Developing virtual worlds: The interplay of design, communities and rationality
First Monday

Developing virtual worlds: The interplay of design, communities and rationality by F. Ted Tschang and Jordi Comas

This paper examines the evolution of virtual worlds from the developer’s perspective. It asks two questions: What are the motivations of developers? What are the specific challenges of the governance of user–generated content? User–created virtual worlds may be characterized according to their degree of design or emergence. On one end is the ‘designer as god’ perspective and on the other is the unforeseeable and perpetually emergent ‘user creativity.’ Utilizing a theoretically derived sample of virtual worlds, we illustrate how governance is complicated by designers contending with three major issues. In general, across all three worlds, developers had to come to grips with the limits of their ability to design virtual worlds for premeditated outcomes. Secondly, communities forming within worlds, as opposed to atomized users, are central to the (creative) building, usage and governance of virtual worlds. Developers have a range of choices for how to interact with communities ranging from arm’s length monitoring to engagement. Thirdly, developers have to manage instrumentally rational aspects of their business which can lead to tensions with the design and community goals, and, ultimately, lead to the failure of a world’s business model. A fuller accounting of governance will have to accommodate the complex interplay between purposeful design, emergent community, and the logic of the marketplace.


1. Introduction
2. Existing perspectives on virtual world governance and development
3. Data and methods
4. The design perspective
5. The role of communities and designing to communities’ needs
6. Development in the service of economic rationality
7. Conclusions



1. Introduction

“There never was any grand plan on the Internet, and there isn’t one today. The Net is just the Net … Funny thing: lawless, planless, management–free, they’re figuring out what to do with the Internet much faster than government agencies, academic institutions, media conglomerates, and Fortune–class corporations.”
— Levine (2000). The Cluetrain Manifesto.

Our objective in this paper is to understand the governance issues that arise within and subsequently affect virtual world development. We want to shift the focus from popular massively multiplayer online role–playing games (MMORPGs) like Everquest and more recently, World of Warcraft. We do so by examining the user created (or user generated) variety in more detail, including Second Life and its brethren. Pearce (2009) places these at one end of a continuum that ranges from “ludic” which are goal–oriented to those that are open–ended or “padaic” [1]. We agree with Pearce (2009) in thinking of the whole continuum being one of created, synthetic worlds. They are all massive, multiplayer, online, worlds (‘MMOWs’ so to speak). Morevoer, as others have discussed, several elements seem common to worlds: avatar interface spatial organization, in–world synchronous and asynchronous communication. In short, as Lessig (2006) describes, there is a distinct placeness that one notes when in entering a cyberspace. Virtual worlds are a subset of cyberspaces more generally.

In general, we can think of this latter type of virtual world as being open ended in use, allowing users to create their fancies, and to interact in freeform manner with one another, be it by social “chat” or deeper online relationships [2]. These worlds are typically composed of: (a) a platform with features and “tools” to empower user creation, (b) the user customization of those features and their applications, (c) user–created content, and (d) user behaviors. In studying these worlds, we will also go further than the typical starting notion of governance as “community management”. This stems in part from theories of game design. For example, Salen and Zimmerman (2003) point out “As a game designer, you can never directly design play. You can only design the rules that give rise to it.” [3] We seek to illustrate how developers embrace interactions with communities in order to better develop their worlds. We are in dialogue with the discourse about governance of virtual worlds (Barbrook, 1998; Bartle, 2006; Malaby, 2006). We understand governance broadly as the sum of rules, policies, norms and other sources of regularity and control of behaviors and interactions. Governance is not reducible to formal elements, but, rather, is always a lived governance.

Scholars of games (Castronova, 2005; Taylor, 2006a; Taylor, 2006b; Malaby, 2009; Malaby, 2006; Pearce, 2009) have enabled us to take as our starting point the continuum of worlds from heavily designed and goal–oriented to co–created and open–ended (i.e., “sandbox” in nature) (Figure 1). These types of worlds map onto typical governance categories of top–down or even “god” designed worlds on one end to purely emergent or even anarchical at the other. We will broaden this continuum in two ways. First, we will broaden the idea of open ended worlds past one–on–one relationships between developers as a group and users as individuals or aggregates of individuals. We argue for including in design approaches the emergent characteristics and uses of communities of users, to understand how worlds adapt or change. The second way is by describing the processes or mechanisms between developers and users and communities of users over time to capture the unique manner of their interactions.

Our guiding empirical question is of how software development practice and the broader development process unfold in user–created and experienced environments such as user–created virtual worlds. Along these lines, we wondered: do general patterns of development emerge across various user–created virtual worlds, and to what degree is development conditional on the type of world? Said another way, is divergence or convergence in development practices occurring across these worlds? The central question of development across worlds is mediated by broader trends. As scholars have noted (Castronova, 2005; 2007; Pearce, 2006), a hard boundary between real and virtual worlds and interactions is a misguided premise. Furthermore, as we argue elsewhere (Comas and Tschang, 2010), the ‘industry’ of virtual worlds is experiencing great turbulence at the firm and industry level. Taking these two facts: the porous boundary between real and virtual and the turbulence of the field of virtual worlds, we argue that the process of world development must take into account these larger trends and unfolding contingencies. Castronova (2005) makes a similar point:

“… synthetic worlds now emerging from the computer game industry, these playgrounds of the imagination, are becoming an important host of ordinary human affairs. There is much more than gaming going on there: conflict, governance, trade, love … Even if you haven’t paid much attention to multiplayer video game worlds up to now, soon enough, I think, you will. We all will.” [4]

As we show in this paper, these worlds also coexist, compete, and corroborate with other forms of social media, with the Internet at large, and ultimately, with the myriad actions and interactions that we call real life.

We will compare selected user–created virtual worlds, including one established world (Second Life), and two worlds under development (Metaplace [closed in January 2010 at the tail end of our data collection for this paper] and Tirnua), to determine the manner in which their overall design developed and evolved. These three worlds vary across particular dimensions. The last two are recent worlds which we engaged recently in their open “beta testing” phase (i.e., operational prototypes under early testing by non–developers). Given that these latter two are still nascent, this allows us to capture “in world” phenomena in a more generalized way, and also to observe high levels of developer–user interaction. Taylor (1999) made one of the earliest cases for using in–world interactions as essential data. We will look in particular at how developers’ or users’ actions, “govern’ the development and operation of these worlds. Our goal is to bridge the theories of how world governance is viewed in positive terms, to more normative notions that can help in better understanding future worlds’ development.

In section 2, we will examine the relevant literature. In section 3, we review our data and methodology. In section 4, we examine how developers have designed and developed virtual worlds, partly as a “planned” process, but increasingly, by taking into account the emergence of new and unforeseen applications. In section 5, we examine the fundamental role of communities in the governance and evolutionary process. In user–created worlds, communities also play a series of roles through creating content through to the governance of and assisting in the design of the world. At the same time, developers’ engagement with communities helps takes emergent phenomena into account, so we will also examine the modes by which developers and communities interact. Finally, in section 6, we examine the broader “real world” that virtual world developers inhabit, in terms of the rational and other commercial interests that cause them to focus on growth and revenue prospects. While each of these perspectives is intricately connected to the other, we feel that their separate treatment allows us to better come to grips with the central issue of governance.



2. Existing perspectives on virtual world governance and development

In this paper, we posit a way of thinking about governance that enfolds not only the “normal” development of the virtual world, including its planning, design, and contingencies, but also the operation of the virtual world, including social and creative behaviors in all their ‘gory’ forms. As part of the development process, there is the general need to enable user creativity, or user–creating behaviors. The community of a user–created virtual world enacts its own forms of play — be it creative or non–creative. Some users seek self–fulfillment while others seek recognition from the community, i.e., they are driven by status. Some users play by creating content, while others do it by exploring, socializing, or acts of display (Ondrejka, 2004a; 2004b). Taylor (2006a; 2006b) describes this as the “co–construction” of games. “Most radically put, the very product of the game [or world] is not constructed simply by the designers or publisher, nor contained within the boxed product, but produced only in conjunction with the players.” [5] While community governance and developer monitoring and enforcement of the community is an important function for enabling play, at least for the vast majority of users, we are also interested in how the play affects the developers, and ultimately, the world, as other, more serious, users come online.

Creativity, community, and play

Creativity is thought by some developers to be the most exalted activity for users and developers to aim for in user–created worlds (Ondrejka, 2004a; 2004b). The developers’ empowerment of users’ creativity — directly through design and indirectly through signaling their openness to user creativity — is crucial to a user–created world’s viability. This is enhanced to the extent that the world’s business plan depends on creativity. A host of more mundane user activities are also necessary to sustain the world and to make it viable for creators and non–creating users alike — a working economy so to speak. Hence, the fostering of creativity can be interpreted as part of the world’s governance.

Two ends of a spectrum: Design, and emergence (from communities)

In a user–created virtual world, governance involves both enabling users’ creativity as well as ensuring viable user experiences (often by the management of communities). Both developer and community perspectives of governance have been discussed in the literature to some extent. There are two ends of the design–control spectrum. At one end of the spectrum, which we call the “design perspective”, the developers are assumed to have powers over the virtual world, and have even been characterized as “gods” in that they provide the basic design, parameters by which the world operates, etc. (Bartle, 2006). To some degree, this resonates with the “technology push” perspective in innovation studies, as well as the ‘code as law’ half of Lessig’s (2006) work. While Lessig makes clear we face choices about what gets coded as law, there is no option for a law or governance–free space. “Code will be the central tool of this analysis … There is no middle ground. There is no choice that does not include some kind of building.” [6] There is a significant academic literature on virtual world design (Bartle [2004] provides an almost canonical text). Within this view, users adopt the platform, and designers design and re–design iteratively either with users’ experiences and feedback in mind, or with the active participation of users. In this tradition, virtual worlds are seen as top–down designed products (Bartle, 2004). Bartle argues that “players know nothing about virtual world design”, in large part because “their knowledge is too personal.” [7] Further, he notes “That’s why most players aren’t good at design. They still sense the magic. Unfortunately for the live team [8], that’s not quite how the players themselves see it … The pressure [of players] can be phenomenal … At this point, the live team often surrenders. Top–down design gives way to bottom–up experiment.” [9]

However, there is another emerging view that is consonant with the rise of social media, and its inherent qualities of social use, manipulation, diffusion, collaboration and even playfulness. This “communities” perspective seeks to understand how communities use the media and innovate within the media itself (Pearce, 2006). The modding of games by users also illustrates a form of user community–based creativity (Jeppesen, 2005; Kow and Nardi, 2010). Designers face a variety of on–going governing challenges stemming from what emerges from the on–going interactions between on one hand, the design, and on the other, the communities’ creativity and use of the design.

Communities of users emerge in difficult to predict ways. The communities perspective seems to be missing from legal studies of gaming like Lessig (2006) as well as the design literature (Bartle, 2004). The later suggests that users will adapt, but does not fully account for communities of users as its own relevant unit of analysis. Pearce’s (2009) formative Communities of play discusses communities of play as distributed networks of players with emergent characteristics and group dynamics. On the other hand, to analyze the role of communities in the development process, or as source of world governance, requires a design– or development–centered approach.

Consonant with Pearce’s view of emergence are recent experiences with user–created virtual worlds like Second Life. While developers have viewed virtual worlds as architectures to be planned and coded, in practice, users have used virtual worlds in ways that modify and even subvert the developers’ original intent. For instance, both Malaby (2006) and Au (2008) discuss how Second Life users developed new uses for the technology that were unanticipated by the developers (Au, 2008). Morningstar and Farmer (1990) observe a similar experience with Habitat — considered to be the first virtual world. Related to this is Malaby’s (2009) idea that community–led (or user–led) phenomena are a function of the “contingent” and performative nature of play. We already know from first hand observations of user–created worlds like Second Life that they are bound to evolve past their initial conditions, and that usage at the “community” level, as opposed to only the individual, can be highly emergent (i.e., creating unexpected outcomes). For instance, the “creative” reactions of one part of the community to another can lead to an unexpected outcome. Given this, our research approach has been to understand in a grounded way how these worlds are formed and evolve from the combination of developers’ actions and the community’s behaviors, collective creativity, and involvement in governance.

The performative nature of play can cause tensions between users, as one user’s notion of play may conflict with another. For instance, a common occurrence is the appearance of “griefers” or non–conforming users (Bartle, 1996). In Metaplace, individuals could “walk” into a user’s created world, essentially a semi–private “room”, and start engaging in forms of conversation less acceptable to the residents of that room. Another example of this includes the various types of avatars found in Second Life, ranging from “furries” (avatars based on animal creature forms) to child avatars. One graduate student in one of our classes encountered a user at one Second Life social gathering spot assuming a cat avatar, and doing nothing but communicating with cat noises and gestures during the entire group conversation.

Partly as a consequence of all these differences in usage, virtual worlds require active management. Second Life (as with other worlds) has experienced many episodes of both constructive and destructive user and community behaviors (Malaby, 2006). Some of this required active community management by Linden Lab (Second Life’s developer). Community self–management is not discussed as much in the academic literature, although it is a significant part of the Second Life experience. Usually, island owners assume controls over their own territory. Perhaps the first significant case of a community taking matters into its own hands (with the help of the operator–managers) was that of LambdaMOO, an early Multi User Domain (MUD) (Dibbell, 1993).

Dealing with complexity by balancing perspectives

We argue in this paper that virtual world development, and hence, governance, must be viewed as not just the designers’ intent, nor as adaptive use by individuals. World development is more complex. It is due to a delicate interplay between the developers’ design, the emergent uses that come from communities of users, and as importantly, the interactions between those two. In a user–created world, designers can only fashion a world that enables user–generated creativity, but they do not create the world itself. Specifically, we seek to understand the patterns of governance in a virtual world from three perspectives, each of which shapes developers’ thinking at one time or another during the development process:

(1) Design perspective: The design perspective illustrates how developers treat the world in the manner of a rational (i.e., planned) design and development process. We show that different virtual worlds do converge in particular practices since there is a common core experience to virtual world development, even as they learn from one another. At the same time, most worlds are bedeviled by a particular issue: that of dealing with emergent trends that come from users playing and creating with one another. This leads to our second perspective.

(2) Communities and emergence perspective: Communities are essential to virtual worlds for a variety of functional purposes. Perhaps most importantly, in user–created worlds, they are as responsible for the world’s makeup (in terms of content and behaviors) as its governance. This requires developers to be sensitive to community needs, which they do with a variety of mechanisms for “engagement”. Thus, while each world eventually defines its own uniqueness, it also has to provide more of what communities commonly want. (This by itself causes some convergence to occur across worlds.)

(3) Economic rationalization perspective: The third perspective (typically less recognized in the academic literature), is that of the need for virtual world developers to demonstrate viable revenue models for stakeholders such as their financiers, or value propositions for other potential clients or stakeholders (such as those in the health, government, or education sectors). This may in turn require a virtual world to be pragmatic in its treatment of communities and their preferences. In general, social media tends to have a difficult time convincing users to pay for services [10]. Ultimately, this perspective would have developers balancing the interests of users and communities with the needs of revenue generation and financial sustainability.

In a market space, between users and business rationale, users are consumers; they demand quality products which shape business rationale (Tschang, 2007). Both users and companies have at their disposal specific measures of governance (see Figure 1). These measures are linked to the ethical mechanics that exist in these communities (Ito; Kow and Nardi, this issue). Governance measures limit the extent of contingencies and bring stability to the product in use, in addition to furthering the community’s own interests. The empirical data seem to have implied that governance exists only in the participatory space. But rules and norms are always present in social relations, and companies are also social communities.

It is worth noting that the design perspective is more oriented around a “methodology” to promote the designer’s plan for action, whereas “community” is an organizational concept that is “in–world” (i.e., pertaining to what happens inside the virtual world), and “rationality” is a concept that emanates from the virtual world developer’s “real world” (consisting of stakeholders and need to sustain the stakeholders’ and developers’ businesses). Thus, even as developers designed worlds in a particular “planned” manner, they also learnt to engage with communities and to adapt to emergent phenomena, and to deal with their rational business and other real–world interests. This addressing of multiple perspectives may have sometimes occurred simultaneously, but could have also been addressed in contingent fashion, for example, only when tensions arise such as the tension between the inherently play–oriented nature of user–created worlds and the more rational uses of these worlds. As we will show, none of these perspectives can be adopted independently of the other, as each perspective is necessary for a world to reach and sustain a critical mass of users and activity. Eventually, to be viable, a world will need to find a balance across all three considerations.

Our three–way perspective sheds more light on why development processes can be characterized by the two endpoints identified earlier: designed and emergent outcomes. Design is necessary to put something clear and differentiated out for use, but emergence, always unpredictable, comes about because of the user–generated nature of virtual worlds, and because the eventually dominant “killer applications” are yet to appear [11].



3. Data and methods

Our sample consists of three worlds. Second Life serves as our benchmark, as it has a longer history than the other two worlds — Metaplace and Tirnua. It is a benchmark because it empowered user creativity to the greatest extent possible at the time, both economically (in terms of returns) as well as individually (in terms of the user’s power to create). Metaplace was a “light” (i.e., less immersive than Second Life) but Web–integratable virtual world, and Tirnua is a massively multiplayer online game with even greater limits on “user–generated content”. The latter two worlds are (or ‘was’ in the case of Metaplace) still nascent, but this makes them ideal for study as they are generally developed in a bootstrapping manner, with early features being put into immediate use for “play testing”, and adapted or evolved (e.g., through a long beta testing phase during which much of the development can occur). After our initial round of data collection, Metaplace closed its doors. We address the development process we observed as it is most relevant to a general understanding of governance. There is understandable concern among developers and in industry circles about the future of virtual worlds, although the more game–oriented ones can in general be quite profitable. Even though worlds are forming and collapsing, those events shape others involved in the industry commercially or personally. Of course, high rates of new start ups and failure is consistent with what we know about new organizations and technology (Aldrich and Ruef, 2006).

We have gathered most of our data from in–world chats, including ones conducted or observed at special events, as well as archived data including discussion forums (Metaplace); blogs, discussion forums and wikis (Tirnua); and, blogs and community mailing lists (for Second Life) [12]. We have also interviewed very active user–creators within all three worlds, and key personnel from Linden Lab (conducting three interviews and attending two meetings with Linden personnel). To triangulate on emerging issues and factors, we made extensive use of the archival data most “native” to each world’s interactions between users and developers — namely, their discussion forums. Finally, to confirm the more competitive aspects of virtual worlds and other business environment issues, we also interviewed principal personnel from two other virtual world startups, as well as two organizational users of Second Life. We observed the process of change in each of the three worlds through participant observation (including informal interviews with selected users of varying degrees of involvement) in the worlds and comparing the worlds and communities on a weekly and monthly basis. This paper is written as our data collection efforts continue.

While our research underscores independence of the development process of each world, at the same time, each virtual world exists in the wake of others preceding it (i.e., developers learn from others, and differentiate and build on the others). Thus, the other two virtual worlds in our sample have the benefit of the experiences of Second Life and other predecessors (The Sims Online in the case of Tirnua). This interconnection tends to confound the data, and so where possible, we lay clear the learning effects.

Table 1 indicates that product features, and therefore, the intended user experience (usage), varies somewhat across the three virtual worlds, with Metaplace and Second Life being the most similar in terms of how they enabled user–creativity. Table 1 also shows the major features of the platform (and major shifts in those features over time), the types of play encouraged (i.e., that result from premeditated design actions, or that emerge from user communities), and the nature of creativity (and community behavior relating to that creativity) in that play.

Table 1 illustrates three major points. The first point is that a range of types of social play are supported, and there is no dominant form of play. The second major point that Table 1 illustrates is the power and attraction (to users) of user creativity and personalization, even in an online game like Tirnua. Second Life was one of the first to prove that a user–led creative economy could be attained, albeit with strong ‘real world–like play incentives’ being given to the community (Ondrejka, 2004b). The Second Life growth path has generally plateaued. Even though Second Life’s creators’ content have helped enable a larger following community to achieve high levels of personalization, adornment and playing around anonymously (a form of play with one’s identity) (Ondrejka, 2004a), this by itself has not lured in a mass market following, and technical barriers to play continue to limit mass users from fully using the system, as well as possibly the premise of a fully immersive virtual “life” appealing to a range of users (Ondrejka, 2007).


Table 1: Comparison of the three worlds: Initial design and change as communities emerge [13].
Basic descriptionSecond LifeMetaplaceTirnua
Vision of product
(taglines on Web sites or in founder’s statements)
Your world. Your imagination. The “dream of seeing Metaplace worlds embedded all over the Web: on your blog, homepages, fansites, MySpace page, anywhere!” [14] “Play green, live green! … A friendly world of simulation games”
Basic designThree–dimensional (3D) perspective. Highly immersive and flexible virtual world2.5D (isometric) perspective. “Light” (simple) virtual world emphasizing socializing2.5D (isometric) perspective. Emphasizes social play on a multiplayer online game with limited user creation
First point
Types of play supported or that emerged
Various types of play emerged, including business ownership, socializing, exploring, role playing, hosting of in–world and broadcasted real–world events.Socializing and exploring was emphasized. User creativity (games and creative uses of space) emerged.Design emphasized environmentally friendly online game. Emergent uses not available yet.
Second point
User creativity supported
Greatly emphasized users’ ability to customize, e.g., via avatar and other content, and “intelligence” of objects.Limited user creation (e.g., changing user’s “world” and avatar clothing (more limited than Second Life).More emphasizing game play and user personalization, but some user creativity (clothing) allowed.
Means of enabling user creativityIntellectual property rights supported creative economy. Strong suite of user creation tools (building tools and the Linden scripting language for imbuing objects with behaviors).Currently (second year) at beginning of a creative economy. Introduction of tools for user creation (building tools).Started with none (user–content creation later improved to personalize play by using “Blender”, a third party tool, to personalize content).
Third point
Basic design features as the evolve from initial to adapting to emergent communities
Design features:
Initial features: general principle and tools to empower users’ creativity
Later features: Enhancing immersion (e.g., voice communication)
Initial features: building tools for world construction.
Later features: Things to do (e.g., mini–games, events, world building contests)
Initial and later features: mini games (to obtain currency), game systems, e.g., seeds (spawning flowers and plants), energy, etc.
Features responding to community, users wants:
Enhancing user experiences (ease of use), customization of platform to new and specific users“Fun” features added over time (e.g., game tokens to be used cross–world, avatar behaviors, customization)Customization of avatar (clothing)


The third point is that these products undergo high rates of change in the way they are developed and used, as communities adopt selected features more than others. While Second Life had a similar open–ended creative environment to Metaplace, it faces the issue of turning off some of its communities whenever it had to cater to other communities. This all goes to prove Malaby’s (2009) point that play has a performative aspect, which makes it hard to cater to all tastes. For instance, in Second Life, early communities illustrated the extreme uses of the platform, but these were eventually obscured by a deluge of more mundane activities like socializing in night clubs. This partly occurred as the result of a design change, namely, the creation of a mini map that showed “hot spots” of user gatherings (Au, 2008). These uses are not always emphasized in the early design, and may be emergent in which case they will have to be managed later by developers.


Figure 1: SL interface - Jordi Comas avatar
Figure 1: SL interface (Jordi Comas avatar).



Figure 2: Screenshot of Metaplace Halloween party and user interface
Figure 2: Screenshot of Metaplace Halloween party and user interface.
Source: Authors.



Figure 3: Screenshot of game of tic-tac-toe in Tirnua
Figure 3: Screenshot of game of tic–tac–toe in Tirnua.


While following in Second Life’s footsteps, Metaplace’s developers undertook specific design decisions which limited the degree with which its community could become empowered, creative and successful. Nevertheless, this has still provided the community with enough creative room, and by 2009, the latter started creating its own trajectory in terms of social play, including the scheduling of events for friends and the like.

Tirnua was more bootstrapped (i.e., started from limited resources) than the other two platforms, but by 2009, Tirnua was looking more like a traditional online game or at least, like its predecessor, The Sims Online [TSO], although like TSO, it also did not possess the same kinds of objectives as MMORPGs. Its method of development was not unlike that of Metaplace in that its developers sought to cater the game to users’ preferences as much as it was to the developers’ intent.


We analyze each of the virtual worlds by breaking it down into three elements: the key actors affecting its evolution, the interactions between those actors, and the outcomes (e.g., actions, behaviors, features or objects such as content). Firstly, the key actors are the developers themselves (including designer/developers, head of the organization, and community managers), the users or users and the community they form (including lead users), and other stakeholders such as financiers (all this seen in all accounts). This division into developers and users follows from the conventional user–developer classification employed in the new product development literature, though the types of interactions themselves may vary depending on the product (von Hippel, 1986). Secondly, we will examine the interactions between developers and users, including feedback, brainstorming and monitoring. This includes interviews and discussion forums or wikis where these interactions were seen. Finally, the outcomes are the effects of developer’s actions, either being features of the platform, or the community’s actions. Outcomes may also result from the interactions between developers and users. The design features are divided into platform features that are “coded” (i.e., programmed) into the structure of the platform, and features that exist in the human realm of governance (e.g., policies such as rules on intellectual property and economic exchange, and roles such as those taken on or assigned in the community, e.g., community managers).

From our examination of secondary and other data (including users’ accounts of their experiences) and the literature on Second Life (e.g., Au, 2008), it appears that the features on the platform itself as well as economic (i.e., intellectual property) and regulatory (e.g., policing and banning) policies helped shape the users’ behavior, and therefore, the world. Thus, we will pay some attention to policies as a potential source of issues.

In a sense, virtual worlds are no different than any other product that has a feature set, a product development process, and a focus on users during the development process (see for example, Bartle [2004]). According to Bartle, however, they differ from games in that “Usually, computer games go into beta testing as late as possible. Virtual worlds … go into best testing as early as possible. This allows for bugs and exploits to be discovered well before paying customers leave over them, all the while forging strong community bonds between the beta testers.” [15] The primary difference with this is that in at least two of the virtual worlds we cover, the communities are co–creators in their experience and content in service of that experience. Furthermore, developers are not clear on what the best uses (and most apt designs) of their virtual worlds are, and this is in the hands of the community. Never mind that these best uses may also vary across virtual worlds. As a result, communities are both a challenge to the development process and necessarily involved from the start.



4. The design perspective

We will first examine the development of the virtual worlds in our sample by using a conventional lens: the product development process (akin to the one laid out by Bartle [2004] for general virtual worlds including MMORPGs). Through our data, we develop an approximate understanding of how the designs were “planned” out, the circumstances of that planning, and the results (e.g., whether there were design changes). This helps us to highlight features that are general as well as specific to those worlds.

We describe our three virtual worlds according to two “conventional” sets of characteristics.

  1. The first is the overall design approach and product development process. There are subtle variations in the product development model for each world (and a change in model for Second Life), in part involving the way in which platform features are designed or “evolved” emergently, and how designs promote creativity which in turn feed play.

  2. The second is the set of specific design parameters that, amongst other things, will “govern” play. This includes specific parameters that designers set or shape, like policies and “rules”; “tools” hardcoded in the code itself; and, parameters that aid the community in self–governance (e.g., the restrictions on play or on a user’s power that an owner of a parcel of land can set). To be sure, the design itself governs play, as for instance by creating an envelope for possible [emergent] types of play and creativity. Finally, the development process itself can “govern” play through facilitating interactions with community, and through judicious timing and adaptivity.

A summary of each world’s development model and experience follows.

Second Life’s development model: Content and the Wild West

Being among the first products to break ground in the user creativity/empowerment front, Second Life’s early development model was less about designing for specific business goals than about aiming for general principles. The first few years (from 2003 through 2007–08) were arguably more about empowering users’ creativity to the maximum, and less about the type of community “desired”, or the revenue being generated from those communities. Linden Lab was also actively promoting the high growth rate of the world in the public sphere, though possibly more as an advertising line than as an explicit “focus.” As a Linden Lab interviewee noted: “We deliberately stayed away (from communities).” Linden Lab made an overt decision in the beginning to leave the communities alone. This did not mean that communities’ and users’ needs were not taken care of, but just that developers did not actively engage with them.

Linden Lab sought to provide for the emergence of very different types of play in the various communities by providing a suite of powerful user creation tools that allowed users to create not only content but to imbue those with “behaviors” (i.e., through the Linden scripting language). However, even this focus was insufficient to promote a strong user–created bent in the user base until Linden Lab chose to award “user–creators” the intellectual property rights of their creations (Ondrejka, 2004b). Together with the monetary reward to users for engaging in IP creation activities, this dramatically shifted how the platform and its content were viewed by the community. Second Life still derives much of its original cachet from its creators and the content economy. Interestingly, no more than 10 percent of Second Life’s population consisted of the content creating community (according to one informal survey by Au [2008]), though a large proportion fell into the “fashionista” category (i.e., those who purchase content and adorn their avatars for status and other social reasons), as well as “explorers” (those who like to look at environmental content) and “socializers” (those who socialize in communities) (the latter two come from Bartle’s taxonomy for players in multiuser dungeons) (Bartle, 1996).

The limits of design: Reining in the Wild West

At the same time, this maximal empowerment of users led to another outcome: the emergence of new user creations and behaviors — both positive and negative. As noted earlier, more than with most worlds, Second Life has often faced unanticipated or emergent outcomes generated by its user communities (Au, 2008; Malaby, 2006). Second Life has also suffered an inordinate share of “griefing” (users harassing other users) and racy behavior (i.e., “adult content”) — all of which Linden Lab has tackled head–on through policy changes, sometimes enforced in the code base.

Linden Lab’s earlier “ignoring” of its communities led to the latter being so self–referencing and self–involved that Linden Lab effectively had no control over them, and would eventually have even greater trouble relating to them. An interviewee with Linden Lab pointed out to us that “they would like (nothing more than for) us to leave, lock the doors, and leave the servers running.” This gap is reinforced on the community side by community blog posts and even departures of “key” members of the community. There is by now clear recognition by Linden Lab that the gap between the communities’ perceptions and Linden’s intentions is undesirable and has to be addressed.

At the same time that this gap was growing, Second Life’s “hype cycle” peaked, its usability continued to be called into question, and revenue became a preoccupation (Gartner Group, 2008). With the concern for revenue came an increasing recognition of the importance of the potential role of corporate users in that. There was also pressure from other virtual worlds and forms of social media in terms of competing for corporate attention and for users.

As a result of all questioning and competition, in roughly the last two years, the internal view of development has shifted dramatically to one of a more purposive design methodology. By 2008–09, Linden Lab realized that they could not go on as usual with an unscripted development process (personal interview). Linden Lab realized that they had to be in the driver’s seat by strategically engaging specific communities. There was also a view that a change in leadership was necessary to move Second Life to the next level of community size and performance.

Partly as a result of the publicity and to assuage its corporate users and stakeholders, Linden Lab became proactive in its design intentions. As part of this, Linden Lab has started to implement policies that dictate stronger control over content and “objectionable” content for example by moving adult content and activity onto an “adult content”. This process was done systematically, involving user feedback, such as the use of the company blog and in–world “brownbags” (i.e., focus group and feedback sessions) [16].

Metaplace: A working design model?

Metaplace sparked interest in the virtual world community because it was primarily play–oriented, Web–based, and was designed to have a low learning curve. Like Second Life, it also purposefully set about stimulating communities. The simplicity and low learning curve promote a “light” feeling to it. Raph Koster noted in a public interview: “from the get–go, we’ll be a highly social, creative place with great ease of use, and anyone familiar with casual virtual worlds and ‘building’ sorts of games and worlds will feel at home.” [17].

The design for the platform was clear from the onset with feedback provided to players detailing their “status”: Status was represented by a power bar on the user interface. In this sense, Metaplace was more on the “planned” end of the “unplanned to planned” spectrum. The development process involved a fair number of feature road maps being rolled out to users, as well as solicitation of user feedback before major additions or changes. While Second Life appears to offer a contrast by emerging on the unplanned end of the spectrum before moving towards the planned end, the truth is actually trickier. Second Life’s aim to cater to the most creative of creators, and its being the first world to develop this goal, ensured that whatever “planning” it had implemented looked “unplanned” as “user forces” took over (see Au [2008], Malaby [2006] and Malaby [2009] for examples).

It may be posited that due to its limited creative aspirations (or aspirations to provide to its users), Metaplace was able to “control” the outcomes from its design and development process better than say, Linden Lab, in part because Metaplace’s developers largely controlled the nature of user creativity, and thus, the emergent nature of outcomes and behaviors in the world. While there is an apparent limit on user creativity, an interviewee from a competitor to Metaplace did note in an interview with us: “I find it amazing that Metaplace could still get the degree of creativity that it has with its 2.5 D (2.5 dimensions, in effect, an isometric view) platform”.

Limits to the design perspective

On the other hand, despite their commitment to design planning, the developers still had to be adaptive to the community. For instance, developers had to respond to “insufficient things for users to do” by seeding the game world with game tokens or fundamental mechanics which users are then allowed to embed in their worlds in their own creative ways. For example, a “golden egg” was released which cost a substantial amount of in–world credit to buy, but which could be “hidden” as part of game play, and used to increase traffic to a user’s world. By introducing features at different points in the development cycle, Metaplace enhanced a design–centric approach by making it adaptive. These changes did not in themselves affect the user experience too much in part because of the additive nature of their introduction, but also because Metaplace was in an early stage of development, where changes were expected.

Metaplace also appeared to learn lessons from other virtual worlds like Second Life and massively multiplayer online games (MMOs). Metaplace’s head — Raph Koster — is a well known MMO designer; and Metaplace regularly benchmarks against Second Life and other worlds. Metaplace tried to engage creators both in managing the generation and flow of content. A typical post on the developer forum went: “If you’re a former or current player of the following games/virtual worlds, we’d love to have a conversation with you to talk about your experiences with Metaplace … . [18] In summary, Metaplace was intended to have a cleaner birth as a world than Second Life, because it knew what it wanted to be and was designed to be simpler, but also because its development process was inherently flexible, responsive (to users), and adaptive.

Tirnua: Game development on the fly

As mentioned earlier, Tirnua differs from the other two worlds we study in that it is far less about user creativity and more about a designed game. That is, users have to “work” at playing a game to create a desired lifestyle as reflected in their homestead. However, users are given substantial leeway to personalize their world, to engage in various activities, and to socialize with one another. Like Metaplace, Tirnua started its development cycle and beta testing almost concurrently, involving the user community from the beginning of development. But this is where the similarity ends. Despite its more recent origins, the development process that Tirnua has thus far gone through appears to be more ad hoc, or at least, more unpredictable [19]. As noted earlier, Tirnua was a more bootstrapped product in that Tirnua’s team did not obtain funding at the beginning, while Metaplace had venture capital backing from an earlier stage. The platform’s primary experience involves designed game play, and games generally require great amounts of “tuning” or iteration to attain a desirable experience (Tschang, 2005). As such, users’ reactions to design changes can be no more predictable than an open–ended virtual world’s. Part of Tirnua’s development team was coming off The Sims Online project whose user base never reached beyond the low hundred thousands. According to some observers, this was partly due to flaws in the designed experience (Ludlow and Wallace, 2007). Thus, Tirnua’s development process appears designed to handle these flaws at an earlier stage where they could be handled.

Like Metaplace, Tirnua developed features in an incremental fashion, but since some of these are actual working “systems” within what is essentially a game, the developers have had to layer these on as complex and interacting subsystems of circulating flows (e.g., of money, energy, food, etc.). Thus, while Tirnua started out with the intention to foster a “greener world,” the earliest subsystems put in were mini–game playing and a planting module. By late 2009, Tirnua returned to its original greener theme with the proposed addition of newer energy and other subsystems. This has created a whole new set of “game playing” expectations with users, who were previously either gambling, engaged in social chat, or doing in game “work” (e.g., by running the “playing at work” mini–games provided) [20].

Commonalities and differences across worlds

As can be seen from the discussion of the three worlds, while there are obvious dissimilarities in the experience, at least four elements are common to the design and development models of all three worlds.

Firstly, all were designed and developed in an incremental fashion (i.e., features were added to over time) [21]. It is not clear whether this is partly a result of the funding model, being staged or low budget, or because the development process needs as much to be “dynamically managed” as planned out. The development team’s experience with other worlds matter as it provides a reference point for how to proceed with development.

The development process in all three worlds, especially for the two under beta testing, and even for Second Life recently, has been to plan for, design, and implement new features or subsystems over time, incorporating them into the overall game play as world development proceeds. This includes adding features and layered systems, that is, socio–economic systems that are added to the world one at a time. While this model seems common at least to the virtual worlds under our study, it has been almost unheard of for single–player packaged games and even virtual worlds to be developed “on the fly” (Bartle, 2004).

Second, unlike games, it is critical for developers to interact with their communities during development (game development also involves interactions with user communities, but not as a part of the design process). As we will show in the next section, the means of interacting with communities across all three worlds are not that different, though one world or another may or may not use a particular means. What matters is the timing of those interactions. For instance, Metaplace’s proactiveness in addressing this at early stages appears to be leading to a predictability in its development path, although the product still failed. In general, it is differences in the timing of the development process that has led to eventual differences in how worlds look, and therefore, how developers further develop the process.

The third issue is convergence or divergence in design. The worlds and the emergent behaviors from communities will of course vary depending on how the virtual world is designed to enable specific outcomes. However, what seems to be converging in all three of the worlds that we studied is the recognition of the importance of some kind of content creation model, if not an entire economy hinged on user creativity. This focus on content may be as much to do with the common nature of virtual worlds — being spatially–immersive experiences and social in nature. This user creativity adds to the unpredictable nature of communities and their behaviors. This emergent nature of world evolution has to somehow be taken into account by developers in the development process. However, the degree to which user creativity was empowered across worlds varies considerably, and thus, can affect the impacts of users to the world [22]. For instance, Linden Lab’s unleashing of substantial user creative forces caused a wider variance in user behaviors than seen in other user–created worlds, and the more extreme of these behaviors may have ultimately helped Linden to decide to change their development model [23].



5. The role of communities and designing to communities’ needs

In the last section, we depicted the development process of our sample as involving developers making sense of virtual worlds that are emergent in nature, and how the design perspective limited developers from doing so. This is in part due to what communities make of the world. The central role of communities to virtual worlds has implications for how each world’s development model treats the governance of play and creativity. In this section, we will illustrate the key roles of communities, these being: (1) how communities form and shape the value of the virtual world to developers, and (2) how developers engage with communities in order to improve their worlds. As we show, communities not only create the world and are central to users’ experiences (e.g., through participating in the world’s governance), but are also an important source of information and effort to developers and external parties.

The lived–in, user–created, and community–infused nature of these media can be quite different from how communities are treated in traditional Internet platforms like discussion forums, or newer media like blogs or Twitter (Bruns, 2008). This is because of the spatial and (software) architectural nature of virtual worlds, which creates a platform for real–time avatar–to–avatar communication. An earlier call was made to engage in the participatory design of MMORPGs (Taylor 2006a; 2006b). While this was focused on the experience of MMORPGs, we will examine our data on user–created virtual worlds for furthering our understanding of the nature of participatory design, especially in the context of their communities.

Origins of community

The initial question we face has a fascinating answer: where do users and communities come from? While this is less of a direct “design” issue, it is a major consideration, as developers need to know where their users can come from. A common observation is that many members of early communities of many virtual worlds come from other virtual worlds. The reason for this may be that older virtual worlds close down regularly, and new worlds require early users who have enough experience. In Second Life, users migrated in from other worlds, while in Metaplace, some early users came from Second Life; similarly, in Tirnua, many early users came from The Sims Online (TSO), which Tirnua’s founder was the program manager for (Pearce, 2006; Au, 2008). Tirnua has in fact been developed along similar lines to TSO. One virtual world developer that we interviewed commented that their world benefited from another’s closure: “We noticed that our number of user registrations shot up 15,000 the week after The Sims Online closed down.” This despite the fact that their world did not offer users specific things to do, let alone the game–like things that TSO offered in game play. The interviewee went on to say that they also advertised on Facebook, and found it more effective than a leading Web advertising venue, in part because advertising done on Facebook tended to target an interactive–savvy audience, and so allowed them to “screen in” users interested in “interactivity”.

There is a critical mass threshold for worlds to become sufficiently engaging and hence have a greater chance of surviving. Over time, the formation of a community can lead to a “lived in” world, spurring on world usage and user creativity as an outcome. Second Life is the clearest example of this, but developers of other virtual worlds can still rely on the “lived” nature of virtual world existence as one means of binding communities together. The diasporas from worlds such as Uru, TSO, and into other worlds are further evidence that in those worlds, a sufficient level of cohesiveness emerged for such self–organizing communities. That same degree of cohesion provides a lived–in feeling for residents, lending the world a “persistent” quality.

The value of communities to virtual worlds and developers

Communities have at least three functions in virtual worlds’ development. The first (and most obvious) is that creators in the community create the world, that is, by creating content and originating or reifying behaviors. In this respect, developers have to respect the communities as true co–participants in the creation of the virtual world. It is through the users’ toil that cities or other patches of dense activity spring up, acquire lived–in status, are well–maintained, and are socially vibrant. At the same time, when communities are left to their own devices, highly creative and sometimes excessive behaviors can result. This excessiveness can lead to a sense of entitlement or empowerment among users that is actually more than developers and current legal standing allow or enable. These frustrated ambitions are part of the impetus behind such normative perspectives as the declaration of the rights of avatars (Koster, 2000). The excessiveness can also result in communities having such high barriers to enter that newcomers feel unwelcome. It can also involve behavior considered deviant enough to cause public relations headaches for developers, e.g., the issues of virtual banks and sex on Second Life, which ultimately led, respectively, to, a ban on banks without a real–world banking organization behind them, and confinement to a separate continent for adult content.

A second function of communities is the help they provide in governing users’ play or even in self–governing themselves. Naturally, when worlds start to add more and more players, we start to see a diminishment in how much governing developers can do, and we start to see an increase in community self–governance. In the case of Second Life, volunteers as well as island owners alike participate in governance, but Linden Lab still remains as ultimate authority, e.g., egregious cases requiring banning of users. In the case of Metaplace, an increase in new players beyond the hundreds of users also caused developers to try to engage lead users (i.e., advanced users) in governance efforts (the idea of lead users is a general phenomenon identified by von Hippel [1986]).

A third function of communities is that they help to inform developers on how better to develop the world. As pointed out by von Hippel (1986), Jeppesen (2005) and others, an active community of lead users can help developers to “make the system more usable,” or in other words, make the system more user friendly, but even beyond this, they can also suggest improvements or new features. The external “value” (as possible clients of potential paying customers) that communities provide to the virtual world developers is also of importance, but this is treated later in this paper. The community provides some measure of product success, and signals the virtual world’s “worth” to external parties.

Taken together, these three functions of communities make it clear that a healthy community contributes to a “healthy virtual world”. Developers hence have the challenge of ensuring their community’s contentment or satisfaction at all stages of the virtual world’s development. The evidence from our sample is illustrated in Table 2 and in the text that follows.


Table 2: The functions of community.
 Second LifeMetaplaceTirnua
Creating and instantiating the worldMaking the world lived in, e.g., through lead users promoting own events, worlds etc., acting as content creators, business owners, etc.Making the world lived in: e.g., through lead users organizing events.Making the world lived in: Engaging in game play associated with “living in the world”.
Governing the worldGoverning functions (e.g., by island or “sim” owners). Volunteers helping newcomers.
Community manager: present.
Governing: lead users designated by developers to act as “moderators” with some powers (e.g., to silence disruptive users’ speech).
Community manager: very active.
Governing: Lead users volunteer to help newcomers.
Community manager: none.
Users helping develop the worldDevelopment: interactions with developers; lead users directly developing “plug–ins” (new features in code).Development: interactions with developers (often via community manager); lead users directly developing “bridges” to other worlds.Development: interactions with developers.


Creating (and instantiating) the world

Virtual world developers rely heavily on communities, especially early on, to build the world — both in terms of (creating) content as well as in terms of (instantiating) play. The design perspective previously discussed suggested that virtual world policies, user creation tools and socio–economic rationales all can help to enable, incentivize and ultimately, “govern” user behavior and creativity. How Second Life’s content creation tools and scripting language helped fueled users’ creativity is well known by now. Similarly, in starting to foster Metaplace’s content economy, Koster noted: “our virtual goods aren’t necessarily just pictures, but can enable unique behaviors and interactivity.” [24] Thus, tools and the broad types of content available are all intricately related to the creativity of users. It is becoming a convention if not a necessity for a content economy to be incorporated into a virtual world’s design early, so as to function in conjunction with the community to make the world interactive, enjoyable (for status and other reasons), or otherwise playable. Tirnua also illustrates this point as a “limit case” (or case of a product not falling into the typical model, but that still fulfills the condition). While Tirnua started out mainly to be a game that initially did not promote user–created content as heavily as the others, its developers have nevertheless found that some user–customization (as part of the personalization of the user experience) is necessary for users’ satisfaction. Developers’ introduction of a custom clothes creation tool appeared to unleash a flurry of clothes creation: “In just a few hours, thousands of new clothes have been created by Tirnua designers in the game” (as well as engendered a number of positive user comments) [25]. Similarly, Metaplace created content specific to themes like Halloween which helped to create an atmosphere for users at events themed for those times.

In the end, a content economy may be a necessary but not sufficient condition to a virtual world’s success (this latter point being illustrated by Metaplace’s viable design but eventual closure). However, as with any system of game play, incentivizing content creation is also a social process, and requires deep understanding of the community. Newer virtual worlds may sometimes go to the other extreme and enhance creation but do not create the economic means for content creators to “profit” from their creations, or an attractive enough social reason for content creators or users to flourish (which may in turn impact on the participation of content creators). As Table 2 shows (and as others have noted, e.g., Ondrejka [2004a]), users act not only as content creators but also as business owners (who create and distribute content), as monitors (of behavior) and volunteer guides, and as creators of events. Thus, the community members may assume particular “roles” as they play.

The knowledge of lead users has always been considered important in many if not most industries, since such users are very knowledgeable of the various uses of a particular technology (von Hippel, 1986). This is no different with user–created virtual worlds except that the role of the community (and users) is even broader. They create and use content, and police the world, amongst other functions. For developers and scholars alike, we must be aware of not a dyadic relationship between developer and user, but the more complex triad between (lead) user and community, community and developer, and developer and (lead) user.

Communities and governance of the world

Communities can govern in benevolent manner (e.g., by acting as volunteers), as well as in authoritative manner (e.g., engaging in policing action). The former is common to all virtual worlds, for instance: “Users also run classes and events to ensure that new residents understand how to create and customize within Second Life.” (Ondrejka, 2004b) The policing notion of governance was perhaps most commonly seen in Second Life, which was well known for having users act in policing their own properties. Roles such as “security” or “bouncer” were common to social venues attracting people “off the streets”. In addition to communities acting on their own as agents of governance (i.e., acting in self–governance), developers can also directly engage communities. In fact, community managers can be effective in engaging users for this, as happened with Metaplace. For instance, we observed Metaplace’s developers engaging lead users in helping to manage disruptive users’ behaviors. In an in–world discussion on disruptive users, Metaplace’s developers visibly empowered its lead users by going beyond the traditional developer–user boundary and “deputizing” lead users; lead users were asked to become mentors as well as to monitor disruptive users. As of this writing, we had not detected similar patterns of developer–user engagement in Tirnua, but Tirnua’s predecessor, The Sims Online, had also faced serious issues of disruptive user behavior.

It is well understood by now how users in worlds like Second Life help to bring about the first two functions (creation and governance). The third issue is perhaps the least explored in the literature — that of how users help developers in the development process. Given that ours is a developer–centered study, we will focus more on this category in the following subsection.

Designing in relation to the community (or how users help develop the world)

Having established the fundamental importance of community to the world, we now examine how developers actually develop with the community, and in relation to its interests. Table 3 below illustrates the three main ways in which virtual world developers interact with communities.


Table 3: Community interactions with developers.
 Second LifeMetaplaceTirnua
Developers engaging community in design.Active engagement with users on forums and blogs.In–world chats with lead community and active developer participation in–world and on forums, e.g., in–world chat on integration with Paypal [26].Design discussions on wiki and forum (where users have proposed features); lead users said developers appear in world, but prefer feedback on wiki.
Developers monitoring communities.Very significant investments in “analytics” (e.g., “mining” stored data on user behavior for trends and outcomes of different designs), data warehousing, community managers to monitor behavior, etc. (but few in number relative to population).Koster noted in an in–world discussion that they tested for user “stickiness” using various analytical means: “What we do here at MP is to run acquisition tests, largely using AdWords, but also using things like events and remote embeds.” [27]Not observed firsthand or in data.
Developers adapting to emergence.e.g., Focus on community’s interest in events (e.g., performances).e.g., Metaplace visibly increased the number of events (including contests) over time; also modified the central meeting place, user interface.e.g., Developers acknowledged the importance of Facebook as origin of users/users’ alternate social media by linking to and advertising in Facebook.


Developers engaging communities in design

Engaging community involves, at its most basic, making the community feel that its voice is heard. This includes the raising of topics and steering of discussion and even roles of the community towards the developers’ ends. The servicing of communities’ needs can be a difficult proposition as the persistent and lived–in nature of virtual worlds makes it harder for developers to make shifts in their platforms’ directions without upsetting their communities or users. Perhaps equally important to the type of interaction is its timing.

In general, developers also prepare the community for upcoming feature additions and other changes. In Metaplace, four to five regular “development roadmaps” were rolled out during the beta phase to allow the community to see where the platform was headed. Linden Lab did this early on in Second Life’s development cycle, stopped doing this for awhile, but then continued in 2009 with a content management roadmap [28]. Having said that, these road maps were not much more than feature lists, and Tirnua was doing the same, albeit more informally.

Another level of involvement is to involve communities with the process of determining upcoming changes — either by brainstorming, by gaining targeted feedback (e.g., the ‘what do you think of this’ type of question) and by unsolicited input. Feedback is important, but developers may even go beyond soliciting user feedback to engaging users in information retrieval. Lead users do provide feedback to developers, whether the latter want it or not. It is also worth noting that despite its initial hands off attitude, Linden Lab did also attempt to learn from individual users. In its earlier years, Linden Lab started engaging with select groups of users in on–site visits. For instance, an interviewee noted that, up till 2008, Linden Lab was able to fly in several members of the community (from a diverse spectrum) each time to their offices, to get a good “scan” of the users’ needs.

As a later (than Second Life) generation world, Metaplace was particularly adept at garnering feedback, and raised new topics in preliminary brainstorming (with lead users) on topics such as disruptive users and the monetization of users’ creations. For instance, at an in–world meeting on disruptive users, Koster asked a lead user: “Michael, can you send over a little doc on everything you have ever [sic] tried at There, and what worked best? No need to include the bad ideas. ;)”. [29] Topics such as these were deemed crucial to getting the world started properly. Since lead users are generally the most widely exposed to other worlds and experiences, and the most facile with the platforms, developers’ rely on lead users. Lead users’ knowledge from wide and varying experiences can help developers assess the working parts of other virtual worlds in order to help assure themselves of a working model.

Size limits to interactions

As the size and complexity of a world increase, we see developers turning more and more to monitoring communities as opposed to engaging them. The inherent nature of the community–based approach is that relationships between developers and lead users and communities are easier to manage while the virtual world is smaller, but once the world gets bigger, many of these interactive mechanisms for engaging and learning from communities are not scalable. In addition, smaller samples of lead users do not seem to yield as much benefit. Once Second Life became larger and more diverse, and once more target communities with unique needs were identified, the flying–in of lead users as focus groups became less viable for Linden Lab as a means for finding ways to relate to each community specifically. According to one Linden Lab interviewee, after the number of user registrations shot up from 200 thousand to two million in one year, they stopped doing this. On the other hand, Linden Lab’s forums are expansive, and their developers actively engage on them. They also found it easier to use metrics and analytics to measure user satisfaction.

Developers monitoring community behavior

A more advanced form of feedback involves developers’ monitoring of users. Recently, Linden Lab has also turned to interacting more with its communities, and through the use of analytics, to understanding the effects of their policies and other actions on communities. However, it was acknowledged that relating to the communities was still a challenge for developers. As an interviewee pointed out, to help it understand what was happening in Second Life, Linden Lab made a huge investment in 2006 in data warehousing and “analytics” (involving the processing of a huge amount of behavioral data on users). They now rely on this infrastructure to help them to determine the success of potential new features even as they were being implemented. The focus on analytics helped Linden Lab to determine how successful their actions (including those with regards to emergent trends) could become, including the return on investment for various decisions. In other words, Linden Lab has chosen to focus on being more “proactive” with design one it had in place systems for regular monitoring. Similar methods appeared to be adopted by Metaplace, as the metrics used in their user interfaces appear to be derived from quantifiable measures. For example, early on, Metaplace adopted Bartle’s typology of users (the user’s activity in each “role” shown to him or her on the user interface). Table 3 illustrates another example that Raph Koster noted.

Developers adapting to emergence

We have illustrated how communities self–organize themselves in highly creative and emergent ways for play, and how developers have essentially latched on to these phenomena in order to ensure that their virtual world expands its “frontiers”. We can define this emergence as the following: “Emergent behavior occurs when a set of rules interact in interesting and unexpected ways to allow experimenters and innovators to create truly new creations.” (Ondrejka, 2004b) It is worth noting at the outset that such creative play can both, add to and modify the world in aggregate, as well as upset other users’ experiences (or even upset others in the real world). There are many examples of the first category, and of developers’ adaptation to those emergent uses of the platform as pointed out earlier.

In fact, developers can be as awed by the power of users’ creativity as anyone. As Koster notes of the potential of the Metaplace community: “I also expect users to take us in many directions over time — that’s the beauty of enabling user–created content.” [30] At the heart of user creativity in these new environments is a simple observation:

“the distinction between mods and the Ultima Online piano illustrates a critically important point. Mods allow the creators to actually change the behavior of the game. The piano, on the other hand, is not a piano and cannot be played. It may look like a piano, but it is only a stack of crates and fish steaks. This is an excellent example of the difference between crafting and creation.” (Ondrejka, 2004b)

Example of adapting to events: Developers handle emergence in different ways, but the story of how Linden Lab adapted to events suggest how this manner of “designing to emergence” happens. Malaby (2006) cites how in Second Life’s earlier phase, Linden Lab adapted by capitalizing on emergent phenomena to further advertise the platform’s creative power. Events (ranging from product launches to telecasts of real world events, speeches, and concerts) have also become an important activity on Second Life as with other virtual worlds. A Linden Lab interviewee noted that while events had been going on for a few years in Second Life, Linden Lab recently noticed just how major an activity events had become. They responded by designating events as an archetype for further support and promotion. Linden typecast event–goers as one of several types of user. They started to support this activity by making it easier for users, even casual users, to log in to watch events. More generally, while Linden Lab initially simply allowed events to happen, they now explicitly recognized the need to design more systematically and proactively, and especially to focus on what emerges. Finally, we should note that such adaptation is not always appropriate or efficient; what counts as “cool” can vary widely, especially as the world grows. For instance, as noted earlier, the proliferation of dance clubs in Second Life was considered quite banal and a distraction from potentially more interesting user activity (Au, 2008).

As shown in Table 3, Metaplace recognized early on in its product lifecycle that events could drive community interest in the platform. In this regard, Metaplace has been more proactive than Linden Lab. Indeed, Metaplace’s focus on events has led to an upswing in community events. They have also developed tools that facilitate events. Metaplace appeared to be hinging on events as a major application, and both platforms have made it easier to search for events.

Example of adapting to broader emergent trends: A significant issue is how companies connect to other trends happening outside the platform and in digital society at large. According to our evidence, all three — Second Life, Metaplace and Tirnua — are shifting their designs to recognize and accommodate the influence of non–virtual world social networking sites or communication platforms like Twitter and Facebook. For instance, many virtual worlds like Tirnua (with Facebook) and Second Life are starting to provide links to these two fairly dominant social networking platforms as well as across themselves.

Increasingly, virtual worlds are quick to follow emergent socio–technological trends. In the virtual worlds we followed, developers tracked inter–world communication, the emergence of Facebook, and Twitter, as a convergence of media streams (to varying degree, depending on the world). In Second Life as with other virtual worlds, the issue of open sourcing has also been an important means of expanding the frontier of interactions between the community and other communities. For instance, Linden Lab faced a choice to either open source its ‘back end’ code (thereby inviting direct competition) or to open source its standards. It chose the latter, which restricted direct competition, but allowed for particular competition such as the open source alternative OpenSim. The main concern with this is the perceived loss of platform control.

Metaplace also tried to initiate conversations with users on the need for a cash–out system (a system to allow exchange for real–world currency). Part of the motivation was to preempt users’ creativity, as when a lead user built a link to the Paypal payment site. The discussion went: “Looking at Obo’s PaypalActive plugin … His plugin allows you from script to send someone away to Paypal … First of all, we have to be very cautious about paypal payments being tied with Metaplace official systems … .” [31] By this, Koster was referring to the inherent possibility that, amongst other things, users (thinking they are dealing with an in–world currency system) might be linked out to a real payment system without realizing it. Interestingly, the efficacy of all links to other social media cannot be foreseen — again, proving that emergence tends to have a winning hand. Tirnua relied on Facebook for an influx of users or interaction, but some users actually disliked having the link, as they wanted their real–world identity in Facebook to be separated from their in–game identity in Tirnua.

In general, we see that external trends amongst users and technologies affect emergence. Virtual world companies have to adapt to these the best they can, hopefully while recognizing any potential downsides in the process, e.g., expected returns and loss of control over their platform.


We have now illustrated how developers of virtual worlds have handled the need to go beyond design as in “what is in the plan” and into richer and more multi–faceted relations with the communities that emerge in their worlds. All the while, they must contend with the broader set of emergences in cyberspace or the rest of the network society such as the rise of social networking platforms. We have also illustrated how worlds and communities are governed not only by the platform’s design and policies, but also by the way in which communities emerge initially, and by the subsequent interactions among world developers and communities of users. The development process adapts to communities as they engage in creative behavior and world–building — at times by restricting user behavior (in egregrious cases) and at other times by trying to promote these innovative behaviors. As pointed out by an interviewee from Linden Lab, some forms of emergence simply cannot be anticipated by all the analytics they have, and have to be adapted to by developers.



6. Development in the service of economic rationality

As this article was being finished, we experienced a mix of fascination and concern when Metaplace announced that it was closing its public virtual worlds spaces: “Unfortunately, over the last few months it has become apparent that Metaplace as a consumer UGC service is not gaining enough traction to be a viable product, requiring a strategic shift for our company.” [32] This suggests the dominance of rational interests over all others, and the continuing lack of a viable business/revenue model for most virtual worlds, though as pointed out by a developer of another virtual world, and judging from the failure of other worlds such as Metaplace, and more recently There and Vivaty, funding has been generally scarce since the economic downturn (For a summary of virtual worlds as of 2008, see Spence [2008]).

We have thus far been examining virtual worlds through the eyes of developers and the communities of users. Now, to explain this turbulence seen at the industry level and amongst individual firms and their worlds, we will turn to the final piece of our emerging framework: the (real) world external to the virtual world. Naturally, the development companies exist in this world. The companies that create the virtual worlds of Second Life, There and Metaplace typically rely on venture capital, and they have been under increasing pressure to find business and revenue models that work. This aspect of venture financing and the high capital investment involved (in the tens of millions of dollars) means that the developers have to focus on the following factors of economic rationalization:

  1. The growth of the worlds.
  2. The value proposition, or providing of enough value and relevance to consumers, both individually and as members of communities, such that they are willing to pay for it.
  3. Sources of revenue, both internal (e.g., the consumers) and external (e.g., advertisers).

Design goals

Economic rationalization can matter for design goals when both interact to either promote one another, or in causing tensions, e.g., by interfering with communities’ “play” in the worlds, or by requiring developers to adapt to a continual barrage of emergent trends as well as a perceived “natural” growth path.

At the same time, as we will show, developers may have adopted a particular stance that may affect their ability to pursue particular design paths later. For instance, Linden Lab’s initial hands off stance has shaped its future abilities to deal with its community, whereas Metaplace’s proactive stance could be argued to be doing the opposite.

Growth, value and revenue

Growth, revenue and value go hand in hand. Without value, there are no new users (“growth”), and without new customers or value added, the potential revenues may be flat or non–increasing. It is ironic then that virtual worlds have been focused on user population (i.e., community) growth (in part because it is an easy measure of success). As virtual worlds grow, their increasing scale can bring about new issues, as we pointed out. The involvement of communities and external stakeholders creates an intricate balancing act, or web of relations that developers must manage effectively.

One of the biggest challenges in virtual world evolution occurs when a virtual world enters its “mass market” phase. This phase is known to world developers by various names, with “crossing the chasm” being the term used by Linden Lab [33]. The issue of reaching the mass market has become a major issue for many virtual worlds (another virtual world start up that we interviewed also discussed problems with reaching the mass market). In the process of addressing the mass market, developers also have to be adaptive, especially as communities react in unanticipated ways, or even become unintentionally creative.

While from the community issues that we identified earlier, it would stand to reason that virtual world developers have much to lose from not engaging with their community, and everything to gain from engaging with them, this is not so much the case. At the same time, virtual worlds have been in search of a revenue model that goes beyond the dominant “advertising as revenue source” model seen throughout the Web. It appears that only Second Life has succeeded in crossing the consumer chasm, at least for its currently slowly growing but no longer exponentially exploding user base [34].

Metaplace actively sought out newer users over this period. One key way was through a consciously easy to use interface that reduced barriers to new users’ entry. They anticipated that an influx of griefers or other problematic behaviors could result from courting more newcomers. They actively engaged their community of users to mitigate the downside of aggressively courting new users. Also, the developers were considering borrowing Second Life’s system of buying back a virtual currency as a way to grow a revenue stream. Tirnua sought a more traditional online game business model for its design. They experimented with charging for premium customers, but saw an immediate drop–off in users. One reason it took this approach was its assumption that its more game–like design would provide value to users that was more like a traditional game.

Linden Lab’s enabling of a wider range of play has led to the emergence of a plurality of communities, each with disparate needs and implications for the revenue base. Scaling up may be more costly since servicing the disparate needs of different market segments will almost certainly cut into the revenues of a company. Linden Lab has recently shifted its organizational focus to attune Second Life more to the mass and serious markets (and perhaps as a result, has deemphasized its more free–form creative communities). Linden Lab itself recognizes that it needs to develop or lose out on this mass market of users, and amongst other things, it is developing a customizable user interface and experience [35].

With new management, Linden Lab also started identifying high–value customers starting in 2008. It began in 2008 by identifying two major categories of “island owners” — corporations and educational institutions. By 2009 it was investing in a few more categories based on those which were emerging more strongly — including the enterprise market, and the events space. All this suggests that a more practical tradeoff was being made between favoring market segments that could potentially yield revenue or high impact, and therefore, market segments in which costs could at least be covered.

All of this was partly facilitated by internal (to the company) transformations. New team members (80 in the past year out of 350) helped in the move towards more user–centric design in Second Life. New programs they instituted included more formal planning processes, the bringing of the entire organization on board (with the new goal), internal focusing processes to help recognize the worth of communities and clients — current and new, and the development of various tools (e.g., use cases, templates and other methods) to handle users entering with those perspectives.

Indeed, the first serious corporate users of Second Life focused on using the platform on their back end. These corporate users tried it for internal purposes such as collaboration or conferencing communications. They walled off private islands, but left open the “door” for their employees to venture out into the “creative unknown” which is the public part of Second Life. Linden Lab tries to promote this: “As we develop our stand–alone, behind–the–firewall Second Life solution, we’re aware of the opportunity it presents for talented and entrepreneurial content creators to reach more customers in a broad inter–connected 3D marketplace.” [36]

Tensions between play and work

The question all of this raises is: under what conditions are various communities’ needs met while the developer advances the strategic cause of growth and sustainability of revenue? Virtual worlds are about creativity and play, but to developers, this creativity and play unlocks unforeseen behaviors that may or may not be desired, or revenue–generating. Revenue also requires developers to focus on growth issues. This also suggests that a ‘positive science’ understanding of the governance of creativity in a community–based environment should also consider incorporating the broader issues of growth and scaling as they relate to community and engagement.

Linden Lab discovered that providing users and communities with the rights to self–determination and creative engagement, and creating a highly immersive platform for content creation and online socializing, did not attract the ever increasing population as Facebook had. Thus, this could not generate revenues in any convincing way. Revenue growth had to be achieved by catering to specific communities’ needs. As an interviewee explained, Linden Lab is now recognizing the complex issue of managing its original community (engaged in role playing based on its high empowerment of user creativity) in relation to corporate and other “high revenue” users. For instance, such corporate users are happy to be associated with a platform where high levels of user creativity are present, yet the same corporate users will be the first to avoid being tarnished by the seedier image that some parts of Second Life are known for. In recognition of this, Linden Lab has sought to minimize the negative impact of adult content in a new policy that, amongst other things, allows the filtering of adult content, and restricts adult activity to specific islands [37]. In the new direction of purposive design, Linden Lab has rolled out this initiative (along with others) to ensure that key stakeholders are kept on board and provide input to the policy change.

In a way, Linden Lab’s recent “value addition” (on the low learning curve side) and “high value community seeking” actions has distanced it from its original lead users and communities. In retrospect, all this is unsurprising due to the many possible factors that affect product adoption. There are some issues which the community was bound to react to, such as the changing of naming conventions to allow a freer selection of names (as opposed to past practices, where last names were “retired” once a certain number of registrants chose them — Linden Lab had expected to upset some community members who may have been proud of their “older” last names.


While a newer platform, Metaplace also recognizes, or has been experiencing, similar issues (and is an example of the final “devastating” one, of failure of its original business model). Nevertheless, the factors causing this did not appear to include ones of design. Issues that appear to be of concern to Metaplace in its second year are of (1) scaling up while not losing the community, and (2) disruptive users (as well as general features including payment systems and monetizing of the content “economy.”) [38] These are all to be expected to some degree, as Second Life faced them. That certain key features have to be encountered ahead of others suggests a rather predictable evolutionary path. In the first, Raph Koster noted in an interview about the further scaling of Metaplace:

“Metaplace has been in a pretty limited beta for quite a while now, and therefore has been populated by early adopters, pretty much folks like many of you here! But of course, over time, as Metaplace keeps growing, it’s going to grow to have a more diverse set of people in it. That is actually part of our explicit goal — virtual worlds for everyone.”

Thus, there is an inherent tension in growth, as it will be harder and harder to satisfy all communities going forward. Koster notes:

“We can’t be all things to all people, of course, and as a UGC [user generated content] platform, it takes a while for every possible use to come to fruition. Now, this doesn’t mean that we value the early adopters less, of course. We absolutely do not want to cause damage to the very unique culture we have developed here. That said, we do need to scale.” [39]

Disruptive users

Another growth issue has to do with dealing with a possible increase in disruptive users as growth ensues. This issue apparently plagued TSO even at its population height. With regards to this, Linden Lab appears to have effectively put in place software measures to deal with the most disruptive of disruptive users’ acts. Socially objectionable behaviors were either banned (early on) or confined to the new adult continent (i.e., area). Similarly, before it closed down, Metaplace was starting to preempt this with discussions with lead users.



7. Conclusions

Many virtual worlds’ developmental paths are starting to exhibit similar features or to garner similar fates, and we are starting to detect what features of designs work or do not work for different types of virtual worlds. We have illustrated how the design of a virtual world helps that world to distinguish itself, both by allowing it to engage users differently as well as by satisfying user interests (i.e., in use creation and personalization). We have shown that the design perspective, namely, developers having a prior design or road map (that they have in mind to evolve the world towards) is only a part of the puzzle that underlies the development of a sustainable and continually developing virtual world. Governance is also part of the puzzle. Developers of user–created virtual worlds also have to recognize and adapt to the emergence of new uses of the world, including creative and innovative uses. They do this by engaging and monitoring users and communities. A user may play a role as a community organizer, or as a lead user facilitating the community’s interaction with developers (perhaps for status reasons). Communities are central to the governance of virtual worlds, but the common understanding of communities leaves much to be discovered about them. Communities all have different interests, as represented by a broad spectrum of types of play and work possible. These can mingle in a virtual world in conflicting ways. Developers have to reconcile these variant communities and keep open the possibilities for creative interpretations of play. As the virtual world evolves, it may be useful to keep the perspective of the communities in mind. Added to these is the third perspective representing the harsh realities of the virtual world business, made perhaps more brutal by the involvement of venture capital.

Developers have to balance amongst all these three perspectives [40]. Said another way, the idea that virtual worlds will evolve (and sustain themselves) entirely due to creative users may be misguided. For instance, the need to serve economic rationalization became so compelling for Second Life that it reduced its emphasis on the creative user community to focus more on business users. In the overall scheme of things, connecting these three perspectives may lead to more appropriate designs, and possibly even optimal ones. These can help to address the issues of reaching critical mass by sustaining the virtual world and its revenues. If this occurs, based on our experience with current designs of virtual worlds, new forms (and even notions) of governance could emerge out of those designs, the communities they secure, and the on–going interactions among developers, users, and communities of users. End of article


About the authors

Ted Tschang is an Associate Professor of Management at the Lee Kong Chian School of Business, Singapore Management University. Ted’s research focuses on creativity and innovation in the creative and high–tech industries. He has studied creativity in the videogames industry and their implications for organizational theory, product development, and industry structure. In recent years, he has been focusing on how user–created virtual worlds are organized and are evolving as social products. He has a Ph.D. in public policy and management from Carnegie Mellon University, as well as graduate degrees in economics and electrical engineering.
E–mail: tedt [at] smu [dot] edu[dot] sg

Jordi Comas is an Assistant Professor at the School of Management, Bucknell University. Jordi received his Ph.D. in 2008 from IESE/Universidad de Navarra. His scholarship combines organization theory, sociology, and social networks. He also enjoys blogging with students (and without), in addition to hanging out in Second Life as JordiSunshine Takacs.
E–mail: jcomas [at] bucknell [dot] edu



The authors thank the Media Development Authority of Singapore and Ravi Sharma of Nanyang Technological University for their financial support of Ted Tschang’s research. The authors also thank their research assistants, including Joshua Nair and LAM Kee Yan of Singapore Management University, for their invaluable assistance.



1. Pearce, 2009, p. 32.

2. For further discussions of definitions of virtual worlds, readers may turn to discussions in Book, 2004; Castronova, 2005 (Introduction, chapter 2); and, Pearce, 2009 (pp. 16–34).

3. Salen and Zimmerman, 2003, p. 168.

4. Castronova, 2005, p. 2.

5. Taylor, 2006a.

6. Lessig, 2006, p. 6.

7. Bartle, 2004, p. 122.

8. The live team handles operations of the world and extensions of the design made by the original development team (i.e., the team developing the original product before it went “live”).

9. Bartle, 2004, p. 123. We should note that Bartle does acknowledge that selected users can be knowledgeable of design and the objective process of design.

10. This point was reinforced by industry participants at a University of Southern California–Nanyang Technological University workshop on business models for digital media, USC, August 2008.

11. Based on conversations with lead users including from organizations using Second Life.

12. Both authors have spent time in–world as participant–observers in Second Life, one author made regular visits to Metaplace, and we have posted at least one research assistant into each world — Second Life, Metaplace and Tirnua — for participant–observation.

13. Much of this table and succeeding tables is constructed from data from the blogs, discussion forums and participant–observations that we have conducted online.

14. Source: Metaplace development roadmap, 18 June 2009, at, accessed 11 November 2009 (no longer online).

15. Bartle, 2004, p. 92.

16. Source: Cyn Linden, “Adult content changes in summary,” at, accessed 6 April 2010.

17. Source: Lowell Cremorne, Interview — Raph Koster, Co–Founder and President, Metaplace, The Metaplace Journal, at, accessed 11 November 2009.

18. Source: “Are you a former/current player of the Sims, Habbo Hotel or YoVille?” at, accessed 11 November 2009 (no longer online).

19. We judge this from the user’s point of view (based on our own participant observations), though as recently as late 2009, Tirnua’s developers started to better address user concerns by making their process a bit more transparent.

20. This observation is based on our participant observation of Tirnua.

21. The nature of adding features during development appears to have been more dramatic in Tirnua’s case (i.e., testing began even as the product was started, and then continued from there onwards). Metaplace appeared to benefit from the lightness of its original design, which makes it easier to gauge and design for outcomes as part of a plan.

22. This depends on whether the overall play (for a certain portion of users) was centrally about content creation (as was the case with Second Life) or not (as was the case with Metaplace, Tirnua).

23. The parallel to this is the degree with which users are allowed to “affect” the environment in current day video games. With increasingly realistic physics being modeled, games’ built environments are increasingly affected (e.g., by being destroyed), which can in turn drastically affect game play.

24. Ibid. (Lowell Cremorne, Interview — Raph Koster, Co–Founder and President).

25. Source: Tirnua Luc, “Custom content is big news in Tirnua,” Tirnua blog, at, accessed 12 November 2009.

26. Content Creator Paypal Integration Discussion Thread, Metaplace Discussion Forum, at, accessed 11 November 2009 (no longer online).

27. Chat Log from Disruptive Users Town Hall Meeting, Metaplace Discussion Forum,, accessed 11 November 2009 (no longer active).

28. This according to an interviewee from Linden Lab. The most current roadmap is the Content Management Roadmap found on the Second Life blog, at accessed 10 November 2009.

29. Chat Log from Disruptive Users Town Hall Meeting, Metaplace Discussion Forum, at, accessed 11 November 2009 (no longer active).

30. Ibid. (Lowell Cremorne, Interview — Raph Koster, Co–Founder and President).

31. Source: Cashing Out and Fraud Prevention Chatlog, at, accessed 12 November 2009 (no longer online).

32. Source: is closing down, Metaplace forums, at, accessed 23 December 2009 (no longer available online).

33. This was also pointed out in an interview we had with the head of Vastpark, another virtual world company from Australia.

34. Through 2008–2009, Second Life grew from 500,000 to 600,000 repeat log–ins per month. This is a metric that captures on–going use as opposed to registrations. Source: “Exclusive: Second Life starts to grow again,” GigaOm, at, accessed: 6 April 2010.

35. These points were made by an interviewee from Linden Lab.

36. “Our content management roadmap,” at, accessed 10 November 2009.

37. “Upcoming changes for adult content,” at, accessed 6 April 2010.

38. Ibid. (Lowell Cremorne, Interview — Raph Koster, Co–Founder and President).

39. Chat Log from Disruptive Users Town Hall Meeting, Metaplace Discussion Forum, at, accessed 11 November 2009 (no longer active).

40. This is subject to two caveats we must note for this study. One is that while we are studying the development practices of these virtual world companies, we are not saying that these are the best practices as predicted by the success or failure of the firm. What seems to be clearer is that these firms had engaged effective (or evolving to effective) practices, as our participant–observations had indicated. In Metaplace’s case, the practices were more than effective, and its failure was connected to the larger business environment. The second is that while we have predicated the study on creativity and communities as a basis for emergent behaviors that developers must adapt to, it is unclear at this stage as to what degree these highly adaptive (to emergence) practices are also the outcome of: (a) the natural condition of an industry early in its lifecycle, and (b) the nature of virtual worlds themselves.



H.E. Aldrich and M. Ruef, 2006. Organizations evolving. Second edition. London: Sage.

J.W. Au, 2008. The making of Second Life: Notes from the new world. New York: Collins.

R. Barbrook, 1998. “The hi–tech gift economy,” First Monday, volume 3, number 12, at, accessed 19 April 2010.

R.A. Bartle, 2006. “Why governments aren’t gods and gods aren’t governments,” First Monday, special issue number 7: Command lines: The emergence of governance in global cyberspace, at, accessed 19 April 2010.

R.A. Bartle, 2004. Designing virtual worlds. Indianapolis, Ind.: New Riders.

R.A. Bartle, 1996. “Hearts, clubs, diamonds, spades: Players who suit MUDs,” Journal of Virtual Environments, at, accessed 19 April 2010.

B. Book, 2004. “Moving beyond the game: Social virtual worlds,” State of Play Conference 2, at, accessed 19 April 2010.

A. Bruns, 2008. Blogs, Wikipedia, Second Life, and beyond: From production to produsage. New York: Peter Lang.

J. Comas and F.T. Tschang, 2010. “Code rules: Conflicting logics of play, collaboration, and utility in the field of virtual worlds,” Working paper edition.

J. Dibbell, 1993. “A rape in cyberspace or how an evil clown, a Haitian trickster spirit, two wizards, and a cast of dozens turned a database into a society,” originally published in the Village Voice (21 December), at, accessed 19 April 2010.

Gartner Group, 2008. “Hype cycle for social software, 2008,” at, accessed 19 April 2010.

L.B. Jeppesen, 2005. “User toolkits for innovation: Consumers support each other,” Journal of Product Innovation Management, volume 22, number 4, pp. 347–362.

R. Koster, 2000. “Declaring the rights of players” (27 August), at, accessed 7 April 2010.

Y.M. Kow and B. Nardi, 2010. “Culture and creativity: World of Warcraft modding in China and the U.S.,” In: W.S. Bainbridge (editor). Online worlds: Convergence of the real and the virtual. New York: Springer–Verlag, pp. 21–41.

L. Lessig, 2006. Code: Version 2.0. Second edition. New York: Basic Books.

R. Levine, 2000. The cluetrain manifesto: The end of business as usual. Cambridge, Mass.: Perseus Books, and at, accessed 19 April 2010.

P. Ludlow and M. Wallace, 2007. The Second Life Herald: The virtual tabloid that witnessed the dawn of the metaverse. Cambridge, Mass.: MIT Press.

T.M. Malaby, 2009. Making virtual worlds: Linden Lab and Second Life. Ithaca, N.Y.: Cornell University Press.

T.M. Malaby, 2006. “Coding control: Governance and contingency in the production of online worlds,” First Monday, special issue number 7: Command lines: The emergence of governance in global cyberspace, at, accessed 19 April 2010.

C. Morningstar and F.R. Farmer, 1990. “The lessons of Lucasfilm’s Habitat,” First International Conference on Cyberspace (Austin, Texas), at, accessed 19 April 2010; also In: M. Benedikt (editor). Cyberspace: First steps. Cambridge, Mass.: MIT Press, 1991, pp. 273–302.

C. Ondrejka, 2007. “Collapsing geography (Second Life, innovation, and the future of national power),” Innovations: Technology, Governance, Globalization, volume 2, mumber 3, pp. 27–54.

C. Ondrejka, 2004a. “Aviators, moguls, fashionistas and barons: Economics and ownership in Second Life,” Gamasutra (23 September), at accessed 21 March 2010.

C. Ondrejka, 2004b. “Escaping the gilded cage: User created content and building the metaverse”, New York Law School Law Review, volume 49, pp. 81–101.

C. Pearce, 2009. Communities of play: Emergent cultures in multiplayer games and virtual worlds. Cambridge, Mass.: MIT Press.

C. Pearce, 2006. “Productive play: Game culture from the bottom up,” Games and Culture, volume 1, number 1, pp. 17–24.

K. Salen and E. Zimmerman, 2003. Rules of play: Game design fundamentals. Cambridge, Mass.: MIT Press.

J. Spence, 2008. “Demographics of virtual worlds,” Journal of Virtual Worlds Research, volume 1, number 2, at, accessed 19 April 2010.

T.L. Taylor, 2006a. “Beyond management: Considering participatory design and governance in player culture,” First Monday, special issue number 7: Command lines: The emergence of governance in global cyberspace, at, accessed 19 April 2010.

T.L. Taylor, 2006b. “Does WoW change everything? How a PvP server, multinational player base, and surveillance mod scene caused me pause,” Games and Culture, volume 1, number 4, pp. 318–337.

T.L. Taylor, 1999. “Life in virtual worlds: Plural existence, multimodalities, and other online research challenges,” American Behavioral Scientist, volume 43, number 3, pp. 436–449.

F.T. Tschang, 2007. “Balancing the tensions between rationalization and creativity in the video games industry,” Organization Science, volume 18, number 6, pp. 989–1,005.

F.T. Tschang, 2005. “Videogames as interactive experiential products and their manner of development,” International Journal of Innovation Management, volume 9, number 1, pp. 103–131.

E. von Hippel, 1986. “Lead users: A source of novel product concepts,” Management Science, volume 32, number 7, pp. 791–805.


Editorial history

Paper received 8 April 2010; accepted 17 April 2010.

Copyright © 2010, First Monday.
Copyright © 2010, F. Ted Tschang and Jordi Comas.

Developing virtual worlds: The interplay of design, communities and rationality
by F. Ted Tschang and Jordi Comas.
First Monday, Volume 15, Number 5 - 3 May 2010

A Great Cities Initiative of the University of Illinois at Chicago University Library.

© First Monday, 1995-2019. ISSN 1396-0466.