Presence_Copresence_Connected-Presence_010.jpg Figure 10. Presence, Copresence, and Connected Presence

In different media for being there together

Of interest in Schroeder’s model is the comparison of desktop SVE and online computer games. The example given in the model for a desktop SVE is Active Worlds which is a massively multiplayer online (MMO) social virtual world and the example provided in his paper for an online game is Quake, which at the time provided for up to 16 players sharing a common space. Both are virtual worlds, use text chat and sound, and use avatars to project the participant into the virtual world (although Quake takes a first person view exclusively). For the purpose of the analysis the main differences were perceived as the number of simultaneous players sharing the common virtual space and the imposition of clear game driven objectives in Quake, and the absence of those same game driven objectives in Active Worlds. Yet, Active Worlds was seen as providing a higher level of connected presence. So why does Active World provide a higher level of connected presence? The distinction here was seen to be the in the concept of the ‘game’ rather than number of players when you compare it to other SVEs presented in this model above. Active World is a social world in which no plot is provided to measure success or failure of an individual, unlike Quake where the measure of success is clear and the entire activity and function of the environment is the relentless pursuit of that individual success. Therefore it was deduced that a social (game) world provide for more connected presence than that of an individually focussed plot driven gaming virtual world (at least as analysed by Schroeder).


Schroeder observation of higher connected presence in social virtual worlds seems to fit with Heeter’s (1992; 2003) definition of social presence where she defines presence in terms of individual presence, social presence and environmental presence. Presence of an individual is increased when social relationships are formed which is based upon the social component of perceptual stimuli. When an environment or situation is focused on the relationship (rather than killing a monster like in RPGs) a higher social presence will be achieved.[8]


Bartle (2003, p. 42) identifies a system of levels of immersion (which in this paper we have defined as presence[9]) based upon a linear scale of the; Player (the real person), Avatar (the digital puppet), Character (representation in the world e.g. character name, role etc) and Persona (your identity in the virtual world where the player is the Character and is in the virtual world). Persona is similar to the concept presence, if your character is killed ‘you feel like you have died’ there is no distinction between the character and the player, they are one, the Persona. Bartle believes that the avatars and character are just steps along the way to persona. Persona is when a person ‘stops playing the world and starts living in the virtual world’.

2.7 Influences on Virtual Worlds from Art and Literature

2.7.1 Introduction

The concept of a virtual world is by no means unique to computing. The thought of exploring an imaginary realm has captivated people’s imagination throughout time.


“If we define that a virtual world is a place described by words and/or projected through pictures, which creates a space in the imagination real enough that you can feel you are inside of it, then the painted caves of our ancestors, shadow puppetry, the 17th-century Lanterna Magica, a good book, play or movie are all gateways to virtual worlds. Humanity’s most powerful new tool, the digital computer, was also destined to become a purveyor of virtual worlds, but with a new twist: The computer enables the virtual world to be both inhabited and co-created by people participating from different physical locations.”(Damer, 2007, p. 2)


At least with respect to the massively multiplayer online virtual worlds/role playing games (MMOVW, or MMORPG), all of today’s exhibits can trace their paradigms to literature. Some such as Eve, Entropia Universe, World of Warcraft are amalgams of a body of works and ideas while others such as MUD1 (Sword of the Phoenix (Howard, 1932)) and Second Life (Snow Crash (Stephenson, 1992)) are direct inspirations of specific literary works.


Consequently, to properly understand the ‘state of the art’ represented by today’s multi-user, connected together, virtual worlds and the gaming, social and business rules they have adopted to govern them, it is essential to consider the context from which they have been derived, and the art that has influenced their creators. While some operational paradigms in virtual worlds are technology constraints, functional capability constraints can be as much a condition of the imagined world being implemented as a real constraint of the technology of the day. To appreciate this fact one need only compare the camera controls of Project Entropia versus those of Second Life – two environments of comparable age, or the commercial capabilities of these two environments versus those of World of Warcraft. In each case the differences and apparent restrictions are a game design decision rather than a technology constraint.


2.7.2 Virtual Worlds of the Arts

James Pearson (2002) believes that from as early as 30,000 years ago in the Chauvet Cave in France shaman used cave art as a means to document their experiences of travel to the dream world. Packer and Jordan (2002) also draw this similarity in their book on virtual reality describing how the Cro Magnon in 15,000 BC in the Lascaux caves of south-western France used cave art (Figure 11) with candles and the acid aroma of animal fat for a magical theatre of the senses.



image:Cave_Art_BC_011.jpg Figure 11. The caves of Lascaux: Cave Art 15,000 BC

The German composer Richard Wagner (1813-1883) (Figure 12) concept of Gesamtkunstwerk (total artwork) has also been cited as an early pioneer in the concept of immersion and presence in virtual worlds (Grau, 1999; Klich, 2007; Packer & Jordan, 2002). Wagner believed that a “Artistic Man can only fully content himself by uniting every branch of Art into the common Artwork” a synergy that not only includes the performance but all that surrounds so that mankind “...forgets the confines of the auditorium, and lives and breathes now only in the artwork which seems to it as Life itself, and on the stage which seems the wide expanse of the whole World” (Wagner, 1849, p. 184 & 186).



image:Wagner_Gesamtkunstwerk_012.jpg Figure 12. Richard Wagner's Gesamtkunstwerk (Total Artwork)

2.7.3 Virtual Worlds of Fiction and Fantasy

There are numerous examples of virtual world that have been explored though fiction and fantasy. Each has contributed to the illusion of virtual worlds becoming a reality (Bartle, 2003; Chesher, 1994).


In Lewis Carroll’s novel, Alice's Adventures in Wonderland (1865), Alice fell down a rabbit hole to explore a fantasy world inhabited by peculiar and anthropomorphic creatures. Similarly, in Carroll’s follow on novel, Through the Looking Glass (1871), Alice explores a world behind a mirror. Hattori (1991) saw Lewis Carroll’s novels as a paradigm for modern virtual reality systems (Figure 13) blending the physical space with fantasy in a rapidly changing environment. To this extent, Carroll’s works provide a perfect analogy for the design and the development of virtual worlds (Rosenblum, 1995; West Virginia University, 2008). An explorative virtual world was realised as a children’s computer game called The Manhole (1988-2007) where it was based upon Carroll’s novel Alice’s Adventure in Wonderland (Wikipedia, 2008a).



image:Alice_via_Caroll_and_Hattori_013.jpg Figure 13. 'Through the Looking Glass' Carroll (1871) & 'The World of Virtual Reality' Hattori (1991)


Within the fantasy literary genre, a key influence has been the works of J R R Tolkien starting with The Hobbit (1937) and its sequel The Lord of the Rings (1954, 1955) (Figure 14). An adventure fantasy that takes place in an imaginary world called Middle-Earth containing races such as Hobbits, Wizards, Elves, Orcs, Dwarves and Trolls. Tolkien’s literature style was so popular that the Oxford dictionary termed his literature approach as tolkienesque[10].



image:JRR_Tolkein_Book_Covers_014.jpg Figure 14. The Hobbit & The Lord of the Rings by J. R. R. Tolkien (1937, 1954, 1955)

With respect to today’s virtual worlds, Tolkien’s contribution has not been merely in the construction of a raft of characters, racial groups and social concepts for role playing game inhabitants and interaction rules, but most importantly in his deep backgrounding of the imagined worlds. He did not merely describe his characters within the context and flow of the story line, but extended beyond that which was needed to tell a story, into that which was needed to make us believe of the real existence of his virtual worlds, Tolkien provides the reader with immaculate detail and descriptions to immerse them into the world Middle-Earth. Both books contained land maps (Figure 14) and the final book to The Lord of the Rings (released in 3 parts) containing appendices describing chronologies, histories, family trees, languages and translations and a calendar and dating system. Being a professor at Leeds and Oxford University he approached his work more like an academic anthropological study of an imagined world than a novelist (Macmillan, 2008).


In so doing Tolkien demonstrated a fundamental understanding a core strategy in establishing convincing presence – the necessity for a consistent, credible back story underpinning the virtual world. It is an early example of the depth of design that many later virtual worlds would exhibit in order to create a convincing sense of presence for the participant (Bartle, 2003; Schmidt, Kinzer, & Greenbaum, 2007).


A couple of virtual worlds that has been translated from Tolkien’s literature are the online virtual world ‘Lord of the Rings Online’ (2007) and PLATO’s MUD virtual world ‘Mines of Moria’ (1974).


More recently, literature has turned to imagining realities in which computational virtual worlds are a fundamental component of the plot. It is from this group that many of the terms now used to describe aspects and elements of virtual worlds are derived or were popularised, such as ‘avatar’, ‘metaverse’, ‘cyber-space’, etc. Some recent examples of novels containing a plot of computation virtual world are True Names (Vinge, 1981), Neuromancer (Gibson, 1984) and Snow Cash (Stephenson, 1992) (Figure 15).



image:Recent_VR_Literature_Covers_015.jpg Figure 15. Recent Literature True Names (Vinge, 1981), Neuromancer (Gibson, 1984), Snow Cash (Stephenson, 1992)


Vernor Vinge’s True Names is not as well know as other novels in this genre but it was the first to present the concept of a person entering a computational virtual worlds meeting other people in ‘the other plane’ (Kelly, 1995). It was also unique in bringing the concept of anonymity to the digital world with one’s digital persona (handle) being different from one’s real self and where there was a necessity to hide one’s real identity thus your true name (and hence the title). It was translated to the computational virtual world in the form of ‘Habitat’ – the first graphical social networking virtual world (Farmer, 1992).


William Gibson’s Neuromancer a true cyberpunk[11] novel is possibly the most widely quoted in the virtual environment space (Chesher, 1994) . In this novel Gibson coined the term cyberspace with the concept of a viable parallel online world capable of critically impacting events and commerce in the real world.


Neal Stephenson's Snow Crash is where the term Metaverse was coined. Metaverse is a planet-sized city that has one continuous street 65,536 kilometres (216 km) in length where millions of people (known as avatars) travel up and down daily in search for entertainment, trade or social interaction. Although similar, in one sense, to Neuromancer it came from a different perspective in that people actually lived in the Metaverse not as cyberpunks getting up to mischief but as everyday people living a mainstream life real life in the virtual world. In this world real commerce was conducted and virtual artefacts were bought and sold with real world consequences which has been realised in the development of the virtual world Second Life.


Hollywood also contributed to the fantasy of the reality of virtual worlds. Films such as Tron (Lisberger, 1982), The Lawnmower Man (Leonard, 1992) and The Matrix (Wachowski & Wachowski, 1999) (Figure 16) just to name a few gave us the visual of virtual worlds that the books could only describe, and in some cases explored the haptic interfaces now being realised (Chesher, 1994).



image:VW_Films_Tron_LawnmowerMan_Matrix_016.jpg Figure 16. Hollywood Films

Tron (Lisberger, 1982), The Lawnmower Man (Leonard, 1992), The Matrix (Wachowski & Wachowski, 1999)


At the time of their release, the novels and movies discussed above may have seemed futuristic and the concepts unobtainable but today we are much closer (if not already past) with advances in networking, computational processing power and understanding of the sociology of virtual environments. Maybe a ‘jack-in’ device that stimulates our nervous system to travel into cyberspace (Neuromancer, Gibson, 1984) is still a little way off (and may be too intrusive for some), or smelling odours or feeling textures within a virtual world may never be quite the same as the real life experience but what once seemed unimaginable in these works has become reality today. With technological advances and the rapid adoption of internet enabled online virtual worlds many of these concepts are less science fiction and more science fact than they once were.

2.8 The History of Computational Virtual Worlds

2.8.1 Introduction

In a lecture delivered by Ivan Sutherland in 1965 the first steps were made to combine computer design, construction, navigation and habitation of software generated virtual worlds (Packer & Jordan, 2002). Here Sutherland laid down a vision for the development of virtual worlds, as paraphrased by Brooks (1999, p. 16):

“Don’t think of that thing as a screen, think of it as a window, a window through which one looks into a virtual world. The challenge to computer graphics is to make that virtual world look real, sound real, move and respond to interaction in real-time and even feel real.”


The new-born medium of the graphical, digital virtual world experienced a “Cambrian Explosion” of diversity in the 1980s and ‘90s, with offspring species of many genres: first-person shooters, fantasy role-playing games, simulators, shared board and game tables, and social virtual worlds. (Damer, 2007)


The massively multiplayer online virtual worlds of today, with their world-wide user bases, are essentially a consequence of the mass adoption of the internet which commenced in the early 1990’s. Since the internet first achieved general acceptance they have advanced substantially in technical capabilities, graphics and number of subscribers (Figure 17) (Woodcock, 2008). See Appendix B: MMOG Analysis, for a break-down of MMOGs contained in this graph.



image:MMOVW_Growth_Rate_017.jpg Figure 17. Massive Multiplayer Online Virtual World Growth Chart 98-2008


The virtual worlds of today (such as World of Warcraft, Entropia Universe, America’s Army, and Second Life, etc) represent a convergence of several disparate computational, technical and social origins and drivers. Current virtual worlds combine 3D visualisation, game theory, text messaging, animations, context and text sensitive gesturing, natural language processing, spatial voice & audio, artificial intelligence, agency theory, physics, connectedness, persistence, business strategy, sensory hardware and haptic interfaces, telecommunications, 2D image processing, video chroma-keying, social networking and many other influences to achieve their sense of immersion and presence. In this section we explore some of the milestones along these convergent paths.


As many of the influences that have contributed to our latest virtual world are derived from research streams that were concurrently pursued over more than 50 years, we shall look at the history of virtual worlds in six streams:


  1. Hardware based user interfaces and virtual reality environments
  2. Early graphical computer games
  3. Text and Text+ based Virtual Worlds
  4. 2.5 and 3D graphical multi-player virtual worlds, broken down into:
    a. MMORPGs
    b. Social Virtual Worlds
  5. Simulation and Training Worlds


It should be noted that, while we will be considering the history in these streams, some virtual worlds necessarily exist in more than one stream. The grouping is that of the researcher, based on an extensive assessment of the literature, rather than the view of any one author.

2.8.2 Hardware Based User Interfaces and Virtual Reality Systems

2.8.2.1 Introduction

These two areas are grouped together, not because Virtual Reality (VR) Systems are a hardware solution, but rather because the work done in virtual reality worlds has generally aimed for extremely high levels of both immersion and presence and has therefore generally (although not always) been coupled with hardware in the form of purpose built user interfaces, designed to assist the sense of immersion such as headsets, or data gloves, etc.


The importance of the progress in VR systems to virtual worlds is that they have contributed or assisted much of the fundamental graphical rendering technologies, 3D animations studies and spatial awareness research and conceptualised the immersive aspects of virtual worlds.


2.8.2.2 Sensorama

One of the earliest inventions in the genre of virtual world simulators was developed by a cinematographer Morton Heilig. Inspired by the work of Fred Walker’s with Cinerama[12], Heilig presented a paper in 1955 ‘The Cinema of the Future’ (reprinted in Packer & Jordan, 2002). In an extension of Wagner’s (1849) Gesamtkunstwerk (total artwork) concept (Holmberg, 2003), Heilig believed that the logical extension of cinema was to provide the audience a first person experience of film using all their senses – “Open your eyes, listen, smell, and feel—sense the world in all its magnificent colors, depth, sounds, odors, and textures—this is the cinema of the future! (Packer & Jordan, 2002, p. 246)”



image:Morton_Heilig_Sensorama_Simulator_018.jpg Figure 18. Morton Heilig, Sensorama Simulator, U.S. Patent #3050870, 1962

Heilig developed and patented the Sensorama Simulator (Figure 18) in 1962. The Sensorama was a single person simulator that offered the viewer a multi-sensory fully immersive theatre. The viewer could sit to watch a short three-dimensional stereoscopic movie that included stereo sound, an odour generator, force feedback handle bars, chair motion and wind on the viewers face (Rheingold, 1992). Heilig believed that the Sensorama Simulator could be next generation of theatres placed in hotels and lobbies or any small space that could fit his miniature theatre (Heilig, 1955, p. 345).


Heilig also recognised that the Sensorama Simulator offered training and learning potential for educational and industrial intuitions (Rheingold, 1992, p. 58) but unfortunately the Sensorama Simulator never took off, it was “a time when the business community couldn’t figure out what to do with it” (Laurel, 1991, p. 52). This may have been different a decade later when Pong kicked-off the arcade game industry and when education, industry and government saw great potential from investing in virtual world technology as they did with the Head Mounted Display (HMD).

2.8.2.3 Head-Mounted Display

In 1968 Ivan Sutherland presented the first computerised graphical HMD (Figure 19) (Sutherland, 1968)[13]. The HMD had a cathode ray tube (CRT) for each eye with a three-dimension simple wire-frame view of a room with motion tracking when the viewer moved their head. Known as ‘The Sword of Damocles’ based upon a Greek legend of a man placed in a precarious position of luxury with a sword above his head (Oxford Dictionary, 1989), similarly the HMD had a computer suspended above the users head attached by a mechanical arm (Figure 19, right) (Carlson, 2003).



image:HUD_The_Sword_of_Damocles_019.jpg Figure 19. Head Mounted Display first called The Sword of Damocles (Sutherland,1968)


The HMD was a significant milestone in the development of virtual reality technology, which has since been used in a variety of applications in virtual worlds. Holding advantages over a traditional computer monitor such as total head and body movement, non interrupted viewing for total immersive HMDs and simultaneous viewing of real world and virtual world artefacts in ‘see-though’ HMDs or sometimes called Augmented Reality Displays (Rolland & Hua, 2005).


Today’s HMDs are more compact than Sutherland’s 1960s prototype (Figure 20). In the figure is shown on the left a HMD used for mixed reality environments similar to that designed by Sutherland and right a immersive HMD which is compatible with several online and gaming virtual worlds.



image:HUD_See_Through_and_Immersive_020.jpg Figure 20. Today's Head Mounted Displays - Left: See-Though HMD - Right: Immersive HMD

2.8.3 Early Graphical Computer Games

Computer games have had a large influence on the evolution of virtual worlds both in the development and use of the technology. The contribution of games includes computational game theory, 2D and 3D graphics, social modelling, simulation, strategies for achieving presence, artificial intelligence, computational game physics and, possibly most significant delivery of a massive consumer market to fund and drive the investment needed for innovation and technology improvement. By far the majority of today’s online virtual worlds were conceived and/or delivered as games, they have subsequently evolved into general business or training platforms which are sometime referred to as Serious Games (Annetta, Murray, Laird, Bohr, & Park, 2006).


The early computer games that can be traced to a few innovative applications (Figure 21):

  • Tennis for Two: In 1958 William Higinbotham developed the first electronic game simulator using an oscilloscope display that demonstrated a two-dimensional side view of a tennis court. It was a two player game that the user could control the direction of the bouncing ball by turning a knob on a hand held device. Originally developed by Higinbotham to occupy visitors to Brookhaven National Laboratory during open days the game had queues of people waiting to play (Brookhaven National Laboratory, n.d.). Tennis for Two introduced the concepts of a shared multi-player electronic game experience, a rule based environment managed by a machine, and an electronic space where the actions of one player in the shared space affected the experience of another. The attention the game attracted demonstrated the willingness of participants to accept the visual and sensory limitations of a machine managed game environment and immerse themselves in the experience.


  • Spacewar! The idea originated in 1961 by Steve Russell at the Massachusetts Institute of Technology (MIT) by 1962 the game was released with assistance from his colleges. Spacewar! was the first official release of a two-dimensional computer game.[14] A two player game each with a spaceship that would fire bullets at each other before being pulled into the middle by the sun. Developed originally to demonstrate the power of the new PDP-1 computer, this game was a good demonstration of both the graphic capabilities and the process power of the machine (Computer History Museum, n.d.; Markowitz, 2000). Later in 1969 Rick Blomme modified the game to run on PLATO which made this the first game to be networked (Koster, 2002; Mulligan, 2002). While Tennis for Two was the first multiplayer electronic game, Spacewar was the first computer based multiplayer game. It thus contributed the same key concepts and ideas as Tennis, only for the first time on a computer managed environment.


  • Maze War: In 1973-1974 Steve Colley developed the first three-dimensional ‘first person shooter’ (FPS) game Maze War at NASA Ames Research Center. A player would navigate around a maze searching for other players to shoot. As seen below (top right) the player had a first person view, (the eyeball seen in this picture is the other player). This is called a ‘first person’ game, placing the player ‘in-world’ as a part of the game is a significant concept of virtual world games. Maze War also provided other innovations now common to virtual worlds such as instant messaging, levelling and non player robot characters (Damer, 2007). This game which started as a two player game was eventually connected to ARPANET (the forerunner of our current internet network technology) allowing several users from remote locations to play and interact (Colley, n.d.; Damer, 2004). Maze War can therefore lay claim to being the progenitor of virtual worlds but not an actual virtual world because of its lack of persistence.


image:Early_Computer_Games_1958_To_1974_021.jpg Figure 21. Early Computer Games 1958 - 1974


  • DOOM (1993) (II, 1994) a 3D FPS game was influential both on a conceptual and technical level (Friedl, 2002; Mulligan, 2000). In DOOM the concept of Maze War was re-implemented in a much more graphically rich 3D environment. Although only a single player game, the key innovation of relevance was the method used to manage the rendering of the 3D space to allow multiple non-player characters to participate in the 3D environment with the player. The strategy adopted was essentially to divide the world into many small rooms surrounded on all sides by walls (essentially a cave system) by rendering only a single room at a time the entire resources of the computer could be devoted to a known confined rendering space, thus achieving the illusion of a highly detailed rendering with the limited computational resources available on the PC’s of the day. Although higher quality 3D rendered games were available some seven years earlier on the Amiga computers from 1986 (including some utilising real-time ray tracing technology), these relied on dedicated proprietary games architected graphics cards and did not provide a 3D space management paradigm that could be easily translated to the future demands of online 3D games. The Doom model could, precisely because it was architected for the graphically and processor challenged generalised home PC’s of the day, rather than proprietary games machines such as the Amiga. The Doom games engine was utilised in many subsequent games and later formed the basis for the model adopted for the online game Quake (Petrich, n.d.; Wikipedia Doom, 2008).


Around the time of DOOM the game industry realised the importance of connecting people together for online gaming, seeing the opportunity they started adding a modem and LAN play and later TCP/IP functionality to their games that allowed both single and multiplayer connectivity. Early games allowed up to 4 players but today’s games can have up to 64 players in a single game session (Quake Wars[15]). Some of the better known brand names included:

  • Quake (1996, a multiplayer extension of DOOM) saw over 80,000 people connected to 10,000 + simultaneous game session (Mulligan, 2000).

Warcraft (1994) (II, 1995) that eventually would become the basis to the largest MMORPG today World of Warcraft (2004) which now has over 11 millions subscribed users (Blizzard Entertainment Inc, 2008).

2.8.4 Text Based Virtual Worlds

2.8.4.1 Text Virtual Worlds: MUDs

In 1978 the first MUD (Multi User Dungeon) outside of the PLATO system (discussed under Training and Simulators) was created by a Computer Science undergraduate Roy Trubshaw (and shortly afterwards joined by Richard Bartle) from Essex University in England. A text based virtual world, coined a MUD by Bartle was based upon Robert E Howard’s (1932) fictional tale ‘The Phoenix on the Sword’. MUD1[16] was an adventure role playing game, with game levelling and chat rooms which allowed up to 32 players to connect simultaneously over a remote connection (Figure 22) (Bartle, 2003).



image:Bartle_The_First_MUD_022.jpg Figure 22. The First MUD: Roy Trubshaw and Richard Bartle (1978)


Early in the game’s history Essex University on whose computers the game was hosted became a part of ARPANET (the forerunner of the internet) and soon after MUD was distributed through that network and being played on universities throughout the world. Some of these institutions were also open for public access. Although copyrighted many variations of MUD1 were made and distributed freely from what Bartle (2003) describes as either player inspiration or pure frustration with the 32 player limitation which made it impossible to play when dial-in lines were fully allocated.


Keegan (1997) identifies two main classification of MUDs developed during this time (Figure 23) - the Essex MUDs (Trubshaw and Bartle’s) and Scepter of Goth (1978). Unfortunately Scepter died an early death, the game was sold and soon afterwards passed onto the creditors when the purchasing company ran out money (Bartle, 2003). Most MUDs were therefore based upon the ideas and technical structure of Trubshaw and Bartle’s MUD (Bartle, 2003; Keegan, 1997).


image:Basic_MUD_Tree_Structure_023.jpg Figure 23. Basic Tree Structure for MUD classification


MUD1 introduced a number of concepts retained by most of today’s virtual worlds. Among which are:

  • The role and effectiveness of the text based narrative and text communication that contributed to, rather than detracts from the sense of presence.
  • Persistence in game play.
  • Shared game space and cooperative (team based) activity.
  • Non-player artificial intelligences called AI’s (or non player characters) as part of the experience.
  • Region based environment management.
  • Role-playing as a central game theme.
  • Characters and avatars (all be it text based in the early MUDs).
  • Game defined goals but player implemented plots.


Region based environment management is a computational aid that warrants particular attention. It was also used by the DOOM 3D graphics engine to manage multi-user environments allowing the computer to render the shared space in a single discrete region at a time. In DOOM this was a room, in MUD1 it was a cave in more recent virtual worlds it may be as much as a 65,000 sqm area (Second Life). This strategy provides a method of scaling the virtual worlds to many regions by distributing the region management across many discrete servers but imposes practical limits on the number of players that can be present in any given region at an instant in time (Hu & Liao, 2004).


MUD1 had a significant impact on virtual world design and development that dominated the online game space until the mid 1990s therefore MUD1 is often marked as the beginning of the first generation in online virtual worlds (Bartle, 2003). MUD1 can still be played online today at british-legends.com (CompuServe, 2007).

2.8.4.2 ASCII Virtual Worlds

In the early 1980’s pseudo graphical interfaces were added to some MUDs in the form of ASCII virtual worlds. ACSII (American Standard Code for Information Interchange) is the most widely adopted character encoding on western computer systems. ASCII virtual worlds provided a pseudo-graphical display making use of shape symbols and character positioning escape sequences to create crude planar maps of the terrain (dungeon) environment. The maps enhance the description of the room provided by the text.


ASCII pseudo graphical virtual worlds provided the player with a view of the world improved over the simple text prompt and description of MUDs. An example of an ASCII game can be seen below (Figure 24) Islands of Kesmai (IOK). Developed in 1982 and released in 1984 the game provided a player with a 3rd person view - overhead view of the world. Walls were denoted by [], fire **) and the players were letters (Bartle, 1990). IOK was Compuserve’s (USA ISP) best selling game with players paying up to $12.50 per hour to play (based upon connection time not game played) which usually had between 10-60 players online simultaneously (Bartle, 1990). Other ASCII games around this time were MegaWars I & MegaWars III (1983), NetHack (1987 (O'Donnell, 2003)), Sniper! and The Spy (Bartle, 1990).



image:RPG_Islands_Of_Kesmai_024.jpg Figure 24. Islands of Kesmai ASCII Text Role Playing Game (1982-84)

By the mid to late 1980s home computing and online networking service providers opened the gates to huge expansion for on line virtual world. People paid for networking services by the hours, which gave a huge incentive to these providers to get their subscribers hooked on virtual worlds. There was big money to be made with 70% of revenue from one provider (Genie) in the early 1990s being made from games. By 1993 a study showed that 10% of the NSFNET backbone (precursor to the internet consisting of mainly government and universities) network traffic belonged to MUDs (Bartle, 2003).

2.8.5 Graphical Virtual Worlds

The text based MUDs evolved into two different streams: the 3D First Person Shooters such as DOOM and Quake which adopted the room at a time view of the world for 3D rendering, and the 2D graphical online virtual worlds that appeared in the early 1990s. Early examples include NeverWinter Nights (1991-1997), Shadow of Yserbius (1992-1996) and Kingdom of Drakkar (1992-Current) (Figure 25).



image:Graphical_2D_Virtual_Worlds_025.jpg Figure 25. Graphical 2D Virtual Worlds


Unlike Habitat and Worldsaway (discussed under Social Networking Virtual Worlds) that predated these games appearing in the mid-1980’s, the graphically enhanced text based games were fantasy role playing games -- basically MUDs with graphics. Although 2D some of these games displayed isometrically on an angle which gave an illusion of a three-dimensional view for the player, for this reason these games are sometimes referred to as 2 ½D worlds (Bartle, 2003). These games used more sophisticated graphics (than the pseudo graphical solutions) to improve the sense of presence experienced by the players, while retaining the text based narrative.


By the mid 1990s with nearly 10 million internet hosts (Figure 26) (Slater III, 2002; Zakon, 2006) and price wars between providers the internet opened to doors to millions which saw hordes of inexpert computer users wanting to play games (Bartle, 2003). Game design had improved long with the graphical elements of virtual worlds with graphics rendering capabilities on standard PC’s and the emergence of common graphics file standards which made development of virtual worlds possible, practical and more economical.



image:InternetParticipatingHosts_Count_1990_to_1998_026.jpg Figure 26. The Internet No. of Participating Hosts Oct. ‘90 - Apr. ‘98


2.8.5.1 MMORPGs

By the mid 1990s we saw the first 3D virtual world online Meridian 59 (1996-2000 & 2002-Current) although technically it used a pseudo-3D graphics engine (Axon, 2008; Bartle, 2003) providing a first person view where the player could view all angles in the environment (Figure 27). We saw the beginnings of a new era of virtual worlds with a massive 25,000 people signing up for the beta release (Axon, 2008), which unfortunately met with limited commercial success (Bartle, 2003; Friedl, 2002) and was shut down in 2000 but resurrected again in 2002 with the updated version online today at meridian59.neardeathstudios.com.



image:Meridian_59_First_3D_Online_Virtual_World_027.jpg Figure 27. Meridian 59 First 3D Online Virtual World (1996)

The turning point for online virtual worlds was Ultima Online (1997-Current). Ultima had already had met with success with the Ultima computer games series. With its online launch it had 50,000 subscribers within 3 months and was the first online virtual world to crack the 100,000 threshold within 12 months of release (which it did so in under 6 months) (Bartle, 2003; Woodcock, 2008). This added a new dimension to the term multiplayer where it has now come to know as a Massive Multiplayer Online, Role Playing Game or MMORPG. Subscription peaked at 250,000 in 2003 with 75,000 being reported in December 2007 (Woodcock, 2008).


Ultima Online consisting of a 2½D graphical virtual world has remained visual much the same (Figure 28) although recently the client that runs the game (the same concept as a web browser) has had a makeover in 2007 with the Kingdom Reborn (right). This game has received regular expansions to the world, which provides for new challenges and adventures for its player. Back in 2001 the client was upgraded to 3D (Wikipedia Ultima, 2008) but recently Electronic Arts announced they will be de-supporting their 3D client continuing only to support the 2D client going forward (Electronic Arts, 2007).



image:Ultima_Online_028.jpg Figure 28. Ultima Online (1997-Current)

Other MMORPGs that started around the mid to late 1990s, which can still be played online today, are Furcadia (1996, longest running), The Realm (1996, second longest 15 days out from Furcadia), Lineage (1998), EverQuest (1999) and Asheron's Call (1999).


The more recent MMORPGs of today, not much has changed in game design from the original RPGs but technically they have improved and do provide much better graphics for the player (Figure 29). They have also increased substantially in popularity with the largest subscription based MMORPG World of Warcraft recently climbing to over 11 million players (Blizzard Entertainment Inc, 2008). Although these players do not play in one virtual world they are separated into different realms, the same game but with different people. This contrasts quite differently to the social virtual worlds like Second Life where all the users share one virtual world. In the next section we discuss social online virtual worlds which although they can be a MMORPG within the world itself (as mentioned early) their model of a virtual world is very different than the dedicated MMORPGs.



image:MMOZRG_Eve_and_WOW_029.jpg Figure 29. MMORPG's Eve & World of Warcraft

2.8.5.2 Social Virtual Worlds

The first attempt for a commercial large scale multi-user game was made by George Lucas’s Lucasfilm Games. Habitat developed by Chip Morningstar and Randall Farmer started development in 1985 (McLellan, 2004; Ray, 2008; Slator et al., 2007). Habitat was built to support thousands of simultaneous users to run on the home computer Commodore 64 to be distributed via Quantum Link network service providers (later known as AOL). Inspired by a science fiction novel ‘True Names’ (Vinge, 1981) the world contained a fully-fledge economy where citizens of the world could own a virtual business, build a house, fall in love, get married and even established their own self governing laws (Morningstar & Farmer, 1990). Habitat a 2D graphical world looked similar to a cartoon (Figure 30, left) with the avatar (digital self) taking a third person view of the world. The storyline was based upon life rather than the fictional storyline of the MUDs, which placed greater emphasis on the social aspect of the world. Lucasfilm's Habitat was first released as a pilot in 1986 then later in 1988 as Club Caribe in North America which reportedly sustained a population of 15,000 participants by 1990 (Morningstar & Farmer, 1990). In 1990 it was released in Japan as Fujitsu Habitat and after extensive modifications. Habitat was realised again in 1995 as WorldsAway (Figure 30, Right) (Damer, 2007) and again as Dreamscape in 2008.



image:VW_Habitat_and_Worldsaway_030.jpg Figure 30. Habitat (86) First Graphical Virtual World Precursor to Worldsaway (95)


Habitat introduced some key concepts in virtual worlds;

  • The term ‘Avatar’ into the general virtual world community;
  • The idea of focussing on social networking as a key form of game play;
  • An economy where people could trade both in world currency and artefacts; and
  • The most important, the concept that living in a virtual world and leading an alternate life that was not dictated by rules of a game (like with the dedicated MMORPG environments).


More recent social networking virtual worlds include Active Worlds (1995, 1997-current)[17], Second Life (2003-current) and There (2003-current) (Figure 31) – all of which have achieved a significant volume of educational interest as platforms for delivery of learning. The generalised nature of the social networking sites means that they tend to be more diverse in the range of facilities provided and the purposes to which they can be applied than the role playing game systems. They have generally provided participants with some form of content creation tools including the importing and/or exporting of non-virtual world artefacts. In the next section we discuss further the aspect of education in virtual worlds.



image:VW_SecondLife_and_There_031.jpg Figure 31. Social Virtual Worlds: Second Life & There

2.8.6 Simulation and Learning Systems

2.8.6.1 PLATO

PLATO (Programmed Logic for Automated Teaching Operations) was a system designed for computer based education at University of Illinois that started in the early 1960s. Originally developed as a class room course system (Figure 32) with improvements in mainframe technology by 1972 saw up to a thousand simultaneous online users making it the first public online community that featured electronic course delivery, online chat, bulletin boards, 512 x 512 resolution monitors and 1200 baud connection speed (Unger, 1979; Woolley, 1994). With over 15,000 hours of instructional development PLATO was possibility the largest ever investment in educational technology (Garson, 2000).



image:PLATO_Lab_Image032.jpg Figure 32. University of Illinois PLATO Lab & Terminal (1961-2006)


By the mid 1970s games made their way onto the university mainframes with great success. Between 1978 and May, 1985 about 20% of time spent on PLATO was game usage (Woolley, 1994). Games appeared such as Spacewar! (1969 game discussed earlier), Empire (1973, multi user space shooter game based upon Star Trek), DND, (1974, MUD[18] based upon the game Dungeons and Dragons), Mines of Moria (1974, MUD, 248 mazes based upon Tolkien’s Lord of the Rings), SPASIM (1974, 32 multi-user, FPS space ship game)[19], Airfight (1974-75 a 3D flight Simulator precursor to Microsoft’s Flight Simulator), Oubliette (1977, first person 3D MUD) and Avatar (19977-79 first person 3D MUD) (Bartle, 2003; Lowood, 2008; Pellett; Wikipedia, 2008b; Woolley, 1994). See below (Figure 33) for some examples of MUDs held on PLATO. Many of the games on PLATO were recreated for commercial use for arcade or personal computer games (Goldberg, 2002; Mulligan, 2002; Woolley, 1994).



image:PLATO_Popular_MUD_Games_Developed_For_PLATO_033.jpg Figure 33. PLATO: Some Popular MUDs Games Developed for use on PLATO (1974-1979)


By 1985 after going commercial PLATO had established a systems of over 100 campuses worldwide (Garson, 2000). Known as the ‘ultimate electronic information and communication utility’ offering over 200,000 hours of courseware (Figure 34), with local dial-up of 300 or 1200 baud connection speed, access to both a social and educational contacts were among the many advances of PLATO that made it an attractive system for the academic community at large (Small & Small, 1984). Over time, with improvements in technology, and the cost of maintaining old technology the final PLATO system was turned off in 2006 (Wikipedia, 2008b).



image:PLATO_Online_Course_Count_1984_034.jpg Figure 34. PLATO Over 200,000 online courses by 1984


A web site has been established for preservation of PLATO at cyber1.org (VCampus Corporation, 2008) which holds many of PLATO’s games and courseware for public download.

2.8.6.2 SIMNET

Military virtual world simulators started with a project called SIMNET (SIMulator NETworking). SIMNET was a DARPA project that enabled the first large scale real-time networked battlefield simulator. Development and implementation occurred on several levels between 1983 and 1990 (Cosby, 1999; Miller & Thorpe, 1995).


Prior to SIMNET military simulators consisted of immersive virtual reality training devices such as cockpit simulators. Cockpit simulators offered a replicated environment of the ‘real thing’ for example, an aeroplane cabin would be built in its entirety with motion and sensory feedback using pre-programmed software to produce repetitive simulations to provide an individual with mastery skills such as low to ground dog-fighting or missile avoidance training (Miller & Thorpe, 1995). SIMNET provided a cheaper alternative for certain types of training than the cockpit simulators and further offered ‘collective skills’ which Miller and Thorpe (1995) define to be cohesive team operations skills distinguished from individual mastery skills taught in cockpit simulators.


SIMNET a multiuser virtual world (Figure 35) consisted of real battle grounds with manned vehicles (tanks and helicopters), command posts, semi-automated forces where a single operator could control many vehicles in the simulation and the ability to record simulations from any view point (known as the flying carpet) so that it could replayed and statistically analysed and reported upon. At the conclusion of the program there were 250 simulators operating in nine locations (4 of which were in Europe) which provided real-time battle engagements that was directly under the control of the participants (Lenoir, 2003; Miller & Thorpe, 1995).



image:SIMNET_Battlefield_Simulator_035.jpg Figure 35. SIMNET: Battlefield Simulator at Fort Knox USA (1983-1990)


SIMNET had a substantial impact on military training after being recognised as the key success factor in winning the 3 day ‘Battle of 73 Easting’ in the Gulf War (1991) which lead to several projects based upon the SIMNET technology (Figure 36) (Foley & Gifford, 2002) with the USA government commissioning $2,549 million dollars in 1997 for modelling and simulation projects (Lenoir, 2003).


image:US_Military_Networked_Simlator_Projects_1938_To_2001_036.jpg Figure 36. Timeline of US Military Network Modelling and Simulator Projects (1983-2001)


In 1997 a project named Synthetic Theater of War (SToW) commenced which was a program to construct an environment to combine varies simulators into one large-scale distributed battle simulator capable of involving thousands of participants (Budge, Strini, Dehncke, & Hunt, 1998; Tiernan, 1996). This project has since become Joint Semi-Automated Forces (JSAF) (Hardy et al., 2001) which now enables more than 100,000 simultaneous simulations at a time (US Joint Forces Command, 2008). Australia military has also adopted the JSAF platform to build their our own Course Of Action Simulation (COA-Sim) for joint military operations training, exercises and planning (Carless, 2006; Gabrisch & Burgess, 2005)

2.8.6.3 Military Use of Commercial Games Engines & The America’s Army

In 1996, General Krulak of the US Marines tasked the Marine Combat Development Command to explore and approve the use of commercial games engines for military training purposes. One outcome of this effort was the collaboratively developed Marine Doom, based on the Software Id Corporation’s shareware Doom engine and Doom Level Editor. The simulation could be configured for simulation of special missions (such as hostage rescue) immediately prior to engagement and used to rehearse the planned mission (Lenoir, 2003).


In July of 2002 the US Military released a milestone in multi-user training game simulators in the form of America’s Army: Operations (Lenoir, 2003; Zyda, 2005). Based on Epic Games ‘Unreal’ games engine, the game created a virtual world that reproduced aspects of a career in the US Army, including ‘boot-camp’ commencement and weapons and tactical training through to various operations scenarios. Although originally developed and released as a recruitment tool, the game was also claimed to be utilised to improve training outcomes by army instructors at Fort Benning (Zyda, 2005).


Now, with 26 subsequent releases (as of 2008) and available for the PC, cell phone and Xbox, the game has more than 9 million registered users exploring entry level to advanced training, and operations in small units (Figure 37). Beyond a focus on realism that extends to accurate tree placement in training courses at the simulated training camps, the game adds an added dimension of presence to the participants through the active involvement of current and former real-world soldiers as players in the game (designated with a star icon in player profiles), interacting with non-military participants (Department of the Army, 2008).



image:Americas_Army_037.jpg Figure 37 America's Army (2002)


From a training perspective anecdotal evidence from army trainers regarding the game is that sessions in the training scenarios such as the firing range or obstacle courses improve subsequent results in the real-life versions of these activities (Zyda, 2005). The US Army possibly one of the largest investors of virtual world game technology recently announced their plans to spend $50 million USD over the next 5 years to create 70 gaming systems in 53 locations around the world for combat training (Robson, 2008).

2.9 Virtual Worlds for Education

2.9.1 Architecture Considerations

2.9.1.1 Introduction

To appreciate properly the discussion of the literature examining educational directions in virtual worlds, the researcher considers a brief overview of the key architectural differences to assist the reader. This material is based on the researcher’s examination of a variety of game environments and virtual worlds, and discussions with experienced and knowledgeable users of these environments, rather than sourced from the work of other authors. As such the discussion is interpretive rather than authoritative.


Some of these environments have existed for only a few years, and have not yet enjoyed a comparative analysis undertaken by the academic community. As such, this discussion might not normally reside in the literature review, but it is felt that the placement of this discussion in this sub-section will assist the reader in better appreciating the issues explored in the literature discussion throughout the remainder of the section.


2.9.1.2 Considerations of Operational Design

While all of today’s major virtual worlds include capabilities for user interaction, sharing of the environment, persistence, avatars, business rules, streamed audio and text there are substantial differences in the technologies used to deliver the virtual experience. While some of these differences may create marginal differences in the world experience of the casual user, from the perspective of the educator and content creator the differences are substantial.


The major offerings can be viewed under the following groups (note: in each category the researcher has selected only a few example worlds, in most cases other options also exist):

  1. Proprietary closed engine (e.g. World of Warcraft, Everquest)
  2. Client resident closed content and world model with open engine (e.g. Shareware Doom )
  3. Streamed (or semi streamed) closed content and world model with closed engine (Entropia Universe)
  4. Open client resident content and world model with closed engine (Flight Simulator X, America’s Army, Unreal games, Quake, Doom)
  5. Open streamed content and world model (Hipi Hi, TruePlay, Active Worlds)
  6. Open streamed content and world model with out-of-world interfaces (Second Life V1, VastPark)
  7. Open streamed content and world model with out-of-world interfaces and open client (Second Life V1.2)
  8. Open streamed content and world model with out-of-world interfaces, open client and open server (DeepSim)


Architectural Components and Implications in Education

Below are some of the architectural components and implementations on the structure of a virtual education environment.

Architectural Components Implications in Education
Closed Proprietary System A closed proprietary system cannot generally be altered. These systems are generally not appropriate for education purposes unless the existing virtual world itself is built for the purpose of the training (such as a purpose built simulator). Closed systems can be used in education for group interaction and discussions, if not for lectures or anything requiring more than text or audio (assuming the system supports group audio communications).
Closed or Open Environment Whether content and world model is closed or open determines whether the textures, objects and artefacts of the world can be modified or created by users. This ability is essential if the world is to be utilised in education as anything more than a 3D discussion forum.
World Content Whether the content and world model is client resident or streamed goes to the complexity of distributing course content, and the dynamics available in delivery. If the content is streamed, it can be changed in real time, but will usually require a high speed internet connection. Systems supporting streamed content generally also include the tools for developing some if not all of the streamable content. If the content is client resident, client interfacing speeds can generally be slower, but the content must be centrally published and distributed to client systems and installed locally prior to use. It cannot be changed in real time, and content production will not generally be supported directly in the virtual world tool set, and will often require advanced 3D modelling skills in dedicated 3D modelling environments.
World Interfaces The existence of out-of-world interfaces goes to whether content from other sources such as the internet web pages, audio or video, etc can be streamed into the world and integrated with the world content and model. Systems capable of providing this capability with streamable open content offer the greatest potential for in-expensive production of course material and publication distribution of that material to students.
Client / Server Engine Whether the client or server engine is open or closed goes to whether the hosting software itself can be modified. Generally this should not be necessary for education if the capabilities of the engines driving the world are otherwise sufficient. Where the content / world are otherwise closed, but the engines are open, the existing content and world could be replaced by interfacing the games engine to a new world with new content.



2.9.1.3 Options for Content Modification

The ability to modify the content of a virtual world is essential if the educator is to deliver course content in-world beyond that of an interactive discussion, or monologue.


There are essentially three ways content can be modified by end-users in current virtual world environments (as opposed to systems providers or publishers) depending on the operational design of the environment:

  1. Level Editor (eg: Doom, Half Life, America’s Army, Flight Simulator). Applicable to client resident worlds (i.e. systems where the world is stored on each client computer and distributed as a separately published down load). A level editor is a content editing tool that allows an entire simulation to be created including the world model, textures, characters, behaviours, etc. They usually support importation of textures and animations, etc into the ‘level’ and then distribution of the entire level to a central server for redistribution to clients.
  2. Client Content Editing Tool with import/export (eg: Second Life, Vast Park, etc). For environments where building and content creation is part of the ‘game play’ the client will have a content editor provided. These environments provide a simplified model for constructing shapes and objects (e.g. Second Life’s prims) and some means for importing complex objects such as organic shapes, textures, animations, sound, etc.
  3. Out-of-world interface (e.g. Second Life, Active Worlds). Potentially available in both client resident and server resident (streamed) worlds. An out-of-world interface allows for the connection of some aspect of the user experience while in world to be drawn directly and live from an off world location like a web page, internet resident database or streaming SoundCast server, etc.


2.9.1.4 Implications of differential content capabilities

Virtual world are comprised of components (objects) and functions that are managed by the virtual world (or game) engine and together comprise the capabilities of the world. Not all worlds have the same object management capabilities built into their engines. For the purposes of this discussion, the range of capabilities will be considered to be:


  1. Terrain – the land form or map of the virtual space. Essentially all virtual worlds offer some form of terrain map (although the terrain map may not be ground, but rather simply a 3D space.
  2. Avatars – Discussed extensively already, the avatar is the user’s projection into the virtual world and may or may not be customisable.
  3. Structural objects – Including buildings, furniture, ornaments, statues, models, etc. These are the virtual world equivalent to objects in the real world. They may or may not be animatible and scriptable. If they are scriptable they may be able to become autonomous agents, depending in the capabilities of the scripting engine.
  4. Textures – The visual covering of any object, terrain, or even avatars. The ability to display and upload/import textures is (generally) essential to the ability to ability to display lecture materials like slides, etc (but note the existence of streams as a potential alternative).
  5. Animations – An avatar and a non-player character appears to walk, sit, stand, change facial expressions, etc because of the animation it is playing at the time. Without animations an object might move from one point to another, but it will not change it apparent state. The ability to modify animations is advantageous for creating a sense of realism, but possibly not generally essential for the ability to deliver a lecture or every type of simulation. All virtual worlds examined, offered some range of built-in animations within their worlds. Some allow the animations to be imported or modified, or strung together to create more complex animations.
  6. Scripts – Scripting is a capability to programme the objects and behaviours in the world. In worlds modified by level editors and programming language is generally provided as part of the level editing environment and ‘compiled into’ the level before it is published and distributed. In user modifiable worlds, where scripting is supported (like Second Life) the scripting editor and compiler is provided as part of the client application and scripts are dynamically modifiable. In some architectures the scripts are stored in the objects and distributed with the objects (and therefore if the object is moved between worlds/simulators the script and behaviours move with it), whereas in others the scripts are centrally stored controlled for the world/level and not available outside of the world or level or simulator (as appropriate). Scripts govern the behaviour (movement, animations, actions, sounds, appearance, world responses, inter-object communication, etc) of objects. The capability and simplicity of language design of the scripting engine is critical to the options available for educators in building a simulation.
  7. Streams – Streams include any media that is streamable such as audio, video, web-page content, etc. The availability of streams is an extension (or possible an alternative) to the ability to import textures. From an educational standpoint it represents the ability to deliver video or sound presentations, or draw lecture materials directly from the internet. Depending on the world engine, stream content may be able to be dynamically published (drawn down to the client as required – such as in Second Life) or packaged into the client resident world (such as in America’s Army).
  8. Non-player Characters (also called Bots, AI’s or MOBs – mobile objects) - These are essentially characters that look like avatars but are completely controlled and managed by the engine. They interact with players/avatars in a semi-intelligent manner. The availability and capability of these vary significantly across worlds. In HalfLife and America’s Army, the AI capability is available within the engine and has considerable ‘intelligence’ and in some cases the ability to learn and modify their behaviour. In other worlds (such as Second Life) they are not directly supported by the virtual world engine at all. The existence of non-player characters can directly impact on the type of learning simulation that an educator can build as it can provide user feedback and the feeling of presence within the environment for the user (if implemented to provide a realistic experience for the user).
  9. Text Communication - Text chat (including instant messages, group communication chat, etc) is the standard communication strategy in all worlds. It is always instant and dynamic (in that it does not have to be pre-packaged into the world). It is a functional capability rather than an object, and may or may not be logged or copied depending on the client capabilities.
  10. Multi-way Voice Communication - Most virtual worlds do not support voice directly, although this has been an increasingly offered function over the last twelve months. Multi-way voice communication enables a group of players to converse as if they were in a conference call, without the necessity to type all communication in text. It is different from streams, in that every client can be a sound source to every other client, whereas streams are a one-way communication from a point source to many destination receivers. Clearly the availability of voice communication impacts both the type student and the form of discussion that can be undertaken in a learning situation.


In selecting the platform for delivering an educational experience, the extent to which the educator requires any or all of these capabilities within a virtual world will probably influence the decision. Some of these capabilities have only recently become generally available, and others are still in only rudimentary forms. In the literature review that follows, the approaches and content adopted, and the outcomes achieved have necessarily been constrained by capabilities of the technology options available at the time and the architectural constraints of the virtual world used.


2.9.2 Education Applications in Virtual Worlds

2.9.2.1 Introduction

During the 1970’s, 1980’s and early 1990’s, perhaps the most significant multi-user online environment for education was the PLATO system. From the mid 1990’s onwards, the influence of this system waned as it was progressively superseded in user interface capabilities by the emerging 3D online games, social networking systems and custom built virtual worlds for the specific application of subject matter.


Today the use of public online virtual worlds for is gaining popularity with educators with a recent special purpose committee of educators (The New Media Consortium & EDCAUSE, 2007) identifying that virtual worlds will have a significant impact in the future of teaching, learning, or creative expression within higher education. In the next section we will discuss some of the research findings of virtual worlds being used for educational purposes.


2.9.2.2 Education Uses in Virtual Worlds

Early work in education using text based MUDs showed that they offered support for constructive knowledge-building communities that offered affordances of coordinated presence with evidence for interactive learning and collaboration across time and space (Dickey, 2003).


The period of the late 1990’s until today has been typified by educators experimenting with the potential for mass market games engines (and more recently virtual worlds) to be re-tasked as education environments (Annetta et al., 2006; Beedle & Wright, 2007; Gikas & Van Eck, 2004). In some cases, such as America’s Army the ‘game’ environment was built with the specific goal of recruitment and training in mind (Zyda, 2005), or as with MicroSoft’s Flight Simulator a game evolved over time with the assistance of subject matter experts to create an accurate simulation tool for the games audience (Lenoir, 2003). In other cases a games engine (the operating system of a game) has been adapted to create a purpose built learning tool, such as educators and students at MIT utilising the Neverwinter Nights tools to create a historical game based on a battle in the Revolutionary War or MIT's Games-to-Teach Project produced playable prototypes of four games, including Biohazard, developed jointly by MIT and the Entertainment Technology Center at Carnegie Mellon University which trained emergency workers to deal with a cataclysmic attack (King, 2003).


The early 3D virtual worlds with their simplistic graphics bearing little resemblance to the real world provided students with advantages over traditional learning methods whilst fostering collaboration in multiuser virtual worlds. An extensive study of virtual reality technology in education was performed by Youngblut (1998) where she looked at 35 different research studies in education that varied in technology use, subject discipline and age group from 1993-1998. Below is an example of VARI House and Virtual Physics both of which were custom built (Figure 38), VARI a single user virtual world and Virtual Physics a multiuser virtual world. Although studies were mainly research based (as opposed to the application in course work) research showed for both single and multi user environments that virtual world technology in many cases surpassed traditional learning methods in areas such as subject matter understanding, memory retention, student collaboration and constructive learning methods. Some obvious disadvantages were technology constraints, cost and development and usability (Youngblut, 1998) which in most part could be contributed to the infancy of this technology, formative years of computer based learning and the lack of general use of computers by students which had yet to permeate sociality as a whole.


image:Education_In_Virtual_Worlds_in_1950_to_60_038.jpg Figure 38. Education in Virtual World Mid 1990s

2.9.2.3 Online Education Uses in Virtual Worlds

As identified in the architecture considerations section, virtual worlds that are to be used in educational settings must enable content modification if learning is to consist of anything more advanced than an interactive conversation. For the purposes of this research, the researcher is choosing to focus on virtual worlds that support the dynamic delivery or streaming of content (and the building tools are provided as part of the environment), rather than those worlds where a separate level editor is required and a client resident world model must be installed on the client computer prior to use. The literature surveyed in this sub-section will therefore focus on the work done in two such environments – Active Worlds and Second Life.


2.9.2.3.1 Active Worlds

Online virtual worlds enabled educators’ access to environments without the cost and complexity of developing their own custom software. One of the first online virtual worlds that made it feasible for research and development in education (given its architecture qualities) was Active Worlds (1995, 1997). Officially known as Active Worlds Universe because it consists of many worlds, Active Worlds provided educators with the opportunity to rent or buy their own world allowing restricted access to invited guest, building tools and content management capabilities. Below is a screenshot of Active Worlds (Figure 39). As can be seen, the current client consist of four sections; left – communications and navigations options, right – integrated web browser, bottom – chat window and middle – 3D environment. This type of client is generally called a “browser” by the environment developers.



image:Active_Worlds_Universe_039.jpg Figure 39. Early Online Social Virtual World: Active Worlds Universe


Active Worlds Research

During the late 1990s to the early 2000s several educational institutions setup up a presence in Active Worlds for various projects from research to actively using Active Worlds as an online learning environment (see Smith, 1999 for a list of Virtual Learning projects most of which being in Active Worlds). The early research into online virtual world based education using Active Worlds showed promise.


Dickey (1999, 2003, 2005) undertook research into the viability of Active Worlds being used for geographically distant learners for both formal (a university business computing skills course) and informal courses (Active Worlds building course). His research studies showed that the 3D Virtual Word offered advantages in fostering constructive learning, student and teacher collaboration, visual representation of course context and course content and student engagement and participation. Some of the disadvantages identified were essentially environment specific and included a lack of support for collaborative activities like a whiteboard or collaborative interactive writing spaces, chat tool single posting word limitation, a single shared channel for chat tool providing no separation of teacher / student discussion and no ability for turn taking and kinetics (animation) constraints such as hand raising for alerting the attention of the instructor.[20]

Dickey also identified a number of opportunities specifically enabled by a 3D environment. While some of the previously identified advantages (such as collaboration and student management and participation) might be duplicated in other forms of online education tools, the 3D modelling of the course itself (the visual representation of course context and course content) was an advantage specific to the 3D environment.


Course context modelling as provided in Dickey’s research (1999) was a 3D representation that illustrated the structure of the course by the use of individual buildings and plazas (Figure 40). Each building was a topic in the subject, which provided resources to aid learning and a meeting place where students could collaborate for group projects around this topic.


image:Visual_Course_Structure_in_Virtual_Buildings_040.jpg Figure 40. Visual Representation of Course Structure by the use of Individual Buildings

Course content modelling as provided in Dickey’s research (1999) was a 3D representation that the student had to build in order to understand the concept of the subject material (Figure 41).



image:Visual_Represnetation_of_Course_Content_041.jpg Figure 41. Visual Representation of Course Content

These alternative methods provide a good example of the power and adaptability of 3D modelling environment applied to education. The course context provided the student a method by which they could visualise the learning objectives and progression of the course. The student had to visit each building within a specific time frame and complete the contained content. The 3D modelling of course content enabled the learner multiple viewpoints of actual subject material which provided interactive learning that was believed to enhance the student’s understand of the subject topic.


Clark & Maher (2006) looked at the role of place and identity in a 3D virtual learning environment using Active Worlds by the analysis of chat logs and physical locality of the avatars within group discussions. They found that a sense of place can be achieved in a 3D virtual learning environment where identity and presence plays a role in establishing the context of the learning place. The students formed a strong bond with their avatar and indicated that they felt a sense of presence, as measured by a series of subjective scales, within the virtual learning environment. Similar Dickey (2003) found that the 3D virtual desktop world provided qualities of presence similar to that of an immersive virtual reality virtual world.

2.9.2.3.2 Second Life

Second Life (started 2003) consists of two worlds. These are: Second Life Teen Grid and Second Life Adult Grid. The teen grid provides access to 13-17 year olds and educational instructors. The functionality of the teen grid is the same as the adult grid with exception that all content has a PG rating. The Adult Grid is where you find all the universities and colleges for students over 17 years of age. Other educational content in Second Life is an extensive list of museums, galleries, simulations, business product development, role-playing spaces, employee and public business training course, etc. Similar to Active Worlds educators are able to rent or purchase land, allow open or closed access to the public and build and develop on land.


One major difference between Second Life and Active Worlds is that the former has an in world economy with in-built functional support enabling the trading of virtual products and services using ‘Linden dollars’, backed by content copyright and duplication controls and augmented by a provider managed exchange where real dollars can be exchanged for Linden dollars (and vice versa). This fundamental difference provides an incentive for content developers and service providers to actively support and expand the world with content and therefore enables access to a large body of pre-constructed content or access to an entire world-wide industry of content developers at extremely reasonable rates (compared to the real world 3D developers providing the similar content outside of Second Life) (Joseph, 2007). The building and scripting tools are easier to master than traditional 3D rendering tools, and delivered free as part of every user’s world browser and are sufficiently powerful that just about anything imaginable can be constructed (Schmidt et al., 2007).


Second Life’s standard interface as seen below (Figure 42) offers extensive functionality over that of Active Worlds. Some of the more common features as seen in the figure below are built-in world, content and people search facilities (left), a mini map (top right), an inventory library (bottom right), local chat channel (with a standard ranges of 15, 30 meters or 60 meters from text source) and group chat channels (worldwide range for up to 25 groups per avatar), customisable streaming media players (for sound, video and web page content), in world or external web html browser (link for both in world and outside world content), private or public multi-player voice facilities etc.


image:Second_Life_042.jpg Figure 42. Online Virtual Social World Second Life (Circa 2008)

Another difference from Active World is avatar control, Second Life avatars can use roaming camera (whereas Active Worlds only provides First and Third person view). Roaming camera enables the user to use their mouse to control the moment around the world without the need to move their avatar. This functionality once mastered offers the users a powerful tool that provides an easy and fast way in which to navigate objects (that can even go through objects such as walls).


Due these and other technological advances over Active Worlds, Second Life has developed a large education community over the last couple of years. For instance, SIMTeach (June, 2008) the Second Life Education Wiki identifies over 200 Educational Institutions in Second Life of which 138 listed are universities, colleges and schools. The Second Life Education (SLED) list server has over 5,000 world-wide members. The New Media Consortium (NMC, a group that hosts education islands) has over 100 universities on their land and Second Life Teen Grid has over 90 educational projects (Linden & Linden, 2008). Figure 44 p88 provides some examples of the training and learning activities in Second Life representing a mixture of educational institutions, corporations and governments agencies.


The content of Second Life is entirely user created. The availability of content developers and potential students already experienced in using the environment is dependent on the take-up and expected future growth of the environment. In Figure 43 are the user base and economical statistics for the first quarter 2008 as provided by Second Life’s proprietor Linden Lab (2008a). As of November 2008 Second Life had 16,318,063 million users (60 day logons 1,344,215 million). A beak-down of Second Life’s demographics as at November 2008 can be seen in Appendix I: Second Life Demographics.


image:Second_Life_User_and_Econ_Stats_Q12008_043.jpg Figure 43. Second Life User & Economic Statistics for Q1 2008



image:Second_Life_Training_and_Learning_044.jpg Figure 44. Second Life Training and Learning


Second Life Research


Educators are using Second Life for both formal and informal purposes. Some Educational intuitions have set up entire virtual campuses modelling their real world campus while others are modelling purpose built virtual education structures. The relative youth of Second Life means that there is considerable variation in the maturity of educational efforts across the virtual world, and limited peer reviewed studies yet published. Many educators are still experimenting while others, having active support of their institutions are actively using the environment for partial or entire subject delivery. Here we will look at some of the current research at the time of writing that has been undertaken in Second Life most of which has been recently published since 2006 although given the technological advances that has occurred in Second Life since 2007 onwards we will specifically concentrate on the later research.


Martinez, Martinez, & Warkentin (2007) researched the implementation of a lecture to geographically distributed third year university students in Second Life. The lecture was delivered in a conventional lecture room setting using traditional chalk and talk style delivery with lecture slides and the chat channel for instruction, no voice was used.[21] According to the lecturer’s experience using text only delivery, the time to deliver the content was double that of a face to face lecture. This was also confirmed by the students in their survey. In the student survey some admitted they felt distracted by the novelty of the environment and were overly concerned with ancillary aspects such as their avatar’s appearance etc. Others admitted to being distracted by external (to the environment) concurrent activities occurring simultaneously on their PC’s such as multi tasking with other programs (e.g. MSN messaging) whilst at the lecture. Others experienced technical difficulties and could not get back into the lecture after they were accidentally logged out. In spite of these short-comings, when asked to rate the lecture experience on a scale of 1-10 the average student response was 8.5. In this study it was noted that some of these distractions and difficulties could be put down to first time user experience. The lecturer also felt that this lecture could have easily been pre-recorded and delivered online and that active learning techniques could have improved the delivery of this lecture in Second Life (Arreguin, 2007).


Joseph (2007) notes a consequence of using Second Life (or a virtual worlds in general) for teaching is that sessions generally take longer than traditional methods but believes that this is not an issue per se as time to complete the task should come second to the effectiveness of the experience. Joseph also believes (from experience) that the avatar projected on the screen and sense of presence experienced by the participants is more effective for learning than a live image of a video feed.


Kofi, Svihla, Gawel, and Bransford (2007) researched the potential that virtual worlds could provide efficiency and innovation for adaptive learning. In their study, students were present with a maze to navigate that simulated problem solving skills required for learning similar to that in a real life learning scenario. Kofi, et al found that Second Life was able to provide enough functionality and support for the learner to apply new concepts in order to solve presented problems as long as they were provided key indicators of possible outcomes. They also found that the use of 3D learning environments required the same amount of instruction that would be provided in equivalent real world learning and that simply building a model did not provide sufficient information, of itself, for the learner to learn in this instance; they also needed to be continuously prompted and guided in order to reach the end learning objective.


In another example, Second Life was used to support learning objectives of a total of 13 students aged between 19-26 for a third year level college students on a course for Digital Entertainment and Society where the students were geographically distributed around the world (Gonzalez, 2007). Both lectures and assignment work was conducted within Second Life. The lectures consisted of a video presentation and an in world field excursion. Assignment work required some in-world building, an exercise using linden dollars with a student presentation on completion. No students had used this environment before but an acclimation exercise was sufficient in providing them with the skills required to undergo course work in Second Life. At the end of the course students were given a survey with results presented below (Table 1).


Elements that Second Life Added:
Agree Disagree
Enjoyment 100% 0%
Technical difficulties 100% 0%
Interaction with tutor 62% 38%
Interact ion with classmates 62% 38%


Table 1. Survey Results for Digital Entertainment and Society Second Life Subject


The technical difficulties result was explained largely by network latency experienced by the students. Each student used their own computers with an average of 512 Kbs connection speed – not especially fast, nor ideal for the use in the Second Life environment. No mention was made in the study as to whether the student computers met the Linden Lab systems requirements (2008c). As Second Life is streaming virtual world where content is downloaded on-demand from Linden Labs servers located in the USA to the local computer connection speed can an important factor in technical difficulty performance. Other major impacts from a technical perspective include the computer graphics cards and the size of onboard computer RAM. The Second Life browser does offer many settings for optimising performance on low-end machines but if the minimum system requirements are not met then the user’s experience of the virtual world will be reduced significantly with dropouts, lag and poor graphics.

2.10 Learning & Instructional Design Theory

2.10.1 Introduction

Learning in any world (real or virtual) requires well thought out instructional design. Learning is a process of the mind regardless of whether your body is present in the virtual world or real world. Instructional components for learning regardless of medium include (DONCIO et al., 2008):


  • Clear, concise, and appropriately structured content
  • Activities that draw relationships between concepts, challenge learners' thinking and understanding, and reinforce information
  • Evaluative measures that determine if knowledge assimilation and retention have occurred


In this research the focus was on the use of new technology in education as opposed to education applied to new technology; therefore this section only provides an overview of applicable theory required to assist in the instructional design, delivery and assessment of the subject material presented to the research participants in this study. Gagne’s Nine Events of Instruction and Bloom’s Taxonomy of the Cognitive Domain were selected to assist in this task.


2.10.2 Behaviourism and Cognitivism

There are two main traditional schools of thought in learning theory. These are Behaviourism and Cognitivism (DONCIO et al., 2008; Lewis, 2001).

  • Behaviourist (Objectivist) views the mind as a ‘black box’ no consideration of personal or past experience is taken into consideration. The mind starts off with a clean slate where a stimulus produces a response. Only when a change in behaviour is observed learning has occurred. Learning is discrete, measurable and quantifiable.
  • Cognitivist (Constructivist) views the mind as a continuous organism that evolves. Knowledge is constructed based upon from past material and personal experience. Learning is unique to the individual; relating new information based upon pervious knowledge learnt.


The University of Washington, Seattle (2008) compares the two approaches of and a provides a discussion of each in terms of philosophy (Table 2, p93), learning outcomes, instructor role, student role, activities and assessment. The philosophies of these approaches are opposing and therefore produce different methods of instruction (Lewis, 2001; Nash, 2007).


Behaviourism was the first to be defined in learning theory while cognitivism developed later as a response to perceived limitations of behaviourism in understanding and adapting to new learning concepts (Lewis, 2001; Mergel, 1998).


While some constructivists argue the merits of constructivism as a distinct theory, viewing knowledge as a something constructed by a learner through the process of learning other writers view constructivist ideas as an evolution of the fundamental cognitivist school. This position is illustrated in Table 2 where the behaviourist and constructivist-enhanced-cognitivist philosophies are compared using a consistent comparative organisation of views (see Dabbagh, 2006; Mergel, 1998).


Constructivists argue a distinction between cognitive constructivism and social constructivism, in which the former emphasises the exploration and discovery on the part of each learner, while the latter emphasises the collaborative efforts of groups of learners as sources of learning, but for our purposes it is sufficient to distinguish the behaviourist and cognitive approaches. Throughout the years many practical teaching methods have evolved with concepts that encompass both approaches.


image:TABLE_Instructional_Design_Behaviorism_Cognitivism_045.jpg Table 2. Instructional Design: Comparative Summary Behaviorism and Cognitivism

(University of Washington, 2008)

2.10.3 Gagne’s Nine Events of Instruction

Gagne theory of instruction can be divided into three areas (Corry, 1996); taxonomy of learning outcomes, conditions of learning and levels of instruction. There are considerable similarities between Gagne’s ‘taxonomy of learning outcomes’ and Bloom’s ‘taxonomy of the cognitive domain’ therefore a discussion of these will be provided in the next section of this thesis.


Gagne breaks down ‘conditions of learning’ into internal learning and external learning conditions. Internal learning is concerned with previous learned capabilities of the learner and external learning is the instruction or stimuli that will be presented to the learner. While Gagne’s theory takes an essentially cognitivist approach, it recognises both behaviourism and cognitivism influences to instructional learning. For our purposes, it is the ‘levels of instruction’ as outlined by Gagne that are of particular interest which we will explore in this section.


Gagne (1985) presents a systematic approach to instructional design termed the ‘nine levels of instruction’ as presented below in Figure 45 (Clarke, 2000)[22]. These nine levels have been specifically designed for the teaching of intellectual skills.




image:GAGNE_Nine_Steps_To_Instruction_046.gif Figure 45. Robert Gagne's Nine Steps of Instruction (Clarke, 2000)



The nine instructional events with their corresponding cognitive processes can be described as follows (Clarke, 2000; Kearsley, 2008):

  1. Gaining Attention (Reception): Grab the attention of the participant by presenting a teaser in order to get the participant interested and motivate them to learn more about the topic that will be presented. This could be done using methods such as a movie, phrase, storytelling or a demonstration.
  2. Informing Learners of the Objective (Expectancy): Provide the participant with the objectives in order to assist them in organising their thoughts ready to receive the new information that will be presented.
  3. Stimulating Recall of Prior Learning (Retrieval): Provide the participant with any background that my assist them in building upon the new knowledge that they are about to receive. This helps to place a framework in their mind based upon previous knowledge.
  4. Presenting the Stimulus (Selective Perception): This is where the new learning begins. Information should be chunked and organised meaningfully in order to avoid memory overload and assist in the learning of new knowledge. Chunking the information into sequence of learning events and breaking it down into constituent parts with a structure and purpose that spans across different areas of comprehension. The revised Bloom’s taxonomy (discussed in the next section) can be used to assist in forming of the presented information.
  5. Providing Learning Guidance (Semantic Encoding): Assisting the participant to obtain a deeper level of understanding of the new knowledge so that information can be encoded into their long term memory. During instruction try to provide examples, non examples, analogies, graphical representation etc. to assist in semantic encoding process.
  6. Eliciting Performance (Responding): Letting the learner do something with the new knowledge or test their new knowledge to confirm they have a correct understanding of the information.
  7. Providing Feedback (Reinforcement): Analyse the learner’s understanding of the subject matter presented and provide feedback to correct any misunderstood knowledge. Immediate feedback and reinforcement of the new knowledge (e.g. question and answers).
  8. Assessing Performance (Retrieval): Test that the new knowledge is understood and the learning objectives have been met. This could be in the form of a test or a demonstration by the learner to assess if they have mastered the information.
  9. Enhancing Retention and Transfer (Generalisation): Generalise the information so that the knowledge transfer can occur, inform them of similar problems or a similar situation so that the acquired knowledge can be put into a new context.

2.10.4 Bloom’s Taxonomy

The Taxonomy of Educational Objectives also known as Bloom’s Taxonomy is widely used[23] to assist in the preparation of learning objectives and the assessment of learning outcomes. The learning outcomes of a student are the results of their learning experience of a course that should be a direct consequence of the course objectives (Monash University, 2008). Hence the application of Bloom’s taxonomy of educational objectives in forming course objectives provides a measure by which to assess student’s learning outcomes.


The original work of Bloom’s Taxonomy was developed by an American committee of educational psychologists chaired by Benjamin Bloom that presented over a period of time three domains: cognitive (knowledge) (Bloom, Englehart, Furst, Hill, & Krathwohl, 1956), affective (attitudes) (Krathwohl, Bloom, & Masia, 1964), and psychomotor (motor skills) (Dave, 1967, 1970; Harrow, 1972; Simpson, 1972). In forming educational course objectives Bloom’s cognitive domain is applied to assess the knowledge and intellectual component of a curriculum.


After nearly 47 years had passed Bloom’s cognitive domain was revised (Anderson et al., 2001; D R Krathwohl, 2002) by a committee of eight, two of whom had worked on the original published work (committee: Krathwohl and editor: Anderson). The revision was made as a result of many years of application and research and has since been accepted by many educators as a replacement for Bloom’s original work. The changes that were made are as follows (Figure 46) (Anderson Research Group, n.d.; D R Krathwohl, 2002):

  • The names of six major categories were changed from noun to verb forms.
  • Comprehension and synthesis were retitled to understand and create respectively, in order to better reflect the nature of the thinking defined in each category.
  • Create was moved to the highest, that is, most complex, category.
  • The revised Taxonomy is not a cumulative hierarchy.
  • A taxon of remember was devised to replace that of Knowledge, and
  • A two dimensional Cognitive Taxonomy Table was formed by sub dividing the original Knowledge taxon.


image:BLOOM_Changes_in_Cognitive_Domain_047.jpg Figure 46. Changes in Bloom’s Cognitive Domain


2.10.4.1 Revised Bloom’s Taxonomy of the Cognitive Domain

A substantive difference is in the handling of “Knowledge”. The revised Bloom’s cognitive domain as shown in Table 3 was extended to include the dimension of Knowledge. So now the revised Bloom’s cognitive domain consists of a two dimensional table with The Knowledge Dimension and The Cognitive Process Dimension. This table provides the instructor with a tool with which to classify learning objectives where learning objectives are specific and inclusive to the discrete learning outcomes or intended results that are hoped to be achieved by the end of instruction. The instructor defines the learning objectives where these objectives are classified into the appropriate cell in the 2D matrix of cognitive and knowledge dimensions which then assists in instructional design, and assessment and provides a tool to enable balancing of the learning objectives across methods of instructional design.


image:BLOOM_TABLE_Revised_Taxonomy_048.jpg Table 3. Revised Bloom’s Taxonomy Table

(Anderson et al., 2001, p. 28)


The Cognitive Process Dimension

The Cognitive Process Dimension is the column values for Table 3 above. This dimension provides the level of learning and comprehension required to complete a task where each differs in their complexity on a scale from 1-6. Cognitive dimensions are defined as 1.Remembering, 2.Understanding, 3.Applying, 4.Analysing, 5.Evaluating and 6.Creating each of which contain further sub-process with 19 specific cognitive processes in total. Table 4 provides an overview of each cognitive process with their defining verbs. Verbs are used to classify an objective. For example, an objective ‘to recall the 7 states of Australia’ would be classified under remembering. Recall in this instance is the verb that classifies the learning objective into level “1. Remember” of the cognitive dimension.


[[image:Cognitive_Process_Dimension_Processes_049.jpg: Table 4. The Six Categories of The Cognitive Process Dimension And Related Cognitive Processes (Anderson et al., 2001, p. 31)

Bloom’s cognitive taxonomy was solely based upon the values contained in the cognitive dimension (with the exception of the differences previously discussed). Bloom believed that the cognitive process was a cumulative learning process in order to achieve a learning outcome. For example, in order to ‘analyse’ subject matter the student would need to have mastered using the old Bloom’s taxonomy of the cognitive domain knowledge/remember, comprehension/ understand and application/ apply whereas the revised taxonomy of the cognitive domain does not assume this cumulative hierarchy. The early Bloom’s cognitive domain took a behaviourist approach to instruction whereas the revised Bloom’s cognitive domain believes that learning can take place at any level without mastering previous levels. This is a fundamental shift in the philosophical grounding of Bloom’s taxonomy of the cognitive domain where it has moved away from the behaviourist approach of learning.


The Knowledge Dimension

The Knowledge Dimension provides an additional dimension that has been added to the taxonomy by the subdivision (and modification) of Bloom’s original knowledge category, which can be seen as row values in Table 3 above. The knowledge dimension defines how knowledge is constructed which can be Factual, Conceptual, Procedural or Metacognitive. Table 5 provides an overview of the knowledge dimension and their meanings.


The knowledge dimension separates the noun (or subject matter) from the stated learning objective. For example, continuing on from the objective discussed above ‘to recall the 7 states of Australia’ would be factual knowledge where the bolded words make up the noun construct. This noun is factual because the learner either knows the states or they don’t, to know is the basic element required in order to solve the problem.


image:Major_Types_and_Subtypes_Knowledge_Dimension_050.jpg Table 5. The Major Types And Subtypes Of Knowledge Dimension (Anderson et al., 2001, p. 31)

The knowledge dimension has been added as it provides further insight to the type of knowledge a student is required to master. In the original work this assumption was also made as it was the first level in a cumulative hierarchy but the revised knowledge dimension provides the instructor with a greater understanding and assists in defining knowledge as a separate dimension. For example, the objective ‘to recall the 7 states of Australia’ the student needs to Remember Factual Knowledge.


The knowledge dimension like the cognitive dimension is not a cumulative hierarchy, learning can start anywhere within the knowledge dimension.


Using the Revised Bloom’s Cognitive Domain to Assist in Instructional Design

To assist in formulating instructional design Anderson et al. (2001) provides in their book for the cognitive dimension; sample objectives, corresponding assessments and assessment formats (chapter 5) and in the knowledge dimension; specific details, elements, generalisation, structures and models etc (chapter 4). This assists in the formulation of specific tasks and in defining the level of knowledge required of the student. It also assists in ensuring those objectives and testing of those objectives lie across the required range of cognitive and /or knowledge categories and that the student is being fairly assessed in areas that are directly related to the objectives.


2.10.4.2 Bloom’s Taxonomy of the Cognitive Domain Applied to a Digital Environment

Bloom’s Digital Taxonomy of the Cognitive Domain

Churches (2008) has extended the (revised) Bloom’s cognitive domain for digital learning by taking the cognitive process dimension and included verbs for emerging technology. As can be seen below (Figure 47) the words highlighted in blue are the digital emerging technology verbs that have been categorised by using (revised) Bloom’s cognitive levels as the basis for interpretation of complexity. For example bookmarking is a remembering process is simpler than programming (which is a creating process).


image:BLOOM_Revised_As_Digital_Taxonomy_051.jpg Figure 47. Bloom's Digital Taxonomy


Churches further added within his classification system a rubric (scoring criteria) of these technologies similar to that that has been defined in the sub-classification system used in Bloom’s cognitive domain. For example, Table 6 displays the rubric for Bookmarking where it has been broken down from simplest to highest.



image:BLOOM_Bookmarking_Rubric_For_Digital_Taxonomy_052.jpg Table 6. Bookmarking Rubric for Bloom’s Digital Taxonomy



Bloom’s Taxonomy of the Cognitive Domain applied to Games

Wang & Tzeng (2007) proposed using the (revised) Bloom’s taxonomy of the cognitive domain as a method for understanding the application of knowledge in digital games. They believed that players learn in various ways within computer games and recognised how little work (if any) had been done in analysing such e-learning platforms in a structured taxonomic manner and in structuring the implementation and understanding of the cognitive processes. They proposed using Bloom’s taxonomy of the cognitive domain as a method by which to assess cognitive processes in a computer game.



image:BLOOM_Taxonomy_For_Games_053.jpg Figure 48. Bloom’s Taxonomy for Games


The research included using a game called Food Force, which was a problem solving and mission-oriented game. Figure 48 summarises the conclusion of their research. As can be seen in Figure 48, players exhibited both personal and social feedback cross Bloom’s cognitive levels. They found that the players experienced cognitive processes for individuals across all categories of the Bloom’s cognitive model and displayed social interaction for the higher level Bloom’s categories of Analyse, Evaluate and Create.

2.11 Summary

The acceptance of the latest crop of virtual worlds such World Of Warcraft, Second Life, Entropia Universe, There, Eve, America’s Army and others by the internet using public as an integral part of their life style is possibly the most significant paradigm shift to occur in the last 10 years. With the statistics of user volumes and retention rates shows consumption numbers in the tens of millions of users, spread evenly across ages from youth to middle age and an approximately even gender balance (at least in the social worlds) (KZERO Research, 2007; Woodcock, 2008; Yee, 2006). The growth rates of these worlds collectively have been, and are projected (by industry analysts) to continue to be, rising dramatically for the foreseeable future.


With the current convergence of disparate technologies represented by these systems, the general public now have affordable single platform multi-media collaborative environments with sufficient realism to create virtual immersive spaces where presence is achieved at a level sufficient for them to lead virtual existences and establish social networks that rival their real world existence.


The linking of these spaces with the affordable (often free) tools that enable the public to create new 3D spaces and content for these spaces over the last eight years has resulted in a world-wide content developer base that with substantial skills and a highly competitive market for purchasers of those skills at often very low rates.


With the combined market pressures of minimising education delivery costs, improving education outcomes, and reaching as wide a market as possible it is understandable that educators have shown an extended interest over many years in the possibilities of virtual environments for education delivery. So with the advent of the latest generation of creativity focused social worlds like Second Life over the last few years, it is not surprising that the uptake by universities and educators (numbering in the hundreds of institutions) has been as substantial as it is.


A brief retrospective of the work in simulators, virtual reality and 3D games, shows that the potential of these environments extends beyond the virtual ‘chalk-and-talk’ to enabling education delivery strategies for even campus based students that cannot economically be delivered using reality bound means.


With traditional real world learning environments there is an extensive body of tested knowledge that can provide clear guidance as to workable frameworks for the design of course work. The extent to which and how these methods can or should be applied to the virtual world learning space remains an open question.






BackLinks




CopyRight Bishop Phillips Consulting Pty Ltd 1997-2012 ( Real Learning in Virtual Worlds - CHAPTER 2: Literature Review )
Personal tools