1. Introduction
[1.1] By conceiving of video game modding as a form of fan productivity and by focusing on free and open source software (FOSS) modders, this article investigates a phenomenon that is heavily underresearched (Sotamaa 2010a): the meanings that modders attach to their practices and the way that modding influences their relationship with the object of fandom, the video game. In computer games, the word mods (modifications) refers to player-made alterations and additions to preexisting video games. Mod makers, or modders, are typically passionate and skilled players who usually form online teams and cooperate with one another or with game companies to create such mods (Sotamaa 2007; Postigo 2007). Recently, modding practice and modders' creative work have been more commodified by the video game industry. This makes understanding how this form of fan productivity blurs traditional dichotomies between players and game designers, and between play and work (Ducheneaut et al. 2006), more important.
[1.2] Early studies of modding used a conflict model to analyze it, framing modding as, for example, a form of resistance from the bottom against the restrictiveness of corporate game design (Huhtamo 1999; Schleiner 1999), or as a form of exploitation and unwaged work (Terranova 2000; Postigo 2003). More recent studies started framing it in symbiotic terms, stressing corporate attempts to sustain and promote the practice (Sotamaa 2007) as well as the mutual benefits that gamers and game companies can receive from it (Postigo 2007). In our view, two crucial topics common to fan studies and video game studies—the meanings that actors (modders) themselves attach to modding in their everyday activities (Sotamaa 2010a) and how their modding changes their play—are generally ignored in this debate. Here, we investigate the question: How does the practice of modding change modders' relationship with the object of their fandom?
[1.3] Thus, the original contribution of this work to fan studies is evidence that serious modding can radically change the modder's relationship to the game. Modders commonly lose the desire to play the video game, have fewer chances to play it, or even develop a conflicted, love-hate relationship toward modding, an endeavor usually portrayed as both engaging and creative. If these findings are demonstrated to be general, they will be relevant to the study of other important phenomena, such as gamification (Deterding et al. 2011) and playbor (Kücklich 2005). Indeed, this would further suggest that using game-design elements to make nongaming activities (e.g., work or education) more fun or more engaging, and trying to extract labor from playful activities, should be approached more cautiously than is the case now.
[1.4] Play testing is a method of quality control that can take place at many points during the video game design and development process: "It is the design equivalent of bug fixing, though it is considerably less cut and dried. When playtesters look at the game, they try to see if the game is any fun and try to find faults in the game mechanics" (Rouse 2004, 484). In the case of The Battle for Wesnoth, investigated here, the specific activity of play testing video game mods revealed the changes in modders' relationship to the game, and it will be at the center of our arguments below.
[1.5] Our work proceeds as follows: Section 2 defines modding as a form of fans' explicit and expressive productivity. The third section examines the emergence of video game modding and its relationship with FOSS development. Section 4 explains the ethnographic methodology used for the research and describes why The Battle for Wesnoth was chosen as the case to be investigated. The fifth and sixth sections, which are empirical, report modders' statements about how modding changed their relationship to the game, leading them to endure activities that they came to consider anything but fun. The discussion and conclusion sections summarize the most important findings and point toward directions for further inquiry.
2. Video game modding as explicitly expressive productivity
[2.1] Fandoms are now understood to be much more than passive aggregates of passionate fans who merely gather around their favorite TV shows, comics, or video games. Rather, they are now seen as performing active roles in the reconfiguration, appropriation, and expansion of their objects. Theorists such as Jenkins (1992) and Abercrombie and Longhurst (1998) have suggested various ways to frame these more active roles, but Fiske's (1992) tripartite model of semiotic, enunciative, and textual productivity has received the widest attention in fan studies. Fiske's categories serve to classify things that fans produce while consuming their object of fandom; meanings and interpretations of those objects, conversations about them; and written artifacts that expand and are integrated into them.
[2.2] However, analytic applications of Fiske's model to fan productivity in video games have proven to be problematic. First, video games are a participatory medium: they require gamers to play, and playing is widely acknowledged to be both a performative and a productive practice (Wirman 2007). It has been difficult, however, for analysts to distinguish between players' productivity in game and out of game (i.e., where does one locate video game fans' semiotic productivity?). Second, like many current types of user-generated content, modding is manifest mainly online. Therefore, the use of Fiske's model may reinforce an analysis problem identified by Hills (2013). Although, according to Fiske, any fan productive action may well span all three categories, contemporary inquiries into online participatory culture generally use them as if they were mutually exclusive. Further, nearly every online interaction is associated with a (digital) text. Perhaps for this reason, semiotic and enunciative productivity have been neglected, while textual productivity has become the all-embracing label for all types of fan productivity (Hills 2013). For instance, Crawford (2012, 120) labels many aspects of gamers' production as fan textual productivity, including "the production of websites, mods and hacks, private servers, game guides, walkthroughs and FAQs, fan fiction and forms of fan art." However, these three types of productive output require very different skills and commitments on the fans' side. Analysis of fan productivity must demonstrate the specific ways in which each type has different implications for the fan-game relationship.
[2.3] To avoid flattening video game fans' productivity under one all-embracing label, it is useful to define such productivity along two axes. The first distinguishes implicit from explicit productivity. This dimension reflects how most participatory cultures (Jenkins 2006) demand of fans a relatively high degree of commitment, awareness, and mobilization of skills. Consequently, producers often explicitly claim their productions, in a way that demonstrates their self-consciousness. In contrast, in many popular Web 2.0 applications and Web sites, users' participatory productivity is generally only implicit, and is rarely demonstrably self-conscious (Schafer 2011). The second axis distinguishes between instrumental and expressive productivity. The first extreme reflects artifacts that gamers produce to advance more effectively in the game (e.g., documentation, walkthroughs, or tools). The second relates to game-specific outputs that extend, enrich, and reconfigure the video game (e.g., customized player models, skins, game levels) (Wirman 2007).
[2.4] Because modders are video game fans who consciously commit their time and mobilize their skills to create digital artifacts that enrich their object of fandom, we suggest that video game modding be analyzed as a form of explicit, expressive fan productivity. Broadly speaking, a shared feeling of belonging and affiliation with the object of fandom, and enjoyment of the expressive affordances that this object provides, are at the basis of game fans' engagement practices (Jenkins 2006). More specifically, Sotamaa (2010a) suggests that playing, hacking, researching, artistic expression, and cooperation are key motivations for modders. Modders play because they desire to extract as much as possible from the game. They see modding as a manifestation of a hacker legacy. They research because they need to investigate not only the details of the source code, but also the background and contextual information that impinge on and even define the mod. The game is a medium of expression, and modding can be aesthetically (and also politically) motivated. Finally, for some people, modding is a primary way to cooperate and connect with like-minded people.
[2.5] It is safe to say that modifying games through modding has been a recognized part of game industry practice and mainstream gaming cultures at least since the early 1980s. Castle Smurfestein (1983), a total conversion of the video game Castle Wolfenstein (1981), is usually considered the first mod ever created (Kücklich 2005). At that time, video games were relatively unsophisticated software artifacts with simple design and mechanics. They were mainly a niche product, within the reach of only technically skilled computer geeks and hackers. However, even in these early days, a few software companies provided development tools or construction kits to help players customize game characters and levels. An example is The Shoot 'Em Up Construction Kit (SEUCK) available for Commodore 64 video games (Sihvonen 2009).
[2.6] The modding phenomenon received a mid-1990s boost with the arrival of first-person shooters mods, initiated by the public release of the source code for Id Software's Doom. This release was so successful in fostering an emerging modding community that Id Software released a Doom Editor Utility to help gamers produce levels, and did the same for its subsequent successful games, Quake and Quake II. Several mods emerging from these first-person shooters (e.g., Half-Life and Urban Terror) brought new life to the original video games while providing career opportunities to the authors of successful mods. This was particularly visible in the success of Counter-Strike (1999), which was originally circulated as a mod of Half-Life (which, in turn, was a mod of Quake II). Later, as a result of its unexpected success, Counter-Strike was reproduced and sold as a stand-alone video game (Kücklich 2005).
[2.7] Nowadays, video game modding has been consolidated into a game resource that is highly appreciated by gamers and game companies. Video game mods feature several levels of design and implementation complexity and integrate with the video game at different software levels. They can affect the physics of the game's virtual world, modify play, and introduce new story lines and game types. Typically, mods introduce new characters, maps, or levels into the game (Postigo 2007), but they can also provide improved interface components or address shortcomings in game design (Nardi 2010). Furthermore, because modders are not constrained by budget plans and market revenues and are therefore much more free to take creative and innovative risks (Postigo 2007), attempts by the game industry to take advantage of modding should not surprise us. These attempts include nurturing modder communities through sponsored events, such as modding contests (Sotamaa 2007), and integrating modding directly into the design of games, making it a concrete part of game play (note 1).
3. Video game modding and FOSS
[3.1] Contemporary modding clearly converges and partially overlaps (e.g., Doom) with FOSS, either as cultural practice or as a form of software engineering. First, as highlighted by Sihvonen (2009) and Sotamaa (2010a), video game modding displays a hacker legacy of tinkering and engaging deeply with technology production for the sake of self-expression, learning, and harnessing. This legacy is also at the basis of much FOSSing (i.e., tinkering and engaging with the practices of FOSS development and culture) (Coleman 2012). Second, FOSS has always manifested a highly modular and customizable approach to software design. Recently, modding and software extensions have emerged as effective software development methods in FOSS (Scacchi 2011). Third, although FOSS video games occupy a small niche compared to the domination of the market by corporate (proprietary) video games, they represent an increasingly relevant portion of FOSS products. For instance, in one of the most used FOSS development portals, SourceForge.net, "Games" has recently come to constitute the second biggest category of FOSS projects. It is therefore reasonable to contend both that the creation of FOSS games and of related game development tools is a central element in the cultural world of computer games, and that gaming is central to FOSS (Scacchi 2011).
[3.2] As acknowledged by Scacchi (2004), FOSS game development includes all the fundamentals of the FOSS paradigm; that is, it is a transparent, iterative, collaborative, and Internet-mediated development model (Weber 2004). Thanks to specific software licenses, the availability of the software source code, and an infrastructure of shared tools, FOSS programs can be and are developed by communities of volunteer developers and skilled users who commit their skills and efforts to the betterment of the software, "to scratch their own itches" (Raymond 1999, 23) or "just for fun" (Torvalds and Diamond 2001). Two aspects set modding in FOSS apart from modding in typical corporate, mainstream video games. First, FOSS permissive licenses and the availability of the source code allow modders to work on their mods without having to reverse engineer code or enter a legal gray area. Second, because FOSS modding is open and collaborative, it takes place in a space where participatory practices (e.g., reporting and fixing bugs, and providing translations and feedback) are welcome, encouraged, and supported (Scacchi 2004). Indeed, FOSS end-users are allowed, expected, and even urged to take part in projects, and they are often provided with the means (e.g., documentation, tutorials, mentors, tools) to do so. After some time making only marginal contributions, users often find themselves involved in more central development activities (Ducheneaut 2005).
[3.3] Certainly, FOSS core developers maintain more power and control over the source code than do co-developers or participant end users, but development happens in less formal, less clear-cut, and more transparent ways than in the case of proprietary, closed source software. In this latter case the boundaries between companies' development efforts and the gamer community's enthusiasm are much clearer and more impermeable (Nardi 2010). Ultimately, because participating in FOSS means socialization, learning, and career development, we suggest that modding in FOSS can be considered an institutionalized practice for fan involvement.
4. An ethnographic approach to modding in The Battle for Wesnoth
[4.1] But what are the patterns typical of fan modding involvement in FOSS games? To investigate these patterns as they are experienced and performed by modders in daily activities, and to go beneath a surface reading of the digital traces left by their interactions, we developed an ethnographic study (Hammersley and Atkinson 2007) of a FOSS video game project: The Battle for Wesnoth. As our primary means of gathering data, we did 14 months (December 2010 to April 2012) of participant observation, including Internet-mediated interviews as the primary methodology for gathering data (Garcia et al. 2009; Kivits 2005).
[4.2] As the existing literature stresses, the online traces generated by social formation—project Web sites, instant messages, e-mails, and forum messages—supply only a partial picture of FOSS modding. For this reason, our empirical research required deeper and longer engagement than unobtrusive or automated data gathering techniques would have (Demazière, Horn, and Zune 2011; Hakken 1999). Of course, the benefits of focusing on a single case and using these methodologies come at the cost of results that are easily generalized.
[4.3] The empirical research was conducted overtly. We announced our presence and intentions through a few public messages. No complaints were raised in response. The fieldwork involved long periods of observation, participation in the collective infrastructure (e.g., reading forum discussions about the creation of add-ons and discussing, via IRC or e-mail, improvements to add-ons and The Battle for Wesnoth), game play (e.g., multiplayer and competitive matches), and off-line interactions (e.g., a couple of face-to-face interviews and attendance at a FOSS conference).
[4.4] We conducted 29 Internet-mediated interviews (note 2). To put informants at ease and to create an egalitarian research setting that would promote trust, interviewees were able to choose their preferred interview medium. The majority opted for asynchronous communication media (either e-mail or the private messaging system built into the Battle for Wesnoth forum), with five choosing a synchronous medium (IRC). As already documented in the literature, asynchrony promotes a more intimate and reflexive process of knowledge construction than does synchrony, but it has the drawback of lengthier and more cumbersome management work (Kivits 2005; Wirman 2012). All interviews began after informants had read, understood, and agreed to explicit procedures regarding interview management, data handling, privacy, and confidentiality. Informants are referred to by aliases.
[4.5] The Battle for Wesnoth is an original "turn-based tactical strategy game with a fantasy theme" (http://www.wesnoth.org/) inspired by the classic turn-based games Warsong (Nippon Computer System 1991) and Master of Monsters (System Soft 1991). The strategic dimension of the game involves fighting on a favorable terrain, at a favorable time of day, and matching one's forces against weaker or disadvantaged enemies. In game, players build up armies by gathering resources and recruiting new units and then use them to meet the objectives set for campaign scenarios or to challenge other players over the Internet. Two game modes are available: single player and multiplayer. The former provides 16 official campaigns, each one with its specific plot, characters, and number of connected scenarios. The latter provides 54 multiplayer scenarios with different game forms (1 vs. 1, 2 vs. 2, 4 vs. 4, free for all, collaborative). Video 1 is a promotional trailer for the video game that shows different combat situations and the basic game layout.
[4.6] Launched in 2003 as a prototype, The Battle for Wesnoth slowly gained success in the niche area of strategy games and reached version 1.10.5 in November 2012. Currently, only a couple dozen developers are actively working on the project, but 60 have contributed to it over the years. About 200 people are officially credited with making helpful contributions. The official Internet forum, which is the primary support channel for gamers, modders, and project participants, has 24,500 registered members, with an average continuous presence of 35 visitors.
Video 1. A presentation trailer for The Battle for Wesnoth (v.1.8). The trailer provides a basic introduction to the game play and game content availability while showing different battle situations.
[4.7] The possibility of creating add-ons for The Battle for Wesnoth is widely advertised and encouraged through the project infrastructure. The Internet forum is used primarily to support modders' work, and vast documentation for add-on creation exists, in the forms of technical guides, how-tos, and coded examples. Technically, all add-ons are coded in a special markup language, Wesnoth Markup Language (WML), which is developed and maintained by Battle for Wesnoth developers. The add-on distribution and installation system is highly integrated into the video game, and the starting interface allows players to connect to servers able to download existing add-ons or upload new ones. Figure 1 shows the interface for downloading add-ons.
Figure 1. Interface for downloading add-ons. Available content is sorted by number of downloads. The interface shows each add-on's name, version, authors, number of downloads, and file size. [View larger image.]
[4.8] Currently, the add-on servers for the two most recent versions of The Battle for Wesnoth include 400 and 100 add-ons respectively (note 3). The most common add-ons are the mods of single-player campaigns, multiplayer scenarios, and factions (i.e., collections of battle units). Less common ones concern map packs, soundtracks, and authoring tools. This additional content represents a valuable and highly appreciated Battle for Wesnoth gamer resource: it extends the longevity of the video game and expands strategic gameplay. The two most recent Battle for Wesnoth versions introduced only one new official campaign, although developers had formerly provided gamers with more content elements. These add-ons provide players not only with more content to play but also, more interestingly, with new strategy-related mechanics. For instance, add-ons for nonlinear campaigns exist, in which the results of the battles directly influence the story lines that may unfold in the campaign. Nontraditional scenarios also exist, in which the composition of the map terrains and of the customized unit factions drive players to radically change their playing strategies. Finally, a few highly appreciated add-ons emerged recently that blend the pure turn-based strategy design elements of The Battle for Wesnoth with elements oriented more toward role-playing games (RPGs).
5. Passionate gamers converge toward modding; busy modders diverge from gaming
[5.1] When we discussed how they had moved from being gamers to becoming involved in Battle for Wesnoth development activities, particularly creating add-ons, all informants acknowledged that they were initially spurred on by passion for the game, either in general or for specific aspects of its design. They all expressed enthusiasm for The Battle for Wesnoth. Modders were more than just players; they started as fervent supporters of the game.
[5.2] I was first exposed to Wesnoth about 4 years ago, when a friend of mine told me about this awesome fantasy game he had downloaded. When I got home, I downloaded it for myself. I played the tutorial, and found myself completely hooked. I went on to complete Heir to the Throne and Sceptre of Fire. I have been playing/involved with Wesnoth more or less continuously ever since then. (Krellis interview, May 31, 2011)
[5.3] I was instantly drawn to the small download size of the game, and (of course!) the fact that it was free. I downloaded it and was instantly sucked into simple mechanics and game play that led to a wealth of strategies and replay value. Today I am still in awe of not only how Wesnoth manages to produce such fun and at times complex game play and strategies using very simple mechanics and rules, but also how moddable Wesnoth is and how sophisticated the engine has become. To be honest, I really find it hard to say anything negative about Wesnoth, simply because there is very little negative about the game, in my experience. (Cylanna interview, June 1, 2011)
[5.4] Furthermore, as shown in the two following excerpts, add-on creators confirmed that they find modding fun:
[5.5] I would say that the point where I moved from being a player to a member of the community was when I joined the forums in hopes of getting help in writing WML and creating a campaign…There I found extensive, user-friendly documentation, and a forum full of helpful, friendly people. The forum community was extremely welcoming, and gave me a good first impression. It was then that the solitary learning process ended and I began "really" learning and having fun with WML. (Krellis interview, May 31, 2011)
[5.6] I just wanted to make a role playing game [add-on]; one-scenario quests aren't interesting enough, character development needs to be infinite (still haven't figured out how best to handle that ;-) ) and actions should have consequences…For me, part of the fun is that the game engine is not meant for this. ;-) I like that I'm twisting something to a completely new purpose…So far, I haven't found anything that I absolutely needed that I was not able to do with WML and LuaWML. As far as I can remember. There's so much that I want to do and *can* do that I try not to spend time worrying about the things that I *can't* do. (Erlornas interview, June 29, 2011)
[5.7] These two add-on authors' claims are very similar to those of others. Indeed, all informants said that they enjoyed modding for one or more of the following reasons: (1) WML is flexible, allowing them to experiment with innovative add-ons, (2) the infrastructure supports modding, offering documentation and tools for authoring and collaboration, and (3) the environment is supportive and welcoming, with a community of friendly and experienced peers who can be relied on for help.
[5.8] However, when modders were asked about their current relationship with the game, a slightly different picture emerged: modders were spending more time creating add-ons for the game than on playing it and, when they did play, they did so for testing purposes.
[5.9] You can create your own add-ons such as maps, units and even entire campaigns. Before I started making my own content, I had been studying the game, its mechanisms and potential. I used to play regularly…Now I play this game occasionally. Maybe this is because I prefer creating to playing. (Kalenz interview, May 14, 2011)
[5.10] I'm one of those people that's more interested in creating than playing, so it doesn't really bother me if I spend more time working on this campaign than I do playing other campaigns. Granted, it does sometimes feel like I made too much work for myself, but I still enjoy actually working on the project. (Owaec interview, August 19, 2011)
[5.11] The amount of time I spend actually playing the game (aside from play testing my own campaign) has gone down considerably since I started contributing. (Elurin interview, June 30, 2011)
[5.12] Over time, the incompatibility of playing and modding emerged. This is not surprising, as both activities are time-consuming. With limited time to spend, modders make decisions about which activity to favor, choosing the one they describe as more fun. However, as shown in the next section, this "fun" is problematic, because modding also includes activities that are considered anything but fun, play testing above all.
6. "Play testing is hard work"
[6.1] In most video games, game content must be balanced in order for players to enjoy an intriguing experience. Because the strategic dimension of The Battle for Wesnoth's design is largely regulated by the relationship among the variables defined for each content type or element (e.g., the sizes of maps, the composition of different terrains, the characteristics of factions and units), it is important for each element of playable content (both official ones and those provided by add-ons) to integrate smoothly with the others. Players, modders, and developers all stress balance as one of the key points of the game. In general, they praise, support, and call for efforts to improve balance. And play testing is fundamental to achieving it. It is also hard work, as most modders point out.
[6.2] Play testing is a relatively collaborative activity. It is usually differentiated into self-testing, play testing with confidants, play testing with strangers, and play testing with the target audience (Fullerton, Sawain, and Hoffman 2004). However, as this section shows, play testing The Battle for Wesnoth is more complex. Indeed, although Battle for Wesnoth modders are in theory able to play test directly with their target audience, and although such testing is critical for the game mods to be broadly accepted, they still often find themselves play testing with very few confidants or, more likely, alone. Practically, play testing an add-on involves the repetitive tasks of (1) playing the add-on to discover how well it balances with other game elements, as well as how enjoyable it is, (2) tweaking it to improve balance and the playing experience, (3) playing it again to check the effect of the changes, and (4) repeating the sequence again when necessary.
[6.3] Play testing may reveal that map resources have been misplaced, allowing one player to go ahead to the detriment of another. Play testing a faction add-on may make clear that the faction has an unfavorable ratio of cost to skills that will make it unappealing. Modders tweak mods to improve these aspects, but they consider this process boring and repetitive, the least attractive activity in the whole process, the one that can lead them to abandon add-on projects.
[6.4] I agree with those others that "developing" [add-ons] doesn't leave much time for "playing," and that play testing is very repetitive and not much fun. It's not really an exaggeration to say I haven't "played" Wesnoth in years. (Arne interview, May 16, 2011)
[6.5] I'm surprised that people would contribute to the game when they were not enjoying it. I certainly wouldn't! I have never felt *obligated* to contribute. (I suppose I could feel obligated to fix a bug, but I haven't yet.). I have gotten pretty bored with play testing before, and when I did, I did something else for quite a while. (Nym interview, February 3, 2012)
[6.6] Largely, play testing produces only approximate results. As there is no certainty of achieving balance (and thus of being effective), play testing cannot be done well individually; it requires collaboration.
[6.7] In The Battle for Wesnoth, play testing is a trial and error process in pursuit of an undefined goal: balance. It is not the same as debugging, as most bugs are evident, and fixes either work or they don't. During a public conference, a leading Battle for Wesnoth developer described balance as undefined, a subjective goal pursued repeatedly:
[6.8] Balance is mainly a feeling in your community. People say a game is not balanced when it's "not fair," when it's not challenging, but if it's too easy [it] is not balanced either, probably even when [it] is not fun…Balance is a moving target. We have been doing this for years and we are still doing it, because whenever we change something it has also a ripple effect throughout the game. So your game is never balanced, your game is about [i.e., approximately] balanced…It's not a science, you have to try stuff: if you see something doesn't work in release X, then you change it and you see what happens in release X+1. Did it work? Did it make things worse? Did it have side effects you didn't want? You can try it. (Deoran, talk given at FOSS conference, February 5, 2012)
[6.9] Resources can be reduced when they are too numerous or increased when they are too few. Similarly, an unbalanced cost/skills ratio among faction units can be adjusted by changing a unit's cost, or redefining another unit's skills. Nonetheless, changes create a new equilibrium that is not necessarily better balanced. A map overflowing with resources may upset strategy play more than a map on which misplacement of resources puts one player at a clear advantage over others. Therefore, modders play test an add-on to identify balance problems, change some variables, play test it again, change some variables again, and so on, until the add-on seems to be relatively well-balanced. Modders and other players are tightly bonded by their shared perceptions of how well mods are balanced. Indeed, as Deoran claims, "balance is a feeling dispersed in the gamers' community." There is no algorithm to evaluate or achieve balance; gamers and modders collectively construct it.
[6.10] The Battle for Wesnoth modders have good practices for designing and developing balanced add-ons and thus lessening the need for play testing, but these are not enough to avoid it completely. They use the design and the code of existing game content to guide that of their own add-ons. They often examine official game elements and well-considered add-ons and keep them in mind when creating their own, and documentation offers tips and tricks for creating balanced content. Furthermore, game elements are only balanced in relation to official and default specifications, so modders can ignore balance concerns among add-ons and don't have to play test them. However, although modders keep balance in mind throughout the process of designing and implementing add-ons, play testing remains a necessary step in their creation and enhancement. This is because players are the ones who collectively decide whether add-ons are balanced or not. This puts modders in a catch-22:
[6.11] The main obstacles we faced developing [the add-on] was getting art and getting play-testers. I did a lot of the art myself, but I couldn't do the play testing myself, and though the Era had tens of thousands of downloads on the add-on server, we rarely got feedback on balancing, so it was difficult to improve anything or get the balancing up to mainline quality—which was pretty frustrating, since the main reason the "serious" MP [multiplayer] players didn't try out [the add-on] was that they considered it poorly balanced, and thus not as fun to play. (Cleodil interview, June 13, 2011)
[6.12] Involving serious players in the play testing phase and integrating their feedback has become fundamental to The Battle for Wesnoth modding. The most established procedure by which modders gather feedback is to open forum threads calling for it. This is a standard practice, usually covering the whole add-on development process. Modders often use general-purpose threads with titles such as "Feedback thread for [add-on X]," where players can provide bug reports, positive and negative comments, and suggestions for further improvements. However, for the play testing phase, modders start specific threads (typically named "Balancing [add-on X]") that are used exclusively for balance issues (note 4). Using forum threads to gather feedback not only requires players to find the add-on among the many others and decide to play it, but also, having played it, to report balance issues. As the quotation above suggests, finding play testers is not easy: good players want good content to play with and in general avoid unbalanced add-ons, while new, inexperienced players cannot provide reliable feedback.
[6.13] Modders seem to start play testing when their add-ons become playable and attract a few first players. Answering a question about the precise moment one can talk about balance, Deoran replied,
[6.14] I would say: at the point where your game start[s] to be really playable. It doesn't have to be good, but it should not be a prototype. Once you start having players who come and play the game, at this point you should start thinking about balance. (Deoran, talk given at FOSS conference, February 5, 2012)
[6.15] However, although a large user base makes attracting play testers for the official content easy, it is more difficult to find them for add-ons. Add-ons must stand out from hundreds of others to gain renown and be consistently played by gamers. As a consequence, modders often find themselves doing something alone that should be done collaboratively. At the same time, they may receive occasional feedback from random play testers that calls into question some already-completed aspect of play testing.
7. Discussion
[7.1] The arguments presented above suggest that where fans' expressive productivity (modding) is institutionalized (as it can be in FOSS development), what was once experienced as mostly fun can get loaded with expectations (e.g., pursuing balance), responsibilities (e.g., reacting to gamers' feedback), and dependencies (e.g., relying on play testers). For these reasons, the practice of modding becomes a more serious form of work. Far from being cheerful hobbyists, modders become more similar to small-scale game designers. However, unlike such designers, modders remain committed to the idea that modding should be fun.
[7.2] Describing the tenets of game design and development, Rouse (2004) depicts play testing as
[7.3] one of the most exhilarating parts of the game development cycle. It is then that you take the project you have been working on for months or years, during which time only the development team has played the game, and show it to people outside the team…Playtesting is not just a minor stepping-stone to getting the game shipped to the duplicators or uploaded to the Internet. Instead playtesting is a key time during which you can transform your game from average to excellent, from something that shows promise to a game that is truly great. Few games ever came out of the developer's hands in absolutely perfect shape. Ideally, it is the playtesting cycle that gives your game the extra push to be the best it possibly can. (483–84)
[7.4] Our fieldwork showed that, although it is crucial, play testing is anything but exhilarating. Its repetitiveness, together with the difficulty of gauging its effectiveness, make this part of the work quite different from the initial designing or the actual coding of the add-ons, which modders still find pleasant and creative. For some modders, play testing is so boring that they may stop developing the add-on altogether, even though the broader effort is usually fun. For others, it spoils the very activity of playing: even though it involves playing the game, play testing is a completely different way of experiencing play.
[7.5] For gamers who became modders, the video game is no longer an artifact to be played. Rather, it is a platform with participatory affordances. However, even though they enjoy most of the creativity involved in modding, they also face more challenging and less fun phases. In this light, the sociotechnical context in which gaming take place emerges as crucial. This means that authoring and collaborative tools, support channels, and extensive documentation and tutorials, together with a supportive and friendly environment, are necessary to reinforce and support the enjoyment of modding, so that the tedious bits can be better endured.
[7.6] Since I joined the developers I don't play much anymore. Testing the stuff I am responsible for is just enough and I don't even enjoy it anymore. Play testing is hard work. (Drogan interview, May 12, 2011)
8. Conclusions
[8.1] Contemporary participatory cultures increase the significance of fan production. If our conclusions here can be shown to be generally applicable, they will underline the value of attending to FOSS modding. By conceiving video game modding as a form of fans' explicit and expressive productivity, and by focusing on the way FOSS modders experience and reflect upon their work in practice, we suggest the following critical conclusions: (1) modding can seriously alter fans' traditional way of relating to their object of fandom, (2) even though modding involves designing elements of the game, these design activities are quite different from how they are formally described in research literatures, and (3) modding is a complex endeavor that includes several phases. Although some of these phases promote enjoyment and creativity, others do not and can even trigger estrangement.
[8.2] Altogether, these results may bear important implications for phenomena other than fandom that are similarly transformed by contemporary participatory culture, such as gamification, playbor, and professional gaming. Rooted in game practices and culture, these phenomena are reconfigured in ways that the traditional separation between play and work no longer captures. These refigurings call into question, for example, the extent to which making nongaming activities like work or education more engaging by using game-design elements, or extracting workforce and labor value from playful activities by designing them in a work-like fashion, may be efficient and sustainable over the medium and long term.
[8.3] In implicit or instrumental productivity, fans' value expresses itself either unconsciously or as the result of the overcoming of concrete needs. In contrast, explicit and expressive productivity implies a higher degree of awareness and expectation on the fans' side. Indeed, the case of modding in FOSS that we examined here shows how explicit and expressive forms of productivity can benefit the object of fandom by enriching and extending it. However, we also showed that the more fans' productive output is valued and integrated into the object, the more fans' efforts are accompanied by expectations and responsibilities toward the collectivity. In this case, their efforts may cease to be creative and fun.
[8.4] Finally, we note a few directions for further inquiry. First, sociotechnical contexts that sustain fans' productivity but also may lead to their detachment should be better investigated. Understanding this would illuminate what makes fans' involvement either endure or wane over time. Secondly, we need to know whether (and, if so, to what extent) implicit and instrumental forms of fan productivity also promote change in the relationship between fans and their object of fandom. Finally, we need to study modding in corporate, proprietary video games, where fans' productive practices are more marginal to the creation of the objects of fandom, and compare those studies to this one. Such a comparison may provide valuable insight into divergent participatory trajectories.
9. Acknowledgments
[9.1] We thank Stefano De Paoli, Lorenzo Polidori, and Diego Fornarelli for their advice and comments.