Actas da Videojogos2013 Conferência de Ciências e Artes dos Videojogos Arte em Jogo

Transcrição

1 TR 2014/04 ISSN X Actas da Videojogos2013 Conferência de Ciências e Artes dos Videojogos Arte em Jogo de Setembro, 2013, Dep. Eng. Informática, Universidade de Coimbra Coimbra, Portugal Licínio Gomes Roque (Universidade de Coimbra) Editores Ana Paula Varela Afonso Luís Lucas Pereira (Universidade de Coimbra) (Universidade de Coimbra) Rui Craveirinha (Universidade de Coimbra) 1

2 TR 2014/04 ISSN X CENTRO DE INFORMÁTICA E SISTEMAS DA UNIVERSIDADE DE COIMBRA PREFÁCIO Este volume apresenta os artigos apresentados na Videojogos2013 Conferência de Ciências e Artes dos Videojogos, que decorreu na Universidade de Coimbra, a 26 e 27 de Setembro de 2013, sob o tema Arte em Jogo, organizada pela Universidade de Coimbra em cooperação com a Sociedade Portuguesa de Ciências dos Videojogos. As conferências da Sociedade Portuguesa de Ciências dos Videojogos realizam-se anualmente e são um encontro para promover a cultura científica, a investigação e a indústria de videojogos em Portugal. O tema Arte em Jogo foi propositadamente estabelecido como um double-entendre, simbolizando simultaneamente a arte que existe nos jogos, e a ideia de que é a arte que está em jogo quando criamos, estudamos e experienciamos os videojogos como novo meio artístico. Serviu dessa forma de mote para a busca de novas perspectivas e reflexões multidisciplinares sobre o meio, as suas complexidades, ambiguidades e contradições, agregando contribuições das comunidades académica, científica e artística de Portugal. 2

3 TR 2014/04 ISSN X Para aprofundar o tema, foram convidados autores de renome de quem são conhecidas reflexões neste campo. Michäel Samyn e Auriea Harvey (Tale of Tales studio), autores de obras como FATALE ou Bientôt l été, e promotores do manifesto RealTime Art e do movimento notgames, onde analisaram de forma crítica a dialética jogo-arte. Dan Pinchbeck (do TheChineseRoom studio), co-criador de Dear Esther e Amnesia: A Machine for Pigs, videojogos que personificam o âmago do debate entre ludologia e narratologia, e que assistiram na divulgação de um discurso crítico sobre os videojogos e os não-jogos. E como representante da comunidade académica, Miguel Sicart (da IT University of Copenhagen), cujo seminal trabalho sobre a ética nos videojogos e o desenvolvimento de uma retórica anti-jogo e pró-jogar, o colocam na vanguarda da investigação da área. Universidade de Coimbra (Comunicação e Imagem); e Turismo Centro de Portugal (Pólo de Marca Turística Coimbra). Por último, um agradecimento especial aos oradores convidados, ao nosso convidado especial Bruno de Figueiredo, aos autores, aos revisores, aos membros da comissão científica, aos voluntários que ajudaram na organização e aos parceiros da organização local, que com o seu tempo, esforço e dedicação contribuiram para o sucesso desta conferência. A organização deixa aqui o seu agradecimento a todos aqueles que contribuiram para a realização desta VJ Uma palavra de agradecimento à UC, e à SPCV e aos patrocinadores da conferência: SPF TRANSCREATIVA; SUDOE Interreg IV B, UE/EU-FEDER/ERDF; Reitoria da Licínio Gomes Roque Ana Paula Varela Afonso Luís Lucas Pereira Rui Craveirinha 3

4 TR 2014/04 ISSN X ORGANIZAÇÃO: Ana Paula Afonso Ana Almeida Tiago Agostinho João Amsellem Nuno Barreto Hélio Cavalcante Rui Craveirinha Filipe Penicheiro Luís Lucas Pereira Durval Pires Licínio Roque COMISSÃO CIENTÍFICA: Ana Almeida Lynn Alves Valter Alves Bárbara Barroso Sergi Bermudez I Badia Ana Amélia Carvalho Nuno Castelhano Monchu Chen Esteban Clua António Coelho Nuno Correia Rui Craveirinha Cátia Ferreira Mauro Figueiredo Ernesto Filgueiras Maria João Gomes Patrícia Gouveia Helio Henriques Rui Lopes Filipe Luz Carlos Martinho Óscar Mealha Leonel Morgado Filipe Penicheiro Luís Lucas Pereira Luís Pereira Rui Prada Paulo Quaresma António Ramires Teresa Romão Licínio Roque Luís Santos Pedro A. Santos Danielle Silva Frutuoso Silva A. Augusto Sousa Beatriz Sousa-Santos Roger Tavares Patrícia Tedesco Luis Filipe B. Teixeira Ana Torres Mário Vairinhos Ana Veloso Nelson Zagalo 4

5 TR 2014/04 ISSN X PATROCÍNIOS: 5

6 TR 2014/04 ISSN X ÍNDICE Página I. Keynotes (Abstracts) Beauty will save the world 10 Michael Samyn & Auriea Harvey Emotion is Gameplay, Gameplay is Emotion 11 Dan Pinchbeck Overtures to the World: Videogames from Pataphysics to Participatory Art 12 Miguel Sicart II. Full Papers Architecting a Reusable Platform for Pervasive Augmented Reality Games based on Petri Net Models 14 Tiago Agostinho, Fernando Milagaia, Fausto de Carvalho and Licinio Roque An Interaction Design Model for a Petri Net based Behavior Editor 24 Nuno Barreto and Licinio Roque EEG Biofeedback and Brain Computer Interface in Games 35 João Rodrigues, Wenya Nan and Agostinho Rosa The Video Game and Player in a Gameplay Experience Model Proposal 45 Samuel Almeida, Ana Veloso, Licinio Roque and Óscar Mealha Uma Proposta de Arquitectura para Composição Dinâmica de Soundscapes em Videojogos 55 Durval Pires, Valter Alves and Licinio Roque 6

7 TR 2014/04 ISSN X Crowdsourcing y Videojuegos. Los Jugadores Como Productores De Conocimiento 65 Israel V. Márquez Our Ludic Sins A Survey on Belief in Videogame Axioms 76 Rui Craveirinha and Licinio Roque Hábitos de jogo em dispositivos móveis pelos jovens portugueses 85 Ana Amélia Carvalho and Inês Araújo Online GYM: um ginásio virtual 3D integrando a Kinect 95 Fernando Cassola, Leonel Morgado, Hugo Paredes, Benjamim Fonseca, Paulo Martins and Fausto de Carvalho Atividade física, obesidade e videojogos ativos na Escola: estudo de hábitos e práticas de jogos em jovens do ensino básico e secundário Nuno Palma and José Ramos 104 III. Work-in-Progress O Espaço da Criação Independente na Indústria dos Videojogos 121 João Sousa and Carlos Caires Os jogos mobile que os alunos mais jogam no 3º ciclo 127 Cândida Barros and Ana Carvalho Exploração de novas soluções de interação com a instalação IMP.cubed 137 Maria João Pinheiro, Ana Veloso and Óscar Mealha Gamágio: A Game For Plagiarism Prevention 144 Andreia Ferreira and Frutuoso Silva Amnesia The Dark Descent: Experienciar o survival horror sem formas de defesa e a custo da sanidade 150 Ana Narciso 7

8 TR 2014/04 ISSN X Criar um Serious Game Sobre Bullying Escolar 157 Diana Rodrigues, Pedro Neves, Ricardo G. Barroso and Leonel Morgado Transcoding Action: A perspective on the articulation between the player s and system s actions in video games 163 Pedro Cardoso and Miguel Carvalhais E-sports: Um fenómeno da cultura digital contemporânea 170 Pedro Saraiva Playable and Collaborative Art: the MonMazes case-study 177 Ana Carina Figueiredo, Marco Heleno, Nelson Zagalo and Pedro Branco 8

9 TR 2014/04 ISSN X I. Keynotes (Abstracts) 9

10 TR 2014/04 ISSN X Beauty will save the world The 20th century was not friendly towards beauty. Beauty was pushed aside by a disinterested appreciation of the sublime and a fascination with irony. Beauty was enlisted to seduce the public into over-consumption. Beauty was used as an insult for the weak and was considered irrelevant to the strong. Beauty was a lie. Beauty was accused of injustice, of being a tool for evil. Beauty was burned at the stake of humanism and democracy. All this time, beauty remained. Dormant in a hidden has caused the deepest despair, so will the awakening of beauty in this new century inspire us to greatness once again. Beauty will save the world. Michaël Samyn & Auriea Harvey Tale of Tales Studio Tale of Tales is a small independent studio operating out of Ghent, Belgium which investigates videogames as a creative medium to craft beautiful and poetic playful experience. mausoleum, surrounded by a magical forest. The time has come to awaken beauty with a kiss of love. A century without beauty has stimulated sinister sentiments in mankind. A century without beauty has put the entire planet and its precious atmosphere in mortal peril. This is no coincidence. And as much as the absence of beauty 10

11 TR 2014/04 ISSN X Emotion is Gameplay, Gameplay is Emotion The neuroscientist Antonio Damásio made a powerful argument that all rationality is actually based upon emotion and emotional memory in his works Descartes Error and The Feeling of What Happens. This was always a profound influence upon our philosophy of game design. We strive to move beyond mechanical complexity to emotional complexity, focusing on the internal interactions between player and game, rather than forming an experience through complex sequences of physical activity. For us, the dynamic emotional landscape created in real-time by the player s journey through an interactive world is the very essence of feedback. Games are architectures for creating emotional experiences and everything else confuses the tool with the goal. This talk will focus on this central idea behind our work and how it relates to the wider issue of emotional interaction between a user and a cultural object. Dan Pinchbeck, creative designer for TheChineseRoom game studio Dan Pinchbeck has a background in Drama, but he has worked in media, digital and technological arts practice using emergent technologies. He researches the significance of narrative within Computer Games, and the effective archiving of computer games. He is probably best known for his participation in the Dear Esther project. gameplay, rather than cold, logical loops of easily definable 11

12 TR 2014/04 ISSN X Overtures to the World: Videogames from Pataphysics to Participatory Art In this talk I will discuss an aesthetics of videogames that places videogames in a history of art tradition that goes alternative way to address the complex issues of aesthetics and videogames. from the classic avant-garde to postmodern participatory art practices. Taking as a starting point a phenomenological understanding of play and games as a source for the aesthetic experience, I will argue that videogames share with art movements the focus on performativity and dialogue as an essential element of aesthesis. Drawing on a number of contemporary games and art pieces, I intend to propose an Miguel Sicart, Associate Professor at the IT University of Copenhagen Miguel Sicart received his PhD in game studies in December His research has focused on providing a multidisciplinary approach to ethics and computer games, focusing on issues on game design, violence and videogames. He is the author of The Ethics of Computer Games (MIT Press, 2009) and Beyond Choices - The Design of Ethical Gameplay. His current research focuses on understanding playful aesthetics as a creative technological practice. He teaches game and play design. 12

13 TR 2014/04 ISSN X II. Full Papers 13

14 Architecting a Reusable Platform for Pervasive Augmented Reality Games based on Petri Net Models Tiago Agostinho, Ivo Cosme, Licínio Roque Informatics Eng. Dep. University of Coimbra Coimbra, Portugal tacagostinho@gmail.com, ivocosme@gmail.com, lir@dei.uc.pt Fernando Milagaia, Fausto de Carvalho Portugal Telecom Inovação, SA. Aveiro, Portugal fernando-j-milagaia@ptinovacao.pt, cfausto@ptinovacao.pt Abstract Pervasive Games require complex technical challenges to be overcome and their design space is yet mostly unexplored. In this paper we present a reusable platform for quickly designing, deploying and evaluating and managing multiplayer Augmented Reality Games. The games can explore visual and aural AR techniques, GPS and QR interface mechanics, and are structured as a set of activities defined using Petri Net models. These ARGs are typically played outdoors, using a smartphone client application. Evaluation of the platform involved field testing with realistic game designs exploring Cultural Heritage scenarios. Author Keywords Augmented Reality Games; Game Platform; Game Design. ACM Classification Keywords K.8.0 [Personal Computing]: General Games. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. General Terms Game Design; Augmented Reality; Pervasive Games; Petri Net. 14

15 Introduction Pervasive Games merge into real life, combining physical, social, and virtual worlds, exploiting mobile technology - sensing, positioning and frequently touchscreens - as interaction devices. Their potential remains mostly untapped in today s society, even though this kind of technology has become ubiquitous and advanced devices follow us everywhere, especially in urban contexts. Nowadays, the smartphone is a device that has a vast array of ways to interface with our surroundings, and present information to us. We can take movement, location, orientation and camera data to layer visual and aural information onto the real world, to augment it. Mobile augmented reality is starting to populate the mobile market, mostly by means of small but rich applications. Still, the process of creating a mobile AR game can be technically demanding, with many unpredictable interactions, both technically and socially. Moreover, the process of designing a game involves several phases and challenges, from idea to implementation. In this work we approached a previously recognized need to solve [CHI'2013 Reusable??] the problem of designing a general purpose reusable software infrastructure for defining, deploying and testing a diverse set of Pervasive Games exploiting AR techniques. Software, Game and Interaction Design disciplines all resort to diverse modeling techniques to get a better understanding of what to build, mostly involving some form of flow diagrams and state charts, along with models of the software logic. Some authors have shown that game related logic can be modeled to some advantage by using Petri Nets [??]. A Petri Net is a directed graph, where transitions can represent tasks or activities, and places are populated with tokens to define the state of the network representing the game system. Interconnections represent action dependencies on resources of certain subsets of the system state. (which consume a given amount of tokens and specify production of tokens on output). In this work we adopted Petri Net diagrams as a basis for specifying game designs, and used it to map these designs directly into a computable implementation that can be immediately deployed and tested. With this we studied the feasibility of a software system implementing this architecture and the practicality of defining and playtesting pervasive AR games based on a reusable infrastructure. In this paper we propose the architecture for a reusable platform for the quick creation and agile deployment of pervasive AR games, using smartphone technology. We used the multitude of interaction means supported by the smartphone to exploit techniques such as visual and aural overlays, location and direction tracking, time and proximity-based, and QR tag reading, as a basis to define a set of general action types. After defining game interactions they are encoded and their dependencies modeled using a Petri Nets modeling interface. This approach significantly speeds up the process of defining and deploying game variants for shortening the playtesting cycle. Over the next sections we will present the design and implementation of the proposed architecture and an evaluation of the platform that was performed by game design students. The evaluation involved executing field tests with realistic proof-of-concept game designs, that also provided interesting indicators on the quality, design spectrum and pervasiveness of the games supported. 15

16 Related Work Augmented Reality Games Augmented Reality has been applied in games since the early 2000s, with notorious seminal projects such as ARQuake [5] or the Human Pac-Man [7]. However, for several years, it was restricted to the use of complex (and heavy) contraptions of wearable machinery, such as head-mounted displays, large GPS trackers, and backpack computers. With the rising of stronger, lighter and more multifunctional devices such as smartphones and tablets, a new generation of AR games became more practical. Still today AR games are being developed mostly for small scale indoor scenarios involving the physical presence of fiducial markers or other small items that map into virtual objects or characters on a game setting. For instance, we have ARhrrrr [4], an Augmented FPS with an aerial view of a 3D mini town (mapped onto a paper sheet) which is overrun with zombies. Marker-dependent AR games may have prematurely reached their peak potential. Marketization has been achieved with for instance, the PSP game Invizimals [13]. On the other hand, AR games are still taking their baby steps in the field of outdoor games, especially marketwise. These comprise mostly treasure hunting or other types of rather simplistic or obvious interaction designs, limited to the context of the player s location and use of the camera. These limitations are often compensated with incorporation of multiplayer aspects (either for competition or collaboration scenarios). NBA: King of Courts [17], is a smartphone game in which physical places become basketball hoops where points are scored in a multiplayer social environment. Another reason for interface simplicity is that outdoor AR games nowadays are more directed to be played occasionally during mid-short intervals (ex: while commuting). Also, they are played on-the-move, with a smartphone, implying that the GPS, wireless connection, and video feed are already being used, leaving little battery power for such complex methods as contour recognition. Some projects managed to solve this restriction by using other computation devices, but sacrificing mobility. For instance, Carcade [6] is an in-car videogame for the passengers, which captures the landscape s silhouettes (via laptop and camera) and combines them into a racetrack. Pervasive Games Pervasive gaming is a rather new research area and still in the phase of exploration. Of all the topics related to our research project, this had the most innovative published references we could find. However, one can consider that the most confined types of pervasive games (indoor with only basic sensors) have already reached the market (via Nintendo Wii). Player Space Director [11] is a framework to ease the creation of pervasive games, by portraying as an abstraction layer that integrates the inputs of a collection of sensors and the metadata retrieved (e.g. gesture processing). Some proof of concept games were created such as the treadmill-racing game Swan Boat [2]. Once we move outdoors, most researchers aim to put to test highly original concepts. For instance, these can either use the entire Wi-Fi network of a city into the capture-the-base multiplayer game panoulu Conqueror [20], or enhance cooperation by using only locally created ad-hoc networks on the game Transhumance Team Exploration [9]. 16

17 One factor however is vastly common in research designs: Most outdoor concepts end up being loosely based on a treasure hunting mechanic, sometimes with a decent degree of narrative involved. The iperg [14] was a large-scale European project, that comprised designing and testing several pervasive games of diverse types. The initiative lasted almost four years and during that time it included treasure hunts and multiplayer alternate reality games, even featuring actor rosters and scripts. Ingress [12] was Google s recent take on Pervasive Gaming through a complex alternate-reality narrative where a faction is trying to establish portals around the world and the other is trying to stop them via gathering a virtual resource ( Exotic Matter ) and hacking Portals located at some city s points of interest. Petri Net Modeling of Games The feasibility of modeling games with Petri Nets has been proposed and studied previously. Since the use of UML at a conceptual or even specification level lacks formal semantics that prevent them from being used in rigorous model analysis [3], and Flowcharts are limited to the modeling of sequential, non-concurrent systems, with little or no support for conflict or concurrency for resources [3] we explored Petri Nets as formal models that are expressive, easy to learn and can be used as computable models. As proposed in [19], Petri Nets solve the gaps mentioned above. Petri Nets were invented in 1939 by Carl Petri with the purpose of describing chemical processes. The initial diagram concept has been extended in various ways and used in many and diverse domains. Petri Nets became an increasing asset during the rise of business modeling within the industry of Information Systems to represent workflow processes in a simple and accessible way [1]. The main reason behind this is that, with a few extensions, Petri Nets can model virtually any process. Since Petri Nets are graphic representations and composed by creative arrangements of repeated use of such a small set of elements, they become easy to learn and to use. Another advantage is that one can use a divide-andconquer approach by subdividing the process model into smaller Petri Nets, by zooming in and out on parts of a larger network. Definition of games with Petri Nets can be seen in [3], where a strategy game involving the Portuguese Maritime Discoveries was modeled through several subnets. These Petri Nets modeled both the game mechanics (such as the energy and state of the ship) and the flow of player action (the actions players could take at any given moment, considering the existing conditions). The ability to represent concurrency is a valuable asset for strategy games and massively multiplayer online RPGs. Also, as Petri Nets have a mathematically strict formality, the games flow can easily be simulated through existing tools, to help foreseeing issues before implementing the game. With small expansions to the original concept of a place, the authors of [22] were able to design and simulate complex quests for NeverWinter Nights via Petri Nets, before inserting them into the game using the built-in Plot Wizard. The extended Petri Net design was able to model every aspect about Non Playable Characters interaction and item achievement. 17

18 Multiple Answer Yes (7) No (5) QR Collect 7 QR Mobile Collect Image 1 A Petri Net sample Point And Click 5 QR Mobile Modeling AR Games with Petri Nets The structuring of the games on our platform is based on an adapted Petri Net model. Here, for readability, transitions are named actions, places are dubbed as resources, and dependencies that link them are named bridges or connections. Resources, along with the Tokens that populate them, define the state of the Petri Net, i.e. the Actions that can be executed at a given moment. An action can be executed if all its input resources have the required amount of tokens to be consumed and, upon execution, produces the specified amount of tokens in all of its output resources (except for a few macro-actions that encode differentiated flows of execution). Each player has its own private Petri Net network marking (which tokens are at each place, repesenting player state in the game). To enable definition of team gameplaying and other player interactions, some resources can be shared by all game participants state or only among team members. On top of this, the designer can define Goals that can be associated with resources to trigger when a specified condition or amount of tokens is reached. These goals can also be used to mimic the concept of obtainable or collectable items on an inventory. A point counter can also be linked to a resource, which supports the notion of a player's score or even a player ranking system. Player Actions can be of several types, including: Point and Click: If the player enters a proximity zone for a GPS coordinate, she will see an AR overlay when pointing in a given direction. Action finishes by clicking the object displayed; Listen And Click: Player is invited to hunt for the source of a sound with only the hearing sense (louder as the player the closes in in the right direction); QR Collect: collect printed or on-screen QR codes from the environment or from other players; QR Mobile Collect ( to be paired with QR Mobile): The first player reads a QR code directly from the second player s device. Used to model direct player interaction; Dialog Message: shows a message to the player, e.g. for narrative, orientation or other purposes, with an image and a textual component; Dialog (Single or Multiple) Answer: a question and answering interface (useful for narrative insertion and progression checking). Can be specified to produce tokens on different outputs depending on the answer given; Timed Event: an action that is triggered when a time condition is met, enabling other actions by releasing tokens satisfying their dependencies; Enter (and Exit) Proximity: this action fires when the player enters (or exits) the radius of a location; Player Router: automatically produces tokens on only one of its outputs, used to distribute or direct game flows based on random conditions; Player Selector: automatically produces tokens on only one of its outputs (based on some feature such as player s address). The example on figure 1 maps a game network model symbolizing the synergy of an ant colony. First, players would be asked Do you like sugar?, and according to their answer they are redirect towards one of the specified branches. If a player likes sugar, he would 18

19 hunt for sugar heaps via QR Collect (this action will be available 7 times, the number of tokens loaded on its precondition place). If he answers no, he can search for virtual (augmented reality) sticks to build the colony, via Point and Click actions (available 5 times). Once he collects 7 sugar heaps, or 5 sticks, he can then interact with a player that took the other path, and both reach the end of the game. A goal could also be set on the last resource place symbolizing the reaching of a final achievement or prize. The Proposed Architecture Definition The proposed architecture pictured in figure 2 is comprised of two servers (Back Office and Gaming) and an gameplay interface Client currently in Android. A client application runs on Android operating systems. The client communicates with the server via HTTP requests. Each time there is a change in game state affecting a player, the server pushes a notification (via Google Cloud Messaging) to the corresponding client application. The Petri Net models for determining and executing available game Actions is processed in the Game Instance and Player Instance processes. When a Client requests the execution of a given Action, this message reaches the corresponding Player Instance (after being handled by the service layers above). The Action request and its private state dependencies are sent from the Player Instance to the Game Instance for verification and processing of the shared game state, to determine whether the action is executable at the moment and to generate its outcome. If enabled, the Game Instance process performs the removal of tokens from the input Resources, and generate the outputs into the Action s output Resources (both private and shared). It then messages the new private Resource states to the Player Instance. The whole change to the state is replicated to the database via the Storage process (in an asynchronous way, thus guaranteeing better performance). GCM Views Controllers Back-Office Client (Android) Game Dispatcher Game Instance Storage Web Service Gaming Player Dispatcher Player Player Player Player Player Player Player Player Instance Figure 2. Proposed Software Architecture. Legend: Red Components Erlang Server processes; Green Component Android Client; Purple component Google Cloud Messaging Server; Connectors: Continuous messages, calls; Dotted HTTP; 19

20 When states (Petri Nets markings) change, all affected players device Clients are notified by the Game Instance process via Google Cloud Messaging (GCM an Android push notification system) [10]. This approach also creates a useful abstraction for the Client, which only knows the playable Actions for one player at a given moment, and thus handles only the Action presentation and its interface with the player (example interfaces pictured in figure 3). The current server is coded in Erlang [8], which is a functional language structured as concurrent processes that interact via messages (no locking mechanisms and no shared states). In this Erlang implementation, updating game state is done in memory, and therefore, with limited concurrency over the database stored state, that is managed by the Storage process. Dynamically, the runtime processing can be described as follows: The service side architecture has two major functionality blocks. We just described the Gaming component that is responsible for managing all running games. The Back-Office functional block is a Website for creating and editing game definition PN models via a web Browser interface. It is coded in Nitrogen, an Erlang Web Framework. This framework only handles the VC (View and Controller) components of the MVC architecture, leaving the Model component to be handled in the Storage component, which is ideal for this scenario, where we have only one database definition. Implementation Our first attempt at implementing the architecture included a much less scalable Ruby service that managed game definitions and state as Petri Nets (PN) encoded and updated on a relational database. That implementation was abandoned due to a serious bottleneck in database access, that rendered it incompatible with real-time PN execution. Player connections were managed through TCP/IP connections and there were massive state pooling requests that rendered the system quite unstable under intermittent network conditions. The Erlang server receives and parses requests on its WebServer module. It uses MochiWeb [16] as the HTTP server library A message is then sent to the Player Dispatcher (if it is of a private nature concerning just the player s state) or the Game Dispatcher (otherwise) Each online player has its own instance of the Player module, that logs locations and information that is not dependent on the specifics of each game he plays Management of the Petri Net begins at the Player Instance (one process instance for each player in each game holds every player private resource state) and the Game Instance (one for each running game holds everything shared between players of a game and manages game state changes) Every single piece of information, for both online games (running) and offline games (either stopped for editing or without any active players at the moment), is kept stored in the Mnesia [15] database (Erlang s builtin database module), that is accessed through Storage. Player Interface In figure 3, we can see two screenshots of the Client interface: Map view and camera View. 20

21 Map displays player s and visible Actions locations on a map of the surroundings. Camera (a.k.a. Gaming mode) comprises the almost all the visual content of the games. We can see that a total of 5 Actions is playable at the moment. In the bottom, there are 4 execution icons for (from left to right) Dialog Answer, QR Collect, Dialog Message and Listen and Click. The smiley on the center-left of the camera is a Point And Click s Augmented Reality object. It is tethered to a location a few meters away in that direction. If a player clicks the object it will be captured, completing the Point And Click. There are also other Client views for Inventory (related to the Petri Net Goals), Ranking, Chat and Message History, as well as a specific interface for each kind of "activity" that can be coded in the game design. Evaluation To evaluate the reusability of the proposed architecture with realistic game designs, we resorted to 3 student design teams from a Game Design and Development course from an Engineering Masters Program and from a European Masters on Cultural Heritage Studies. The Informatics Engineering game concepts developed were more focused on promoting interaction between players. The EuroMACHS design was centered on adapting a Cultural Heritage scenario into a game format, with a rich multipath narrative. To translate the concepts into the platform, some adaptations were made. These included mapping the game concept into the required Petri Nets models and rethinking some elements to take advantage of the proposed reusable action types. The adventure style narrative game concept became a Petri Net with over 70 Actions encoded, while the other games were significantly simpler and more "circular" or repetitive, with fewer actions and more cycles. After tuning the proposed game designs, the final games deployed were: City by Night - player complete a set of challenges spread throughout the nocturnal establishments of the City. The barmen also played to confirm (using the QRCollectMobile-QRMobile action pair) that certain challenges have been fulfilled (e.g. drinking a shot). PIDE vs Revolucionários - a multiplayer, teambased game inserted on a pre-revolutionary Portuguese context, where players take on one of two roles, either the regime police or the insurgents, and walk to explore the area searching for clues to the identity of their adversaries, whom they have to capture before being discovered themselves. Mystery of Alta - a singleplayer game with a dense narrative, where the player has to communicate with the game characters (helpers/opponents), to act or to decode clues, to explore the University surroundings, acquiring knowledge about its history and traditions. This collaboration culminated with a game exhibit day where several users tried out the applications. This corresponded to our field test, since multiple game instances were running at the same time and it was possible to understand in loco the difficulties posed by some interfaces and the challenges associated with envisioning how the players might act at each specific location. To collect data, all requests to the server were logged, along with some important data, such as, location and player ids. Due to its high Petri Net 21

22 complexity paired with a large array of location-based Actions, the Mystery of Alta game was also used for a more extensive analysis leading an estimation of player mobility. Results The support of three different game designs acted as a first feasibility test for the platform and its reusability. Both City by Night and Pide vs Revolutionários escaped from the trend of treasure hunting that we have seen on Pervasive Games. They were more focused on player interaction than geolocation. These concepts were only made possible within the platform by the introduction of direct player interaction actions (i.e. QRCollectMobile and QRMobile pairs) and routing actions (e.g. Player Selector), thus pointing to the relevance of these action types. Cooperation with the design teams revealed the importance of having a platform to quickly and iteratively deploy game concepts with less coding knowledge, which gave them more time to think on the game elements and scenarios. On the other hand, their usage of the platform served to validate existing functions and as participatory design opportunity as they requested changes/additions such as player routing, random flow, and multiple answer redirection actions that significantly enhanced the spectrum of possible game definitions. The field tests during the exhibit day produced a hefty amount of log data. Some indicators could be drawn from the Mystery of Alta case, as shown in table 1. These values were defined to reflect an exploration rate (an approach to movement detected between actions), since it is based on the distances between locations of consecutive requests. The mobility rate was calculated to comprise all blocks of 10 seconds for which cumulative distances were more than 10 meters, signaling the players were active exploring the space (to discard small movements or GPS corrections). Indicator Mean % Time exploring (player mobility rate) 61% Average exploring speed Table 1. Information drawn from Field Test results 3 km/h In Mystery of Alta, with such large amounts of dialogues and questions (which draw on the player s attention, creating intervals without movement), it was rather interesting to verify that players dedicated 61% of their game time exploring the action s physical surroundings. An average moving speed of 3 km/h is close to human s default walking speed (5 km/h [21]), meaning that exploration was mostly made on-themove, which we think might be representative of what to expect from pervasive games. Also, one should notice that the speed isn t too much on top of the human s walking speed. This means that players did not just walk straight to the Action s target location, but instead deviated from the optimal route between actions, thus also indicating exploration. Currently there is still a lack of reference values to understand what to expect while designing pervasive games, therefore, these indicators can contribute with a first useful approximation. Acknowledgement This research resulted from the AdVenture project developed at IPN, in partnership and with funding by Portugal Telecom Inovação in the scope the company s Plano de Inovação and

23 Conclusions In this paper we presented an architecture for a reusable platform for designing and playtesting pervasive AR games, validated through a set of design cases and field tests. We also presented our approach to modeling pervasive AR games resorting to Petri Net models and a definition of general purpose action types to be reused. We concluded that the approach for defining and computing games with Petri Net definitions is valid and agile, and that it could be well understood by game designers to an interesting spectrum of game models, that were quickly deployed and playtested. By calculating player mobility rates and speeds we validated indicators for the pervasiveness of the game designs. References [1] Aalst, W. and van Hee, K. Workflow Management: Models, Methods, and Systems. MIT Press (2004) [2] Ahn, M., Choe, S., Kwon, S., Park, B., Park, T., Cho, S., Park, J., Rhee, Y. and Song, J Swan Boat: Pervasive Social Game to Enhance Treadmill Running. Proc. 17th ACM international conference on Multimedia, ACM (2009), [3] Araújo, M. and Roque, L. (2009). Modeling Games with Petri Nets. Proc. of the DIGRA2009, London. [4] Arhrrrr Zombies. [5] ARQuake. [6] Carcade. /carcade-in-car-gaming [7] Cheok, A., Fong, S., Goh, K., Yang, X., Liu, W. and Farzbiz, F. Human Pacman: a sensing-based mobile entertainment system with ubiquitous computing and tangible interaction. Proc. NetGames 2003, ACM (2003), [8] Erlang. [9] Gentes, A. Guyot-Mbodji, A. and Demeure, I.. Gaming on the Move: Urban Experience as a New Paradigm for Mobile Pervasive Game Design. Proc. MindTrek 2008, ACM (2008), [10] Google Cloud Messaging for Android. [11] Hwang, I., Lee, Y., Park, T. and Song, J. Towards a Mobile Platform for Pervasive Games. Proc. of the MobileGames 2012, ACM (2012), [12] Ingress. [13] Invizimals. [14] IPerG Integrated Project of Pervasive Games. [15] Mnesia [16] Mochiweb. [17] NBA King Of Courts. [18] Nitrogen. [19] Petri, C., and Reisig, W. Petri Net. Games. Ext. Abstracts CHI 2009, ACM (2009), [20] Tiensyrjä, J., Ojala, T., Hakanen, T. and Salmi O. panoulu Conqueror: Pervasive Location-Aware Multiplayer Game for City-Wide Wireless Network. Proc. of the Fun and Games 2010, ACM (2010), [21] Transafety Study Compares Younger and Older Pedestrian Walking Speeds. [22] Young-Seol L. and Sung-Bae C Dynamic quest plot generation using Petri net planning. Proc. of the WASA 2012, ACM (2012),

24 An Interaction Design Model for a Petri Net based Behavior Editor Nuno Barreto, Licínio Roque Informatics Eng. Dep. University of Coimbra Coimbra, Portugal nbarreto@student.dei.uc.pt, lir@dei.uc.pt Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Abstract In the context of Game Design, to define behaviors and choreographies of actors in a game or simulation context can be a challenging and complex programing task. This paper proposes an interaction model for a Petri Net behavior editor to be used as a visual language for game behavior modeling. A proof of concept prototype implementation of the proposed interaction model was created and validated with formal usability lab tests. In spite of some usability issues, most users were able to complete the proposed game behavior definition tasks using the editor interface. We think this provides evidence to reinforce the case that Petri Nets can be used to advantage in the game modeling process and, when coupled with a runtime simulation, can provide an interesting immediate feedback loop for faster design and experimentation. Author Keywords Complex Systems; Behavior Modeling; Game Design; Simulation; Petri Nets; Visual Language ACM Classification Keywords D.2.2. Software Engineering: Design Tools and Techniques. 24

25 General Terms Design Languages; Digital Games; Experimentation Introduction Video-Game development is a multidisciplinary area which can encompass diverse skillsets. One of such skillsets is programming which has an important role in the development as it translates the game concept into an interactive artifact known as a video-game. Since game programming is what contributes to the creation of aspects such as how game objects (referred from now on as actors) are modeled in an environment and how they interact with each other and with their environment, it becomes necessary to translate every possible course of action as code into actors which can be an error-prone task. This paper presents an interaction model for an editor that is integrated in a solution proposal used to simplify game programming that is centered around the creation of a visual language, based on Petri Nets. More specifically, a language that is intended to model concurrent actors behaviors and choreographies (i.e. how the behavior translates into visual and sonic feedback) in the game world as stage. Petri Nets have been proven accessible and easy to learn by non-programmers and provide an economical way of specifying behaviors in complex systems. As a design tool, they give the advantage of a simple visual language that promotes agile modeling and testing of complex interactive systems. This paper is structured as follows: the Background section contains state of the art research regarding video game development and the usage of Petri Nets as well as their editors and data structures, the Methodology section illustrates the development approach used for the creation of the solution proposal, the Interaction Model section states the model used and its specification, Evaluation demonstrates the evaluation s results and analysis and finally Conclusions will sum up the results and state some further work. Background Developing Video Games Developing video-games has been made easier with the increasing existence of game engines - a system whose purpose is to abstract common, and sometimes platform dependent, computational game related tasks [28] - and other development tools, leaving developers with the choice of creating their game by either building their own engine using tools such as SDL [20] or XNA [26] or using pre-existing engines including UDK [24] Unity [25] or even CryEngine [7]. Most, if not all, of these tools require some knowledge in programming which imposes a learning barrier to more novice developers and are difficult to maintain, as coding is heavily dependent on a good architectural design to be readable and extensible. Nevertheless, to counter programming s inherent steep learning curve and maintenance issues, some tools provide visual languages such as Scratch [19], Stencyl [22], ToonTalk [23] and Agent Sheets [2] and some game engines grant their users graphical modeling mechanisms (an example of this is UDK s Kismet, Unity s Mecanim and Cry Engine s Decision Tree Editor) that are utilized to build some aspects of a games (making them somewhat limited because coding is still 25

26 Figure 1. Steps in Design Science Research required when developing aspects not covered by these mechanisms). These visual approaches rely on either system block building or classical Artificial Intelligence algorithms and data structures, usually present in behavior modeling, such as Cellular Automata, State Machines, Behavior Trees, Flow Charts and Rule-Based Systems [13]. These visual data structures usually present a readability problem and their maintainability quickly decreases with increased model complexity. Not as widely used as the aforementioned graphical tools, Petri Nets [15] have already been used to describe some game aspects including their plot as depicted in [6], level sequences as shown in [14] and even entire game models as demonstrated in [8] and [3]. It was showcased in the latter work that Petri Nets appear to be easy to use and learn and support models validation through simulation. Petri Nets are also extensible and, consequently, have a wide variety of augmentations that add functionalities which, in turn, reduces potential model complexity. Because of their apparent smooth learning curve, validation functionality, extension capability and the demonstrated ability to model several aspects of a video-game, this tool was considered as a candidate for the base of the language described in this paper. Petri Net Editors There are several Petri Net editors available that can be used to model interactive systems. These include PIPE [16] Woped [29], Yasper [23] and many others. Since these tools are used to illustrate a wide variety of systems, they are deprived of semantics and their main application is to develop concept diagrams that demonstrate how systems function. As such, simulation functionalities present in these tools are mainly for debugging purposes. However, there are some editors that allow users to add semantics to their Petri Net models. Examples are JFern [9] and JPetriNet [10]. As an attempt to make Petri Net models interoperable and standard, a description language, called Petri Net Markup Language (or PNML) [5], was created. This language is built upon XML and besides describing how Petri Nets are distributed in a specific model, it also allows to add graphical attributes, which provides visual editors the means to render the models, and to append tool-specific attributes that can only be parsed by designated editors. The editor developed borrows some interface aspects from the tools mentioned above while maintaining simplicity. It also uses PNML to store models persistently in an interoperable way, so that these models can be read on other authoring tools. Research Methodology Design Science Research [11] is a methodology that aims to produce a statement of learning as a consequence of research made through design or, in other words, this methodology s objective is to solve problems with the purpose of producing a statement of learning. The Design Science Research, as illustrated in Figure 1, encompasses 5 steps: Awareness of Problem, Solution Proposal, Prototyping, Evaluation and Statement of Learning, each producing its own artifacts. Since this methodology can be used in any area where design is possible, the following clarification of the methodology s steps states example artifacts best suited in this project s context. 26

27 The first step, or Awareness of Problem, comprises the definition and identification of a problem. In order to help clarify this definition and identification, State of the Art research was made. In the following step, Solution Proposal, a suggestion for the problem s solution is created through abduction drawn from the state of the art research made previously. This step outputs interaction models and software architecture. Prototyping comprehends the actual development. In this stage, software is produced, using the conceptual models created previously. The artifact produced acts as a proof of concept, proving that the proposed solution is possible. The next step, or Evaluation, attempts to validate the prototype created in the previous stage using an evaluation model. When such model is nonexistent, one is devised alongside other Solution Proposal s artifacts. This step also provides feedback, or circumscription, to the other previous steps which permits an iterative and incremental development (the agile development process is used during the iterations Prototype- Evaluation). Finally, in the Statement of Learning step, the project is concluded so that knowledge can be produced. Artifacts produced in this stage include concepts, models, methods and prototypes. In our process a Proof of Concept implementation was produced to study the proposed Interaction Model and associated Concepts. Proposed Interaction Model As written, the main purpose behind this solution proposal was to develop an easy to use modeling tool, based on the Petri Nets (PN), that could be used, by non-programmers, to define actor behaviors and choreographies in a game/simulation environment, whether running on one machine or on a distributed architecture. Intrinsically, this tool was designed to provide the ensuing quality attributes: portability so that it would not be tied to a particular OS, interoperability which would allow to use the tool with different game/simulation engines, usability to complement Petri Net s accessibility, error recovery due to the fact that designers will be working with a language with a defined syntax and, therefore, they must be warned of syntax errors and how to solve them and scalability relative to the number of actors and/or players. Given the solution s objectives and non-functional requirements, the proposal was devised to contain the visual language s specification (syntax and semantics), an easy to use visual behavior editor and a language execution engine used to translate the Petri Net models into in-game/simulation actions. Language The language used in this solution proposal contains a similar syntax to that of Hierarchical Petri Nets [1] with weighted arcs. By utilizing the capability of grouping sub-nets, this language is able to reduce graphical complexity and thus, improve readability. Another important advantage is that it promotes component reutilization, i.e. the same sub-net can be used in different contexts. 27

28 Editor s supported actions: Add/Remove Language objects (places, transitions and tokens) Link places and transitions through arcs Create sub-nets Edit objects properties Filter sub-nets according to keywords Export/Import sub-nets Save/Load Petri Nets to/from disk Undo/Redo actions MODEL A game/simulation model is represented by a root Petri Net that contains a set of Petri Net models, each symbolizing a different actor archetype. It is worth noting that instances of an actor archetype share the same Petri Net model. Every child of the root model is an independent net that is associated to an Actor and follows the language syntax detailed earlier. These PNs cannot communicate directly with one another by means of arcs, but only through explicit messages - tokens moving from an output place of a PN into an input place of another PN. Places can be of four types: Input Places, Output Places, Fused Places or Regular Places (this designation must not be confused with the naming given to places linked to/from a transition on the original Petri Net language). Regular Places share the same meaning as places in the Petri Net language. Input Places act as actor sensors. This means that when a token arrives at these special places, something was perceived by the actor. Output Places, on the other hand, assume the role of announcers, i.e. when they receive a token, it is announced to the game world that something has happened. Furthermore, an Input Place can be used to observe an Output Place. Finally, Fused Places are places inside sub-nets that are linked with outer-net places, acting as proxies for their outer-net counterparts. Whenever an outer-net place receives/loses a token, its Fused Place receives/loses the same one as well. Unlike places, which have four different types, Tokens, just like in the Petri Net language, stand for a condition that was met. Transitions, however, have associated programming scripts that govern actions. When a transition fires, its script is executed, meaning that an action is taking place. Interface The interface s design originated from an iterative process. Initially, a paper prototype was constructed. This prototype was then evaluated through user testing so that it could be refined. After several iterations, the prototype was converted into a mockup representation using Balsamiq [4], as illustrated in Figure 2. This proposal is adequate because the interface s viewport provides the necessary information for the simulation and manipulation of the language s constructs in a segmented way. Because of this, users can easily interact with the editor without having to navigate through menus in order to look for actions. Furthermore, the spatial distribution helps organize the information so that users don t feel overwhelmed. Figure 2. Mockup of the petri net editor's GUI 28

29 As can be seen in Figure 2, the editor is divided into 7 panels or menus: 1. Input Panel. Here, designers are allowed to put places that represent the actor s inputs. These inputs can range from sensors (vision, hearing, etc ) to messages containing information. 2. Function Panel. In this panel, designers can model the actors logic. As depicted in the figure, this logic can contain both places and transitions. 3. Output Panel. Similar to the Input Panel, places set in this panel represent the actor s outputs, or information he transmits to the game world. 4. Group Panel. A panel that lists grouped Petri Nets. 5. Search Panel. A panel used to filter the Group Panel, through the means of a keyword search. 6. Button Sidebar. A sidebar containing the most important action buttons. From top to bottom, these buttons are: Play, to simulate the given petri net; Add Place, as the name indicates, inserts a place onto one of the panels 1 to 3, as given by the cursor; Add Transition, functioning as Add Place but adding a transition instead; Group, used to gather selected sub Petri Nets into one transition; and Validate to check whether the petri net is valid or not. 7. Menu Bar. A menu bar where designers can save/load petri nets to/from disk and import/export groups, if implemented. Proof of concept implementation The editor s backend was built using Java due to its portability and the interface was made with Java s GUI API: Swing. Initially, the editor was meant to be based on the JFern Editor because it provided JFern s Petri Net threaded simulation mechanism, data structure and PNML exporting/importing API as well as views and controllers for their visual representation. This idea was discarded because the GUI s code was poorly documented, confusing and some of its classes were not made available as source code. Consequently, only the simulator, data structure and PNML parser were used. The reason behind using JFern is that besides offering the previously mentioned set of tools, it was the only tool from the ones researched that allowed to introduce code to be executed when a transition fires, thus reducing some programming effort when developing the editor s execution mechanism. This tool was modified to provide some additional attributes to the Petri Net s objects. Figure 3. Initial version of the editor Evaluation Evaluations were concentrated on the interface as it was a crucial part of the application and encompassed most of the required attributes and objectives devised for the project. The type of tests chosen to evaluate the 29

30 Description of the test s tasklist: 1. Read crash course and explore the editor for 5 minutes. 2. Read the game s design doc and start a new project. 3. Build a chronometer mechanism. 4. Create a score update mechanic. 5. Build the player s navigation system. 6. Create the player s shooting mechanism. 7. Make a bot spawning mechanic. 8. Devise the enemies AI. 9. Integrate the score update with a local scoring system. 10. Make a winner announcement system. 11. Save the project. 12. Open a project and answer some questions regarding the language. interface were usability tests [21]. The main goal of these tests was to verify how users experienced the creation of video game definitions using the editor. Test Setup As previously written, the application s interface was evaluated by means of formal usability lab tests. These tests had an expected time of completion of approximately 1h30m, however subjects were free to surpass this schedule. They consisted in individual sessions where each voluntary tester was prompted to setup and define the behaviors in a video-game using the thesis application in conjunction with the Unity game engine and pre-existing graphical assets (level and character 3D models). During those sessions, testers were accompanied by an evaluator, whose job was to clarify any rising questions and to take notes of events that could happen during the test. In order to help document any event that might escape note taking, audio was recorded. Each session followed a predefined script. Primarily, test subjects were introduced to the project s context and test objectives. Secondly, they were asked demographic questions for later analysis of the population performing the test; the actual test started afterwards, when testers were given a document with information regarding Petri Nets and were encouraged to explore the interface for 5 minutes, after which they were given the game s design document and a list of tasks that contributed for the creation of said game. After each task, test subjects estimated its difficulty in a scale of 1 to 5. Subsequently they were interviewed to detail their overall user experience and performance and were requested to list the top 5 best and worst aspects of the interface, according to their opinion. The game that test subjects were supposed to create was, as stated, described in a pre-made design document and detailed in a task-list. This document defined the context of the game, its rules, actors and sensors and scripts that were available to them. The task-list helped guide the users in the completion of the game by dividing it into tasks. The first half of the list contained a step-by-step guide while the second half was only comprised of objectives. This way, testers, during the first half, could learn the basics of the application as well as its quirks. Overall, the game consisted in a competitive first person shooter where players and AI-controlled bots had to toss balls at each other in order to increase their team s score. The following screenshot illustrates the game, as made by one of the test subjects. Figure 4. Screenshot of the test game Spheres of Steel as created by one of the subjects Tests were performed using 11 subjects [17] and the data collected from the recorded audio, demographic questionnaire, interviews and notes, was categorized into three classes: demographic information, user performance and usability issues. Although 11 testers participated, only 10 completed the development of the 30

31 Figure 5. Total Time per User for the complete exercise. Figure 6. Average Perceived Difficulty per Task (scale 1-5) game and, therefore, the information aggregated from the subject who had to abandon the test midway, due to personal reasons, was only used in the demographic information and usability issues as there was not enough information necessary to compile in the user performance category. Demographic questionnaires required subjects to state their age, sex and highest academic degree. They also inquired users to rate their experience, in a scale of 0 to 2 0 meaning never heard of the term and 2 denoting highly proficient - in textual programming (TP), visual programming (VP) and game development (GD). The reason for this is that textual programming introduces people to algorithms; Experience in visual programming would make the subjects used to the mannerism required to manipulate a visual language; and experience in game development would make users accustomed to the steps involved in creating a game. The population sample is composed of 2 females and 9 males, with an average age of 26. Their qualifications range from Bsc student to Phd student. From this information, it can be deducted that the highest degree achieved by the test subjects range from High School to a Msc coinciding with the academic education that game designers often have. The average experience in textual and visual programming and game development is, as self-stated, 1.18/2, 0.45/2 and 0.72/2 respectively. This means that subjects are familiar but not proficient in textual programming, barely know about visual programming but have a little knowledge of game development. Overall, the subjects constituting the population sample were selected in a manner that allowed for a heterogeneous sample, in means of qualifications and experience levels. This way, in theory, it would increase the amount of issues found by subjects. Results and Analysis USER EXPERIENCE In this context, user performance consists in the overall time subjects took to complete the game s definition and each individual task and their relative perception of the difficulty of every task. This was extrapolated from the audio recordings and ratings that testers gave after finishing their tasks. From the data presented Figure 5, it was concluded that on average, testers completed the test in 1h33m, only 3m above the expected time, and their perception of the test s difficulty was, on average, 2/5 this was derived from the values available in Figure 6. This means that users thought the test they made, while using the application, was easy. Nevertheless, only 1 out of 10 subjects did not require the evaluator s assistance. USABILITY ISSUES The notes and interviews originated a list of usability issues. These issues, after compiled and normalized, were categorized according to their importance [17], occurrence frequency, type and occurrence by task and by user. From the usability tests, 406 occurrences, distributed across 88 different event classes, were found. There were only three importance levels given to issues: High, Medium and Low. These levels were attributed according to the issue s degree of prevention in completing a certain task. Each level corresponded to a number: High corresponded to 1, Medium to 0.66 and Low to

32 In total, there were 9 types used to classify the issues [12]. These types were Functional Error (FE), Affordance (A), Feedback (FB), Perception of System State (PSS), Naming Interpretation (NI), Instruction Interpretation (II), Representation Interpretation (RI), Mappings (M) and Domain Knowledge (DK). It was assessed that most events lie on the category of Mappings. This means that during the tests, most recorded events were comprised of discrepancies between their users intentions and the interface s available actions. This results in a value ranging from 0 to 1 because all variables were normalized beforehand. By multiplying these factors, it is assured that, for instance, issues that appeared frequently during tasks, were encountered by most users and tasks, and hindered the completion of said tasks are given more priority than issues that, for example, were not as frequent or important. A chart detailing the priority levels per problem, sorted by value, is presented in Figure 7. The revised interface is illustrated in the screenshot in Figure 8. Figure 7. Priority Level per Issue USABILITY CORRECTIONS We used a metric to help identify the most critical problems sorted accordingly to their potential impact, to elaborate a correction plan. This metric consisted in a two part algorithm. In the first part, a value, referred to as priority level, was attributed to each issue by calculating the arithmetic product between the its frequency per user (IFU) and per task (IFT), its relative frequency (RF) and its importance (I). The formula is given by the expression. Figure 8. Screenshot of the editor interface (post usability corrections) The most notorious differences between Figure 3 and Figure 8 rely on the tool bar. In it, some toggle buttons were introduce to provide a better indication of the system s state. Another difference is that the buttons were sorted and grouped with separators to avoid button pressing mistakes and to decrease the time it took an user to look for a specific button. Finally, some label names were changed to avoid confusion 32

33 Conclusions In sum, this paper introduced an interaction design model for a behavior editor used to edit constructs of a language, based on Petri Nets, intended to define actor behaviors and choreographies. This design was instantiated as a proof of concept and validated through formal usability lab tests. We concluded that in spite of pending usability issues most users were able to complete the proposed game behavior definition tasks using the PN editor interface. We think this provides a good evidence to reinforce the case that PNs can be used to advantage in the game modeling process and, when coupled with a runtime simulation, can provide an interesting immediate feedback loop for fast design experimentation. For future work, it would be interesting to develop a look and feel more appealing to the editor s target audience, as this attribute was not considered during the development of the editor, since the main focus was to create a prototype to illustrate the interaction design model. Acknowledgements We would like to thank all the support from lab colleagues for the help they provided through the development of this editor, and the test volunteers for their availability to help validate this work. References [1] Aalst, W. Hierarchical Petri-Nets. Eindhoven University of Technology, [2] AgentSheets. [3] Araújo, M., & Licinio, R. Modeling Games with Petri Nets. In Proc. Digital Games Research Association (DiGRA) on Innovation in Games, Play, Practice and Theory (2009). [4] Balsamiq. [5] Billington, J., Christensen, S., van Hee, K., Kindler, E., Kummer, O., Petrucci, L.,... Weber, M. The Petri Net Markup Language: Concepts Technology and Tools. Applications and Theory of Petri Nets (2003), [6] Brom, C., & Abonyi, A. Petri Nets for Game Plot. In Proc. AISB on Narrative AI and Games workshop (2006). [7] CryEngine. [8] Dormans, J. Machinations Framework. hp?title=machinations_framework. [9] JFern. [10] JPetriNet. [11] Hevner, A., & Chatterjee, S. The General Design Cycle. Design Research in Information Systems: Theory and Practice (2010), [12] Hourcade, J. Usability Principles. University of Iowa, ecture2.pdf. [13] Milligton, I., & Funge, J. Artificial Intelligence for games. Morgan Kaufmann, Burlington, USA, [14] Natkin, S., Vega, L., & Grünvogal, S. A new methodology for Spatiotemporal Game Design. In Proc. CGAIDE'2004, Fifth Game-On International Conference on Computer Games: Artificial Intelligence (2004). [15] Petri, C., & Reisig, W. (2008). Petri Net. [16] PIPE

34 [17] Sauro, J. 10 Things to Know about Usability Problems. [18] Sauro, J. Applying the Pareto Principle to the User Experience. [19] Scratch. [20] SDL. [21] Sharp, H., Rogers, Y., & Preece, J. Interaction Design: beyond human-computer interaction. Wiley, USA, [22] Stencyl. [23] ToonTalk. [24] UDK. [25] Unity 3D. [26] XNA Game Studio. [27] Yasper. [28] Ward, J. What is a Game Engine? s_a_game_.php. [29] WoPeD. 34

35 EEG Biofeedback and Brain Computer Interface in Games Rodrigues, João. Rosa, Agostinho. System and Robotics Institute, System and Robotics Institute, Instituto Superior Tecnico, Instituto Superior Tecnico, Technical University of Lisbon, Technical University of Lisbon, Portugal. Portugal. Nan, Wenya Department of Electrical and Computer Engineering, Faculty of Science and Technology, University of Macau, Macau Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Abstract This work proposes a complete structure of an EEG biofeedback platform focused on an efficient way for its users to learn how to self-regulate their cortical activity and use it as an input for Brain-Computer interfaces. This platform was tested in 20 subjects that underwent 20 short sessions of training. Although the main focus of this study was to understand the positive effect this training has in short term memory and how EEG selfregulation can be achieved, in this work we analyze the results in the perspective of a Brain-Computer interface training tool. Therefore, besides introducing our protocol, we present the mental strategies subjects used to control their EEG, their improvements along training sessions and the delays related to this kind of interface. The positive results obtained by this training, along with the time delays necessary to voluntarily produce changes in the EEG made us consider the usefulness of this ability as input for a videogame. We briefly discuss how it is being used and why it may not be challenging or interesting for gamers and propose different uses of this additional mean of gameplay. Author Keywords Brain-Computer Interfaces, EEG Biofeedback, Game Interfaces 35

36 EEG Biofeedback Platform [8] Select Channels: must be channels from locations we know that can produce voluntary EEG and exist in available portable hardware. Select feedback feature: Power from a certain frequency band in a certain channel, some operation between bands from different/same channels. Connectivity between brain sites. Muscle or eye movement. Goal 1: the feedback feature value surpasses some threshold. Goal 2: Goal 1 is being achieved continuously for more than a certain amount of time. What is shown?: the feedback parameter value can be shown but Goals 1 and 2 achievement must be more evident. Still they should not disturb focus. Session structure: trials and intervals, fixed or not. Introduction EEG biofeedback consists in making an individual aware of his cortical activity by presenting him this measured signal. The aim is allowing its conscious control but, as the EEG is a rather complex signal, any success in voluntarily influencing it can be unperceived, making it impossible to acknowledge any change. To ease the task of understand the EEG behavior, it can be divided into several frequency bands: delta (0.5 to 4 Hz), theta (4 to 8 Hz), alpha (8 to 12 Hz), the sensorimotor rhythm (SMR) (12 to 14 Hz) and beta (14 to 26 Hz). EEG biofeedback is usually focused in the amplitude of some of this frequency bands rather than the entire signal. An EEG biofeedback system can be seen as a closed loop enclosing neuronal activity and a brain computer interface (BCI). This BCI consists in a signal acquisition device connected to a computer that processes the EEG information and translates it into stimuli to be perceived by its user. This translation can be done by calculating different characteristics of the EEG such as the EEG coherence value between different locations in the brain, the power spectrum or amplitude of a specific frequency band or electrical slow cortical potentials (SCP) [1]. The result is presented to the user through the computer screen, in the form of visual stimuli, or speakers, in the form of auditory stimuli. Ideally, after an initial learning period, the subject is able to voluntarily induce changes in his EEG that are reflected by a specific outcome in the feedback parameter. This way the subject is able to modulate any external system, using the EEG as input, alternatively or complementary to motor skills. Subjects suffering with neurological diseases that result in motor disability that lose their capabilities of communication and expressing justify these interfaces as it might be their only way of communicating [3]. These applications are limited to binary choices as they re based in high or low values of a specific frequency band or signal amplitude. SCP control has been proven to be possible by human subjects in several BCI [4] and EEG biofeedback studies [5]. A cursors vertical position can be determined by this potentials amplitude after it is calibrated to the user s amplitude range [6]. SMR rhythms are usually recorded over the sensorimotor cortex and its amplitude is related to sensory input, movement or motor imagery. Its control by human subjects is also proved to be possible and the cursors movement vertical position is also dictated by the band s amplitude. Left to right movement with a constant rate can be introduced to give access to the extra dimension [3]. The selection of a character, for example, can be done using a 6x6 matrix filled with characters. The rows and columns of this matrix flash with a certain constant rate and the user is instructed to count the number of times the flashing row and column crossing point coincides with the desired character. When this happens, because of the attentional demand similar to an oddball test, an electrical potential is evoked. At every flash, the EEG is recorded and associated with one character. After a certain amount of flashes, each EEG segment is averaged and the averaged signal with the highest amplitude is probably the one where more electric potentials were evoked leading to the conclusion that it s the option desired by the user [3]. Still, in all strategies the maximum communication bit rate is low (from 0.1 to 3.5 bits/minute [7]). EEG Biofeedback Training Because self-regulation of the EEG is not a usual skill, several training sessions are needed for the user to understand what cognitive strategies lead to the desired changes in the EEG. Because these sessions are based on 36

37 operant conditioning, their protocol, according to M. Sterman and T. Egner, must follow certain rules in order to be effective [1]: Each training session should provide discrete trials separated by brief pauses. When the produced changes in the EEG meet the required condition (for example, increase band amplitude until a certain threshold) a reward stimulus must be presented. There must be minimum delay between the reward situation and the reward stimulus for optimal learning to occur. The reward stimulus must have the highest reinforcement effect. Also, the electrode placement should be determined by the International System of Electrode Placement since it is based on the location of cortical regions and uses relative metric, given that head sizes vary. Platform Features The developed platform [8] was designed to allow different EEG biofeedback trainings and their evaluation. Besides the flexibility in defining different biofeedback protocols, it allows linear and non-linear spectral analysis, labeling the signal to check for event related potentials during different cognitive assessment tests and EEG biofeedback and tracking individual frequency bands. Training Structure Each training program is composed by several sessions. Based on the rules recommended by M. Sterman and T. Egner for the training effectiveness [1] each session can have several trails separated by intervals. Each trials length and number can be fixed (Figure 1) or controlled by its user (Figure 2). Letting the user decide the number of trials and how long each one lasts can be useful in the initial learning period. Here, at the end of each trial, the user can write down what cognitive strategies were used to change brain activity. This is helpful in the next sessions because the user can acknowledge what strategies produce better outcomes and use them in those sessions. After the learning period, protocol defined sessions can be used. Figure 1. Number of trials is defined in the protocol as well as their duration and intervals in between. Figure 2. Adaptation session structure. User decides when to start and end trials. TRIAL OBJECTIVES Each trial is guided by two goals. The first, Goal 1, consists in the comparison of the value of the feedback parameter with a predefined threshold. Here, two choices can be made in the protocol: the goal is only achieved when the feedback parameter value is above 37

38 a certain threshold or it is only achieved when the feedback parameter value is below a certain threshold. The second goal, Goal 2, is related to the period of time the first objective keeps being achieved continuously. If Goal 1 is being achieved continuously for more than a predefined period of time, Goal 2 is accomplished. Each trials score is based on these two goals. It is also possible to introduce trials where the user must not try to accomplish these two goals but the feedback parameter is still being fed back to him (Figure 3). By comparing the results of these trials with the results from those guided by goals it is possible to see if the user is really voluntarily changing his EEG towards the objectives. Figure. 3: Session with control trials where the subject must not try to achieve any goal. EEG Measurement and Translation Because each brain activity pattern has its source in different cortical locations the placing of the recording electrode in the scalp must be identifiable. It is possible to use any of the electrode placements defined in the International System of Electrode Placement. The training consists in the increase or decrease of the amplitude of a certain frequency band. The calculation of the feedback parameter value is done in the frequency domain by calculating the amplitude of the frequency band that is being trained relative to the whole EEG amplitude - relative amplitude (Equation 1) - or another frequency band. bandamplitude ramplitude (1) EEGAmplitude In this case, the EEG frequency of interest is considered to be between 0.5 and 30 Hz. Using the amplitude spectrum instead of the power spectrum prevents excessive skewing that results from squaring the amplitude values which increases statistical validity [1]. Individually Adjusted Frequency Bands The boundaries of the different frequency bands referenced previously are standardized by averages of the normative population. As a consequence, each individual measure will suffer with this standardization and, more representative results would be obtained if the determination of individual frequency bands was possible. Therefore in this platform it is possible to determine the new individual boundaries for each frequency band by determining the individual alpha frequency band (IAF) and the peak alpha frequency (PAF). The PAF reflects the dominant or most frequent oscillation in the alpha band and it s a necessary value to adjust this frequency band between individuals. Activity in the alpha and theta band respond in different and opposite ways, when one synchronizes usually the other desynchronizes [2]. With increasing task demands, theta synchronizes while alpha desynchronizes the same way alpha synchronizes and theta desynchronizes when closing the eyes. This way, by plotting the EEG spectrum of a recording during a demanding task against a recording during a resting 38

39 period it is possible to identify the boundaries of the individual alpha band as well as the PAF which is the frequency with the highest amplitude inside these boundaries. Another simple way to get both previous results is by plotting the spectrum of a recording where the subject has his eyes closed (alpha synchronizes and theta desynchronizes) against a recording with the subject having his eyes opened (alpha desynchronizes and theta synchronizes). See Figure 4 for an example. Figure 4. Individual alpha bands: IAFL lower alpha; IAFU upper alpha; PAF peak alpha frequency. The calculation of the boundaries from the other frequency bands is based on W. Klimesch method [2] where fixed length bands are applied before and after the IAF band or by the same method of labeling events that induce increases or decreases in specific frequency band amplitude. For example, the individual sensorimotor rhythm (SMR) could be calculated by plotting the EEG spectrum during an event where the subject is asked to maintain motionless (SMR increases) against the EEG spectrum during an event where the subject is allowed to move his limbs if he wishes (SMR decreases) [1]. The individualization of the frequency bands should not only be done between individuals but also between different recording sites as EEG frequencies vary between them [2]. Feedback Display The display was created with the aim of producing a simple visual feedback but at the same time an immersive environment should be created to minimize undesired distractions. For this effect, the Microsoft DirectX TM library was used to draw tridimensional objects that would respond to the value of the calculated feedback parameter. The display contains two objects; a sphere and a cube (see Figures 5 and 6). The sphere is where the feedback parameter is reflected. Its value is directly reflected into the spheres radius and if it reaches the threshold (Goal 1) the sphere color changes. This sphere is constituted by several slices and the more slices it has, more smooth it looks. Initially, the sphere is only constituted by four slices, which is the minimum number possible, and while Goal 1 is being achieved slices are slowly added to the sphere. When Goal 1 is not achieved, the sphere loses slices slowly until it only has four of them again. The cubes height is where Goal 2 is reflected, making it rise until Goal 1 is no longer being achieved. Then the cube starts falling slowly until it reaches the bottom or Goal 2 is achieved again. So, the best outcome would be having the cube as high as possible. 39

40 procedure were interpreted to them. Participants were pseudo-randomly allocated to training (20 subjects) and non-neurofeedback control groups, controlling for factors such as age and gender. The protocol was approved by the Research Ethics Committee (University of Macau). Figure 5. Display when neither Goal 1 nor Goal 2 is achieved. EEG recordings During the experiment, the participants sat in a quiet room. training was done on channel Cz-M12 (M12 is the average of M1 and M2) of the International System of Electrode Placement with sampling frequency 256 Hz, the ground was located at forehead. The signals were amplified by a 24-channel system (Vertex 823 from Meditron Electomedicina Lda, SP, Brazil) and were recorded by Somnium software platform (Cognitron, SP, Brazil). Circuit impedance was kept below 10kΩ for all electrodes at all times. Figure. 6. Display when Goal 1 and Goal 2 are being achieved. Materials and Methods This study is published in [9] and focuses in memory improvement associated with this training. Here we analyze the results concerning EEG control. Participants A total of 32 students (22 males and 10 females, aged years: mean=23.28, SD=3.11) took part in the experiment. Informed written consent was obtained from all participants after the experimental nature and Results EEG Control over Training Sessions The average relative amplitudes across all participants are shown in Figure 7. The relative amplitudes for IAF, IAFL, IAFU, IAF/theta, sigma and alpha showed an increasing trend with the training sessions while theta and delta decreased over the sessions. The increase in IAF/theta ratios resulted from an increase in IAF amplitude and simultaneous decrease in theta amplitude. Furthermore, the relative amplitudes of the different bands had linear correlation over the sessions. The R square was between 0.78 and 0.91 for the different alpha bands. This results show the increase in control over this specific frequency bands as the number of sessions advance.. 40

41 Figure 7. The average relative amplitudes across all participants over sessions, the straight line results from a linear regression and indicate a linear long-term change. Mental strategies In an attempt to help participants find out efficient strategies for self-regulating their EEG, after each training session, they were asked to write down which mental exercises or thoughts were used. Their effectiveness is represented in Figure 8. Among them, the most efficient thoughts or strategies were friends (1.625), love (1.4) and family (1.1); the worst were anger (-2.0) and calculation (-0.15). The effect of some positive strategies sub types like love (lover (1.67)), nature (hometown (1.5)) and family (brothers (2.0)) stood out. Figure 8. Strategy types scores taken from [9]. 41

42 Response time. Response time is the amount of time required to voluntarily produce a significant change in the EEG activity. We measure it as the amount of time since the beginning of the trial and the accomplishment of Goal 2. As Goal 2 is only achieved after 2 seconds of uninterrupted achievement of Goal 1, 2 seconds is the minimum possible response time. Therefore, in the following analysis we must always sum this additional 2 seconds to the results. From the 20 subjects that participated in the training, 14 were able to decrease their average response times. Figure 9 shows the distribution of these delays for a successful subject and Figure 10 depicts the distribution of response times in all trials and all subjects in order to show the orders of magnitude we are dealing with (mean of 6 seconds). Figure 10. Distribution of response times across subjects and sessions. Figure 9. Distribution of response times until the achievement of Goal 2 from a successful subject. We can notice that the delays of 0 milliseconds are very common and, in the subject from Figure 9, they start apearing in more advanced trials. This occurs because the subjects start changing the EEG in advance during the trial intervals and, as the trial starts, after the 2 seconds, Goal 2 is achieved. 42

43 Example of a BCI Game Training: Before actual gameplay the gamer can train his/her EEG in the similarly to our platform. Training should be short but effective. EEG in Gameplay: The gamer should perceive the EEG characteristic being monitored by some effect that does not affect gameplay (color or sound). Achieving Goal 1 also does not affect gameplay but must be noticed. Goal 2 (lasting achievement of Goal 1) produces a gameplay change. Goal 2 Gameplay change: This gameplay change should affect a single player game (or multiplayer if all players have BCI) but the game must not depend on it so it is also attractive for people without BCI. It can potentiate some ability that is triggered with the conventional controllers. Example: A spell is stronger or has an additional effect; a jump is higher; accuracy of a shot increases; running faster; slowing time Discussion Using this platform, it was possible to observe voluntary control over a specific EEG rhythm, the IAF band, and its increase with training. Also, in some cases, besides producing more EEG activity, the time required to produce this response also decreases although this effect is not stable and may depend on several unmeasured factors. Nevertheless, these values are centered in 6 seconds, a value that does not seem adequate for situations that require fast responses. Some projects have been using hardware like OCZ s NIA TM or Emotiv as additional remote controllers to interact with games that require precise or fast responses (for example, PONG or changing weapons with NIA TM or a recently funded Kickstarter project Throw Trucks With Your Mind! ). Moreover, these games tend to ignore existing controllers and leave all the effort for the BCI what can be very frustrating for someone who does not have the chance to train with proper feedback or simply, has difficulty in controlling his EEG. This, not only shortens the number of potential consumers of these interfaces but may even make them unpopular and undesired. In light of our results and our knowledge of EEG-Self regulation we can imagine BCI being used as promoted by Emotiv, to read ones emotions and translate them in the gamming environment or, to use this input as an additional experience. For example, while playing a game with a keyboard and mouse or a gamepad, the BCI (if successful) input can trigger some special effect of and ability called with the controller (ex: jump higher, stronger spell, more accuracy, etc ). This might not be so frustrating when not achieving, as the game is still playable, and can be immensely gratifying when achieved. This can be easily implemented in a game and, the game can still be enjoyable for gamers that do not possess a BCI device. As a first step for BCI in gaming, we believe this is a safe and very amusing one. Acknowledgements This work is supported in part by FCT SFRH/BSAB/1101/2010 and PEst-OE/EEI/LA0009/2011 grants and the Macau Science and Technology Development Fund under Grant FDCT 036/2009/A and the University of Macau Research Committee under Grant MYRG139(Y1-L2)-FST11-WF. References [1] Sterman, M. B. and Egner, T., Foundation and practice of neurofeedback for the treatment of epilepsy, Applied Psychophysiology and Biofeedback, 31, 1 ( [2] Klimesch, W., EEG alpha and theta oscillations reflect cognitive and memory performance: a review and analysis, Brain Research Reviews, 29, 3 (1999), [3] Schalk, G., McFarland, D. J., Hinterberger, T., Birbaumer, N. and Wolpaw, J. R., BCI2000: A generalpurpose, brain-computer interface (BCI) system, Ieee Transactions on Biomedical Engineering, 51, 6 (2004), [4] Martinez, P., Bakardjian, H., Cichocki, A., Multi- Command Real-Time Brain Machine Interface Using SSVEP: Feasibility Study for Occipital and Forehead Sensor Locations, Springer Netherlands (2008), [5] Birbaumer, N., Elbert, T., Vonbulow, I., Lutzenberger, W. L, Rockstroh, B., Canavan, T., Linden, A. and Wolf, P., Biofeedback of Slow Cortical Potentials in Epilepsy - Preliminary-Results of A Controlled Double-Blind Outcome Study, Biofeedback and Self-Regulation, 13, 1 (1988),

44 [6] Iversen, I. H., Ghanayim, N., Kubler, A., Neumann, N., Birbaumer, N., and Kaiser, J., A brain-computer interface tool to assess cognitive functions in completely paralyzed patients with amyotrophic lateral sclerosis, Clinical Neurophysiology, 119 (2008), 10, [7] Blankertz, B., Dornhege, G., Krauledat, M., Müller, K., Curio, G., The non-invasive Berlin Brain Computer Interface: Fast acquisition of effective performance in untrained subjects, NeuroImage, 37, 2 (2007), [8] Rodrigues, J., Migotina, D., da Rosa, A., EEG training platform: Improving Brain-Computer Interaction and cognitive skills, Human System Interactions (HSI), 3rd Conference on HIS (2010), [9] Nan, W, Rodrigues, J., Ma, J., Qu, X., Wan, F., Mak, PI., Mak, PU., Vai, MI., da Rosa, A., Individual alpha neurofeedback training effect on short term memory, International Journal of Psychophysiology, 86, 1 (2012),

45 The Video Game and Player in a Gameplay Experience Model Proposal 1,2 Samuel Almeida Licínio Roque 1 CETAC.MEDIA 2 CISUC Dep. Communication & Art Dep. Informatics Engineering University of Aveiro Polo II Campus Universitário University of Coimbra de Santiago, , Coimbra, Portugal Aveiro, Portugal lir@dei.uc.pt sja@ua.pt Ana Veloso Óscar Mealha CETAC.MEDIA CETAC.MEDIA Dep. Communication & Art Dep. Communication & Art University of Aveiro University of Aveiro Campus Universitário Campus Universitário de Santiago, de Santiago, Aveiro, Portugal Aveiro, Portugal aiv@ua.pt oem@ua.pt Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Abstract The video game medium, embedded within one of the most lucrative industries today, is developed to entertain and create satisfying experiences for the game player. Extensive work has been developed on these experiences, exploring concepts such as immersion or flow, or centering on specific experiencerelated models. However, we consider that these do not fully portray the nature of the gameplay experience a dynamic interplay between a video game and the player. This work summarizes a Gameplay Experience Model proposal centered on the dynamic interaction that exists during video game play. We describe the development of the proposed model, centered on a literature review process and complemented with two focus group sessions where the gameplay experience and its characteristics were discussed. Posteriorly, the conceptual model is explored in terms of its various elements and dimensions, in addition to its applicability in game contexts. Author Keywords Gameplay Experience, Video Game, Player, Model ACM Classification Keywords K.8.0 [Personal Computing]: General Games 45

46 Introduction Similar to many other media industries, the success with video games relates to the resulting experience from interacting with the game. The discussion of these experiences is a widely debated topic, and has been contemplated under the names of gaming, player or gameplay experience, and how these are formed and how they can be measured. Widely used concepts to describe game experiences are immersion, flow and presence, for example. Moreover, other studies [1 3] have looked beyond these concepts and built specific models that characterize the nature of the gameplay experience. Our interpretation of the gameplay experience suggests that it is the result of a dynamic interplay between the video game and the player. A video game is developed based on a group of characteristics that play a role in the player s consumption of the media. Furthermore, the player contributes to the creation of his experience by bringing to game playing his personal motivations, expectations regarding the game and developed abilities that influence the overall experience. We consider that there lacks a model which equally balances and characterizes the dynamic interaction described above. The work presented here seeks to fill this identified gap, by presenting a conceptual model to characterize our interpretation of the gameplay experience a twofold experience, where it is both the process and the outcome. Here, we explore in some detail the development of the proposed model, built upon a literature review process which considered mainly characteristics related to immersion and flow; and posteriorly complemented with information gathered through two focus groups. Related Work and need for a new Model Multiple studies [1 7] have reflected on the global concept of the user experience in video games, appropriating concepts such as player, gaming or gameplay experience. These studies focus on the essence of these experiences, how they are formed and how they can be measured. Takatalo et al. [7] suggest the user experience is a term that includes concepts such as immersion, presence, fun, involvement, engagement and flow. Briefly looking into some of the ideas explored in these studies; Brown & Cairns [8] divide immersion into three levels: engagement, engrossment and total immersion; McMahan [9] explores immersion as based on three conditions: users expectations, users actions and conventions; Ermi & Mäyrä [1] present a gameplay experience model focused on immersion divided into three components: sensory, challenged-based and imaginative immersion. Considering flow, a state in which individuals are so involved in an activity that nothing else seems to matter [10], work on the concept and the optimal experience are based on the ideas introduced by [10]. Various flow related studies [11 13] are based on the basic ideas of the concept and further explore it in terms of its development and measurement [13] or game related contexts [11,12]. However, the gameplay experience is more than these concepts, individually or considered simultaneously. Other studies [1 3] have focused on presenting models that characterize and represent the gameplay experience process. The SCI Model (Sensory, 46

47 Challenge-based and Imaginative Immersion) [1] focuses on immersion and its diverse manifestations in different games and with different players. The model details player-related characteristics involved in the forming of the experience. However, it lacks clarity on how these characteristics influence the experience as well as how referred game structures can influence a player s attitude towards a game. The CEGE Model (Core Elements of the Gaming Experience) [2] reflects on the necessary conditions to provide a positive experience while playing video games, presenting the core elements of the interaction process that build the experience. The basis of the CEGE is the game and the interaction between it and the user, which they call puppetry. Video game is a guiding element for Gameplay and Environment, while Puppetry is a guiding element for Control, Ownership and Facilitators. Fernandez [3] contributed with a Game Experience Model, defending a framework that further clarifies the relationship among game components. The model suggests the experience is built upon three moments (before, during, and after the experience) where model elements act upon and influence other elements, and that fun is the result of the experience. Despite the important contributions of each of these studies, we consider they do not fully grasp the extent of the gameplay experience and our interpretation of its key actors: the video game, the player and the multiple relations that result from their interplay, working towards creating ideally a satisfying gameplay experience. Gameplay Experience Model Development The development of the proposed model followed and is grounded on a two stage process: an initial literature review followed by two focus group sessions. Initial Gameplay Experience Premise The initial premise of our model considers the gameplay experience as a result of the interaction process between a video game and the individual that plays the medium the player [1]. These two elements are the cornerstones of the model, from which the rest of the model is constructed. In order to further characterize each of these elements, supporting model characteristics were collected from literature review and focus groups. Literature Review The first stage of development was based on the analysis and collection of gameplay experience related characteristics from a literature review process. Various authors [1,8 14] were considered, focusing primarily on the concepts of immersion and flow, two widely used concepts in describing the gameplay experience. Each of these studies is also built on a specific development process, including literature reviews, grounded theory method, observation and interviews. Looking initially at Immersion, McMahan [9] defines three conditions that create a sense of immersion in a virtual environment or a 3D computer game, summarized in the concepts of (i) user expectations, (ii) user actions, (iii) consistency. Brown & Cairns [8] affirm that immersion can be divided into three levels engagement, engrossment and total immersion where certain barriers must be overcome in order to move to the next level. The barriers to each level 47

48 include time, effort and attention for engagement; visuals, tasks, plot for engrossment; empathy and atmosphere for total immersion. Ermi & Mäyrä [1] structure the gameplay experience according to three dimensions: sensory immersion, with relevance to the audio and video aspect of the game; challenge-based immersion, referring to the balance of challenges and abilities; and imaginative immersion, related to players becoming absorbed with characters, story and the game world. Considering Flow, existing work embraces the original ideas presented by Csíkszentmihályi [10]. Flow the optimal experience considers eight factors [10]: (i) a challenging activity requiring skill; (ii) a merging of action and awareness; (iii) clear goals; (iv) direct, immediate feedback; (v) concentration on the task at hand; (vi) a sense of control; (vii) a loss of selfconsciousness; (viii) an altered sense of time. Additional studies [11 14] have picked up on these eight factors and adapted them to their own needs, while remaining core characteristics that can promote a state of flow. From this initial process, a preliminary gameplay experience model was developed based on the multiple characteristics identified. The model distributed the highlighted characteristics among two vertical axis player and video game and three additional horizontal partitions immersion, flow or immersion + flow. Focus Groups To further explore the gameplay experience and develop on the preliminary model, a second moment of work was carried out. The second stage of development consisted in focus group sessions. Two focus group sessions were held during October 2012, at two different Portuguese Universities. Both focus groups gathered a heterogeneous set of individuals with different game-related activities (e.g. some were enthusiastic game players; others frequently play videos games, but also actively develop work and research game-related topics). One group consisted in 10 participants (seven men, three women); the other group in six participants (four men, two women). Sessions lasted on average two hours. The focus group sessions were divided into four parts. Part one (i) consisted in an introduction of the objectives of the focus group. Participants were introduced to the topic of discussion video games and the gameplay experience. Participants were asked to reflect on game and player-related characteristics that can contribute or define the gameplay experience. Part two (ii) served to discuss the characteristics participants indicated related to video games. Part three (iii) served to discuss the characteristics participants indicated related to the player. Part four (iv) served to discuss possible relationships from the characteristics gathered in parts II and II, as well as a possible categorization of these characteristics. From each focus group, characteristics related to Video Games and Players were isolated for posterior analysis. More than 100 different characteristics were identified in both focus groups as pertaining to the gameplay experience, either related to the Video Game or the Player. Concerning the video game element, from this group of characteristics, interface, actions, learning and art were referred multiple times. Interface, for example, was 48

49 described as the ease of use, as the technological support, or as the means through which interaction is established with the game. The experience of the game should not be the experience of dealing with the interface and dealing with the problems of the interface, instead of playing the game. Balance, one of the few characteristics mentioned in both sessions, was presented by one participant as: [There should be] balance between the interface and the interaction that is proposed to the player if the interface is really good but without adequate interaction, there is no balance. On the predictability of the game, one participant explained: When you already know everything that is going to happen, the experience is lost. That is why in a game where something new happens even if small the pleasure is maintained. In addition to balance, coherence, sociability and expectations were the other three characteristics mentioned in both groups. Looking at the Player element, multiple characteristics were also identified. From this group of characteristics, motivation, expectation and interpretation were referred on more than on one occasion. Player motivations were described as important in order to become engaged with the game. Expectations, for example, were described as: It essentially is related to life experience. Related to a game, what is expected from a game, what we get from playing the game is highly conditioned by what we ve played before. During both focus group sessions, an exercise of characteristic organization was proposed in order to group similar characteristics into a possible single representative characteristic. While this stage of the focus groups did not yield expected single representative characteristics, multiple relationships between the collected game and player related characteristics were established. Regarding video games, balance, coherence and consistency were always grouped together given their similarities; interface and feedback were grouped on one occasion, as was interface and sound in another relationship. Objectives and mechanics were paired in one relationship; rules and rewards in another. As for the player component, skills and background were grouped in one relationship; experience and skills in another relationship. Multiple relationships were identified for the various collected characteristics, demonstrating a significant versatility of the concepts. Given the nature of the participants of the focus groups, this process proved to be important in uncovering various characteristics not identified through literature review. The active involvement with video games from these individuals provided an important contribution to the development of the model. 49

50 Basic Model organization Posterior to the development stage, the model was organized according to the multiple data sources used. From the literature review and complimented with ideas from the focus groups each of the two main model elements (video game and player) were divided into several dimensions. The video game element of the model was divided into three representative dimensions Mechanics, Interface and Narrative based on the interpretation of [15], and supported by an additional dimension consistency which bonds the previous three dimensions. The player element is also divided into three dimensions Motivations, Expectations and Background based on the ideas of [1]. A Renewed Gameplay Experience Model The gameplay experience model explored here is a conceptual framework for understanding the various elements and characteristics that can play a role in the gameplay experience. The model considers the gameplay experience as a result from of the interaction of a player with a video game during game play. The model proposed here defines the gameplay experience as a twofold experience it is both an interactive experience and an emotional experience. Borrowing the ideas of Dewey (1938) as presented in [2], this model considers the gameplay experience to be both the process and the outcome: it is an interactive experience (the process of playing the game) and an emotional experience (the outcome of playing) an emotion (or a group of emotions) that results from playing. During the act of video game play, these experiences can influence one another and are shaped by the multiple characteristics of the model. The chemistry of the interactive and emotional experience defines the product of the gameplay experience. The interactive experience is the manner in which a player operates and approaches a game. It is how the player explores the game space or level; how he or she interacts with other players, non-playable characters or objects; and how the player makes decisions. This process is framed and limited by the game itself, but influenced by the player s background, motivations and expectations. The interactive experience is also influenced by a player s current emotional experience. While playing, a player can also be characterized by a current emotional experience that may possibly vary throughout the game, influenced by the game s directives or the outcomes of a player s actions as he or she progresses. The interactive experience is frequently being influenced by a player s current emotional state: if a player is anxious, he may be less capable of paying attention which reflects on his ability to play; if he is relaxed and stress-free, he may be in a state of flow where everything runs perfectly. This process will affect the gameplay result, which refers to the visible consequences of the game. These consequences can then influence the emotional gameplay experience such that when positive, they can produce positive emotions within the player (e.g. satisfaction, enjoyment, excitement); if negative, they can produce the opposite (anger, despair, indifference). It is also possible that positive or negative consequences alter a player s approach (interactive experience) by changing player motivations (motivations) and actions (expectations), for example. This bi-directional relationship can explain why occasionally players can feel enjoyment and frustration throughout the duration of the game. 50

51 Figure 1 represents a holistic view of the proposed gameplay experience. possible that differences in the games may condition the player s experience. The ambient setting can also limit the experience such that the place where a player engages in playing will limit the possible platforms that can be used to play. The Video Game element of the model incorporates four dimensions: (i) Mechanics, (ii) Interface and (iii) Narrative, each with one or more core characteristics, and supported by an additional (iv) Consistency dimension. Figure 1. Representation of the proposed Gameplay Experience Model Model Elements, Dimensions and Characteristics In addition to its two main elements (video game and player), the model considers a Gameplay Situation, a global setting in and through which the act of play takes place. It is exterior to both the player and the game, and is defined by an ambient setting and a platform setting. The ambient setting can be related to the time of day or the place where the game is played; the platform setting refers to the game platform used to play (e.g. console, PC, mobile device). Considering some games can be played across platforms, it is The Mechanics of the game are its rules of organization and operation [15]; the mechanics define what a player can and should do in the game to progress through it; they are the game s rewards and supplements to the player as a result of his actions in the game. Therefore, the Mechanics include Goals, Rules and Rewards. Goals (i.e. objectives, tasks and challenges) are what a player must do and accomplish in order to progress (or win) in the game. Rules are an imperative that overlook and limit the interactions occurring within the game and the outcome of these interactions [16]. Rewards which can differ according to the game and the situation are something a player receives in return for completing goals or specific objectives, tasks and challenges. The Interface is the look and feel of the game. It is what the player sees and hears, but also how he interacts with the game and is informed on his actions. The interface includes Visuals, Audio, Input and Feedback. Visuals are related to how the game looks, in two or three dimensions; it can also be the Heads-up display (HUD), where much of the game information is displayed (e.g. remaining time, quantity of life, game 51

52 resources). Audio is the game s sounds, sound effects and music, equally important when compared to visuals in creating atmosphere and for player feedback [15]. Input relates to how a player physically interacts with a game through technological support (e.g. keyboard and/or mouse, a joystick, a gamepad, physical movements that are captured, or others. Feedback is the game s response (visual or audio) to players actions. Narrative refers to the part of the story that is told to the player. All video games can tell a story, even if the story is present in the individual s imagination and created while he plays. Consistency is transversal to the other three dimensions (mechanics, interface and narrative) and refers to the harmony and balance established between the three dimensions. The consistency of a game is what makes it a solid and enjoyable product, rather than a sum of various parts that do not relate. The Player element of the model incorporates three dimensions: (i) Motivations, (ii) Background and (iii) Expectations, each with one or more core characteristics. The Motivations of a player refer to the necessary conditions that must be met for a player to become motivated to play, or to continue playing. Motivations include Motivation, Attention, Effort, Time and Connection. Motivation is the single or multiple incentives for a player to initially play a game and posteriorly, can influence how the gamer plays. Attention (or concentration and focus) is a state where players place all their cognitive and/or physical effort on a specific goal, objective or challenge. Effort refers to the physical or mental investment and energy a player puts towards the game or learning to play. Time refers to the quantity of time a player is willing to invest in playing a game, which is more or less according to the resulting motivation to play. Connection refers to the players possible emotional engagement with a game, characters or other elements. Considering the uniqueness of each player, the motivations and reasons for a player becoming connected may be different and have distinct origins. The Background of a player is related to his or her personal history, related or not to video games, but which can be applied while playing. Background includes Preferences, Ability and Knowledge. Preferences are associated to the aspects of a game a player enjoys the most, related to game platforms, genres, visuals or others; and are normally dependent of players past experiences with games. Abilities are a player s collection of learned skills motor, cognitive or perceptual which are applied while playing a specific game. Knowledge is a repertoire of information that may or not be collected from playing games, but that can be applied during game play. The Expectations of a player refer to what a player anticipates finding within the game, related to game mechanics, interface or narrative, based on his background with similar games or other sources of information. The Expectations dimension includes Expectations, Action and Control. Expectations refer to the collection of things a player anticipates and hopes to find in a game. However, while a player may expect something from the game, his experience can still be satisfying if what he finds and did not expect is 52

53 enjoyable. Actions are what the player can do in the game and should, consequently, have some impact in the game and how the player and game progresses [9]. Control is related to the freedom the game gives the player to manipulate his game characters or playable objects, as well as the possibility of a player being in a situation where there are no preoccupations of the outcomes of one or more actions. The Model in Context The developed model, in addition to its utility as a framework for better understanding the multiple characteristics related to a game and a player that can contribute to the gameplay experience, is accompanied by a gameplay questionnaire specifically developed to assess the presence of the multiple model characteristics in a gameplay context. A 27-item questionnaire assesses player s opinions on the presence of model characteristics related to the game and their experience. From the questionnaire results, an analysis can be steered in the desired direction, focusing on the desired model dimensions or specific characteristics. Furthermore, the model also considers the interactive experience which can be assessed according to game metric analysis (where possible, if the game supports metric extraction). The simultaneous analysis of questionnaire results with game metrics can provide insight regarding a players or group of players gameplay experience. Final considerations In the absence of a gameplay experience model that equally balances player and video game, we propose our interpretation of the referred experience. We introduce a conceptual model proposal that equally highlights the dynamic interplay that occurs between these two elements during the act of game playing. We also explore each of these elements dimensions and supporting characteristics, which together shape the interactive and emotional experience we consider to be part of the experience. In this work we present the process behind the development of the model, supported on a literature review and two focus groups where the gameplay experience was discussed. The work and model discussed here presents a comprehensive look at the gameplay experience and can further be considered a tool in the analysis of the experience in a defined game playing context. Acknowledgements Special thanks to those that participated in the focus groups. Our acknowledgments to the Portuguese institution Fundação para a Ciência e Tecnologia (FCT) for funding this project with the PhD grant no. SFRH/BD/66527/2009. References [1] L. Ermi, F. Mäyrä. Fundamental Components of the Gameplay Experience: Analysing Immersion, in: DiGRA (Ed.), Vancouver, Canada, [2] E.H.C. Gámez, P. Cairns, A.L. Cox. Assessing the Core Elements of the Gaming Experience, (2010) 288. [3] A. Fernandez. Fun Experience with Digital Games: a Model Proposition, in: O. Leino, H. Wirman, A. Fernandez (Eds.), Extending Experiences: Structure, Analysis and Design of Computer Game Player Experience, Lapland University Press, Rovaniemi, 2008: pp [4] E.H.C. Gámez, P. Cairns, A.L. Cox. From the gaming experience to the wider user experience, (2009)

54 [5] L. Nacke, A. Drachen, K. Kuikkaniemi, J. Niesenhaus, H.J. Korhonen, W.M. van den Hoogen, et al. Playability and Player Experience Research, in: Breaking New Ground: Innovation in Games, Play, Practice and Theory, DiGRA, [6] R. Bernhaupt. User Experience Evaluation in Entertainment, First, Springer, [7] J. Takatalo, J. Häkkinen, J. Kaistinen, G. Nyman. Presence, Involvement, and Flow in Digital Games, in: R. Bernhaupt (Ed.), First, Springer, 2010: p [8] E. Brown, P. Cairns. A grounded investigation of game immersion, in: ACM, Vienna, Austria, 2004: pp [9] A. McMahan. Immersion, Engagement, and Presence: A Method for Analyzing 3-D Video Games, in: M.J.P. Wolf, B. Perron (Eds.), Routledge, New York, NY, 2003: pp [10] M. Csíkszentmihályi. Flow: The Psychology of Optimal Experience, Harper Perennial, 1990.[11] J. Chen. Flow in games (and everything else), Commun. ACM. 50 (2007) [12] P. Sweetser, P. Wyeth. GameFlow: a model for evaluating player enjoyment in games, Comput. Entertain. 3 (2005) 3. [13] J. Nakamura, M. Csíkszentmihályi. The Concept of Flow, in: C.R. Snyder, S.J. Lopez (Eds.), Handbook of Positive Psychology, Oxford University Press, New York, NY, 2002: pp [14] C. Bateman, R. Boon. 21st Century Game Design, Charles River Media, Hingham, MA, [15] A. Rollings, E. Adams. Andrew Rollings and Ernest Adams on Game Design, New Riders Games, [16] S. Egenfeldt-Nielson, J.H. Smith, S.P. Tosca. Understanding Video Games: The Essential Introduction, Routledge,

55 Uma Proposta de Arquitectura para Composição Dinâmica de Soundscapes em Videojogos Durval Pires Department of Informatics Engineering University of Coimbra Coimbra, Portugal Valter Alves Polytechnic Institute of Viseu, Portugal & Centre for Informatics and Systems of the University of Coimbra Coimbra, Portugal Licinio Roque Department of Informatics Engineering University of Coimbra Coimbra, Portugal Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Resumo A imprevisibilidade associada à interactividade presente no meio videojogo, faz com que a dificuldade de fazer sound design para este contexto seja extremamente elevada. Como consequência, é possível encontrar situações em que o contexto de jogo e a soundscape que é ouvida pelo jogador não se complementam. Neste trabalho é apresentada uma arquitectura, informada pela teoria de Acoustic Ecology, que permite testar uma nova abordagem ao problema da composição dinâmica de soundscapes em videojogos. O principal elemento do sistema apresentado é um módulo de composição dinâmica que executa em run-time. Também como parte da solução proposta, é fornecida uma API que permite que os designers especifiquem elementos e contextos sonoros. Durante a execução do videojogo, o módulo monitoriza a soundscape e aplica algumas técnicas de composição, tendo em conta as especificações dos designers, definidas por via da API. É apresentado um exercício que consistiu na reimplementação de um jogo usando a arquitetura aqui proposta. Este exercício confirmou a exequibilidade da solução, reforçando a ideia de que as técnicas 55

56 desenvolvidas podem realmente ser úteis nas diversas situações de jogo. Author Keywords Acoustic ecology, dynamic soundscape composition, game audio, game design, middleware, healthy soundscape, sound design, sound engine, soundscape. Composição dinâmica de soundscapes, desenho de videojogos, desenho de som, Ecologia Acústica, som em videojogos, middleware, motor de som, soundscape, soundscape saudável. ACM Classification Keywords H.5.2 [Information Interfaces And Presentation (e.g., HCI)]: User Interfaces auditory (non-speech) feedback, evaluation/methodology, standardization, styles guides, theory and methods; H.5.1 [Information Interfaces And Presentation (e.g., HCI)]: Multimedia Information System audio input/output, evaluation/methodology; H.5.5 [Information Interfaces and Presentation (e.g., HCI)]: Sound and Music Computing methodologies and techniques; K.8.0 [Computing Milieux]: Personal Computing General Games. General Terms Design, Standardization, Languages, Theory. Introdução O sound design aplicado aos videojogos enfrenta o enorme desafio de lidar com a natureza dinâmica do meio [7, 8, 16]. Os jogadores, ao executarem variadas acções sobre diferentes entidades presentes no jogo, tornam-se parte integrante da composição da soundscape. Além disso, como cada uma dessas entidades pode ter variadas expressões acústicas, cada sessão de jogo poder ter uma composição diferente das restantes. Esta imprevisibilidade, se não for devidamente abordada, pode resultar em composições que comprometem as intenções do designer, quer em termos estéticos quer em termos semânticos. A indústria tem vindo a procurar soluções para este problema, embora maioritariamente recorrendo a ferramentas e técnicas originárias de meios lineares [7]. Ainda assim, existem ferramentas de Audio Middleware que oferecem funcionalidades que tentam alterar essa tendência, [3, 4, 5] e oferecer novas capacidades. Porém, o seu custo, quer monetário quer em termos do know-how necessário para as utilizar, faz com que estas ferramentas não sejam viáveis para todos os criadores de jogos [10]. Por outro lado, mesmo com estas ferramentas, é muito difícil para os sound designers conseguirem prever, e preparar, todas as situações passíveis de ocorrer no cenário de jogo. Um conceito que pode ser valioso na busca de novas abordagens ao problema é o conceito de soundscape, originário da teoria de Acoustic Ecology [17] e introduzido por Schaffer [17] e aprofundado por Truax, nomeadamente em termos da sua composição [18]. Este conceito enquadra o som num ambiente que deve ser tratado como um todo, em que todas as relações se influenciam mutuamente. O equilíbrio funcional entre a variedade e a complexidade deste sistema de relações entre diferentes entidades e o ambiente, é fulcral para a manutenção de uma soundscape saudável. Adicionalmente, este tipo de soundscapes deve permitir que os sons presentes nela consigam transmitir de forma clara a semântica que a eles está associada. 56

57 Layer Ambiance Dialogue Music Foley SFX Intenção Sons do ambiente onde o jogador se encontra. Oferece noção de lugar. Qualquer forma de discurso presente no jogo. Oferece vários tipos de informação. Qualquer forma de composição musical presente no jogo. Ajuda a definir o tom emocional de cada momento. Sons reais que caracterizam uma entidade ou evento. Sons imaginários, que são colocados na cena para ajudar a realçar a expressão sonora de alguma entidade ou evento. Tabela 1 - Sound Layers propostos por Nick Peck [15] O conceito de soundscape já foi utilizado em contexto de videojogos, nomeadamente para efeitos de definição de ambientes acústicos [6]. Porém, a definição de soundscape defendida por Schaffer e Truax argumenta que o contexto no qual um som é ouvido é parte capital da sua percepção [18]. O mesmo som, em diferentes contextos, pode assumir diferentes significados. Transferindo este mindset para sound design para videojogos, os diferentes contextos oferecidos durante a jogabilidade, exigem diferentes abordagens de sound design [1,2]. Sistema de Composição Dinâmica de Soundscapes Embora reconheçamos a dificuldade inerente a esta tarefa, acreditamos ser valioso desenvolver formas de permitir que a composição da soundscape tenha em conta tanto as intenções do designer, como as acções executadas pelo jogador [20]. Também se torna importante que as possíveis soluções encontradas estejam ao alcance de todos os criadores de jogos, mesmo os que à partida detenham menos conhecimentos em termos de sound design. Além disso, em acréscimo às suas potencialidades, o facto de novas abordagens serem dadas a conhecer pode ajudar a sensibilizar os criadores de jogos para o potencial que o som pode ter no processo de game design. Como tal, nesta secção apresentamos uma proposta de um módulo de composição dinâmica que faz uso de uma abordagem holística, inspirada nos conceitos de Acoustic Ecology. A base desta solução consiste num sistema que modera a soundscape de acordo com eventos que recebe do jogo. Ao receber os eventos, as características dos elementos sonoros presentes na soundscape são analisadas, sendo de seguida aplicadas heurísticas de acordo com essa análise. Este sistema procura oferecer um conjunto de técnicas de composição que mantenham a soundscape saudável. A caracterização prévia dos elementos sonoros é feita através de uma API também desenvolvida de raiz para esta solução. Esta API procura oferecer acessibilidade e permitir uma abordagem que seja mais próxima, em termos do discurso, de pessoas do domínio do sound design. Nesse sentido, os elementos sonoros, ao serem criados, podem ser classificados de acordo com o seu nível semântico (aqui designado de Layer) [15]. Existem 5 Layers diferentes: Ambiance, Dialogue, Music, Foley e Sound Effects (SFX) (ver Tabela 1). De forma semelhante, as fontes também podem ser associadas a personagens do jogo (Agents), e também podem ser associadas a tipos de explorações de som (Patterns) [1,2]. Na Figura 1 é apresentada a arquitectura geral do sistema proposto. A caixa no canto superior esquerdo representa todo o código que é específico de um projecto de jogo. É aí que está a lógica de jogo, por exemplo. A caixa no canto inferior esquerdo representa a API criada, com as classes e métodos que implementam os recursos oferecidos (por exemplo, o código que cria uma fonte de som ou um contexto). A ideia é que os programadores chamem esse código a partir da lógica de jogo. O módulo de composição dinâmica é representado pela caixa à direita (DSCM). A comunicação entre o código da API e o módulo de composição é assegurado por mensagens OSC (Open Sound Control). OSC é um protocolo standard para comunicação por mensagens com um formato semelhante a um URL. Este protocolo 57

58 ! recursos necessários para o processo de composição (criação, eliminação e edição de elementos). Este tipo de tarefas pode exigir cooperação entre o Maintainer e o Run-Time Player. Os recursos geridos pelo Maintainer são armazenados nas estruturas que ele controla: Sources (contém Sources representação de cada fonte sonora da soundscape), Contexts (ver Heurísticas de Composição), Listener (estrutura que mantém informação sobre a posição do ouvinte no mundo de jogo) e Patterns (tipos de exploração de som, que podem ser associados a cada Source, que o módulo de composição tem em conta no momento de decidir como actuar sobre a soundscape). Estes objetos são criados quando o designer os inicializa através da API, no código de jogo, sendo mantidos em espera, prontos a serem usados sempre que for exigido. Figura! 1 - Arquitectura geral do módulo de composição é optimizado para ser utilizado por rede, oferecendo um alto nível de interoperabilidade, precisão, e flexibilidade. Durante a execução do videojogo, o OSC Receiver encaminha as mensagens recebidas (enviadas pelo OSC Sender) para o Handler. O Handler analisa as mensagens recebidas, e encaminha-as, de acordo com o seu propósito, ou para o Scheduler, ou para o Maintainer. O propósito do Maintainer é para fazer todo o trabalho de background; ou seja, preparar todas os O Scheduler coordena todo o processo de composição. Oferece resposta a pedidos de start/stop relacionados com Sources e Patterns, bem como com alterações no contexto activo. Sempre que é solicitado que uma fonte seja tocada, o Scheduler responde de acordo com a aplicação de um conjunto de heurísticas, que por sua vez levam em consideração o Current Context e as fontes actualmente activas no Contextual Score. Esta última é uma estrutura que mantém uma categorização das fontes activas de acordo com a sua semântica. As heurísticas também podem avaliar outros aspectos das fontes sonoras, tais como o agente e a Pattern a elas associados. Conforme o processo de composição vai decorrendo, o Scheduler pode encaminhar ordens para o Run-Time Player, de forma a este renderizar o som de acordo com as heurísticas. No nosso projecto, o Run-Time Player adoptado foi a biblioteca de som FMOD API. 58

59 Figura 1 - Exemplo dos diferentes tipos de contexto É importante referir que, durante o processo de composição, o Scheduler faz uso de três pautas (Scores) diferentes: Score, Contextual Score, e Over Density Score. A estrutura chamada Score armazena todas as fontes disparadas pela lógica do jogo, e que não terminaram ou foram solicitadas a parar, independentemente do Scheduler ter decidido que deveriam ser audíveis ou não. Esta estrutura permite que o Scheduler tenha, em qualquer momento, uma visão completa da soundscape que foi requisitada pelo jogo. Por seu lado, Contextual Score é uma estrutura que armazena apenas as fontes actualmente activas, que fazem parte do Current Context. Na prática, esta é a estrutura que sustenta a soundscape que está realmente a ser ouvida pelo jogador. Finalmente, o Over Density Score é uma estrutura que, como o nome sugere, contém as fontes que, apesar de fazerem parte do Current Context, não podem ser ouvidos para não aumentarem demasiado a densidade da soundscape (número de fontes em simultâneo). Heurísticas de Composição Com o objectivo de lidar com a natureza dinâmica do meio, foram definidas um conjunto de técnicas para guiar o módulo de composição. Estas técnicas consistem numa lista de práticas comuns em sound design para jogos, que acreditamos serem instrumentais no sentido de alcançar alguns princípios de Acoustic Ecology, nomeadamente, uma soundscape saudável. Este conjunto de técnicas, representado em termos computacionais sob a forma de uma lista de heurísticas, poderá ser modificado ou ampliado. O papel destas heurísticas passa por monitorizar as fontes que a lógica de jogo determina que deveriam tocar, e decidir se, e como, essas fontes devem ser tocadas. Estas decisões têm em conta as características dos elementos presentes na soundscape, nomeadamente, a Pattern associada a cada fonte. De uma forma simplificada, podemos dizer que as heurísticas são comportamentos que servem de resposta às Patterns que podem ser associadas às fontes. As heurísticas que implementámos, nesta fase da investigação, são as apresentadas nos parágrafos seguintes. Naturalmente, este conjunto de heurísticas não permite contemplar a complexidade de todas as situações passíveis de ocorrer num contexto de videojogo, mas serve o propósito de testar a abordagem proposta. Context O conceito de contexto serve para distinguir entre sons relevantes num determinado instante e sons não relevantes. Esta heurística serve para dar ênfase aos sons do primeiro grupo, e atenuar ou silenciar os do segundo grupo. Existem 3 tipos de contexto: contexto do tipo Layer, onde as fontes dentro de contexto dependem dos layers seleccionados; contexto do tipo Agent, onde as fontes dentro de contexto são escolhidas de acordo com o agente a elas associado; e contexto do tipo Ad-hoc, onde as fontes dentro de contexto são seleccionadas livremente através do seu nome. Na Figura 2, estão representados os 3 tipos de contexto: Agent, Layer e Ad-hoc. Os círculos são fontes que, segundo a lógica de jogo, devem ser tocadas. As cores dos círculos representam agentes, sendo que cada coluna representa um layer. Ao lado de cada círculo está a representação do volume de cada fonte, estando directamente relacionado com o facto dessa mesma fonte estar dentro de contexto, ou não. Thoughts Esta heurística oferece uma possível representação do efeito voz dentro da cabeça, 59

60 vastamente usado para representar pensamentos de personagens. Silence A noção de silêncio pode ser interpretada e implementada de inúmeras formas. Esta heurística procura oferecer uma possível abordagem, na qual todas as fontes são atenuadas quase na totalidade, exceptuando as fontes de Diálogo e Foley. Awareness Em muitos jogos, os game designers recorrem ao som para chamar a atenção dos jogadores para certos elementos relevantes. Esta heurística oferece esse efeito ao dar relevo à fonte sonora que vai ser tocada durante um período de tempo, voltando a soundscape ao seu estado normal uma vez findado esse intervalo. Dialogue Esta heurística visa dar sempre relevo a fontes sonoras que representem elementos de diálogo, silenciando as fontes de Foley e SFX, embora deixando presentes as fontes de Ambiance e Music (ainda que bastante atenuadas). Footsteps Footsteps são um tipo de Foley bastante importante, pois dá personalidade e distingue as diferentes personagens, além de ser importante em certos aspectos de jogabilidade. Quando a movimentação do jogador não produz feedback sonoro, a sensação de imersão do jogador pode perder-se. Como tal, esta heurística não permite que as fontes que lhe estejam associadas tenham o seu volume atenuado abaixo de um certo valor. Contextual Music Música contextual ajuda a caracterizar diferentes contextos, espaços e outros elementos dos jogos. Analogamente, esta heurística, tal como a anterior, não permite que as fontes associadas a ela tenham o seu volume mais atenuado que um certo valor. Achievement/Failure/No Can Do Achievement, Failure e No Can Do são tipos de explorações de SFX, vastamente usados em jogos [1,2]. Devido à importância semântica que costumam ter, esta heurística modera a soundscape de forma a que eles sejam sempre ouvidos, com o objectivo de não deixar o jogador perder informação importante para a jogabilidade. Encoded-Embodied Inspirada pelo modelo psicoacústico apresentado por Walter Murch [13], onde é apresentado um espectro de cores de som, esta heurística mapeia cada uma das 5 cores do modelo para as 5 Layers que usamos [15]. Seguindo as leis do modelo, e com o objectivo de manter uma densidade de soundscape aceitável, esta heurística não permite mais do que 2 fontes de cada cor (layer) a actuar em simultâneo. Prova de conceito Blindfold Depois de concluído o design e a implementação do módulo de composição dinâmica, foi verificada a exequibilidade do sistema num cenário de jogo. O cenário de jogo escolhido foi o Blindfold, um jogo de aventura audio-only (sem componente gráfica) previamente desenvolvido por nós. O jogo consiste numa aventura enigmática que tenta evocar emoções no jogador, sendo a audição o único sentido que pode guiar os jogadores durante a experiência. A Figura 3 apresenta um screenshot com fins de debugging, aqui apresentado apenas para uma facilitar a caraterização do cenário de jogo. O jogo Blindfold foi desenvolvido em XNA [12], tendo a implementação da sua componente sonora sido feita originalmente através de XACT [11]. 60

61 A experimentação de exequibilidade consistiu então na reimplementação do Blindfold, mas desta vez com toda a implementação e renderização de som feita através do módulo proposto neste trabalho. É importante referir que o objectivo desta experiência não era uma comparação directa de resultados, mas sim testar a exequibilidade da abordagem proposta, assim como verificar os comportamentos de cada uma das heurísticas. O jogo Blindfold, devido à sua variedade de explorações de som, revelou-se um cenário ideal para testar o potencial oferecido pelas diferentes heurísticas implementadas. Figura 2 - Screenshot de debug do jogo Blindfold A título de exemplo, a heurística Context, na qual é dado relevo a algumas fontes, e são atenuadas as restantes, foi particularmente útil para ajudar a distinguir diferentes momentos de jogo. Uma delas verificou-se na situação, durante o jogo, em que os jogadores recolhem um bebé que se encontra abandonado no mundo de jogo. Após esse evento, e apenas exigindo uma linha de código para activar o contexto necessário, a soundscape passa a dar relevo aos sons do bebé, e da mãe que chora desesperada à sua procura. Desta forma, o jogador recebe uma pista sobre a tarefa a fazer de seguida, ao mesmo tempo que a composição o ajuda no processo de ecolocalização da mãe. A heurística de Awareness também se revelou bastante útil nesta experimentação. Esta heurística permitiu que, por exemplo, sempre que novos elementos entrassem dentro do raio de audição do jogador, estes elementos tivessem maior relevância na soundscape durante alguns segundos. Já as heurísticas de Dialogue, e Thoughts, permitiram, respectivamente, garantir que as linhas de diálogo presentes no jogo tivesse sempre relevo na soundscape, e permitir que mais tarde pudessem ser recordadas pelo jogador, sob a forma de pensamento. Finalmente, importa referir que a heurística relativa aos SFX (Achivement/Failure/No Can Do), teve também extrema utilidade. Neste jogo, que é audio-only, o som tem o papel de dar todo o feedback acerca das acções do jogador, pelo que é fundamental garantir que este tipo de sons com alto valor semântico seja sempre ouvido. Em função desta nossa experiência de aplicação das heurísticas, argumentamos que se confirmou a exequibilidade da solução proposta neste exercício em particular, assim como saiu reforçada a ideia de que as heurísticas desenvolvidas podem realmente ser úteis em diversas situações de jogo. Defendemos que inúmeras situações de jogo exibiram o proveito retirado das explorações de som já cobertas pelas heurísticas implementadas, e que permitiram realçar aspectos importantes da jogabilidade em diferentes momentos. 61

62 Em especial, e na linha do que já foi referido, a heurística de Context permitiu que de forma quase instantânea, a soundscape pudesse ser adaptada a contextos diferentes que se iam moldando de acordo com as acções do jogador. Além disso, este exercício demonstrou que é possível alcançar um comportamento acústico bastante interessante através de uma implementação significativamente mais leve, usando a API desenvolvida. O facto do módulo de composição poupar o trabalho de implementar vários comportamentos torna o código bastante mais limpo, devido aos comportamentos relacionados com som que desaparecem do código da lógica do jogo, por já estarem codificados nas heurísticas do módulo de composição. Por outro lado, esta separação entre lógica de jogo, e a lógica de comportamento do som, leva a uma melhor legibilidade do código do próprio jogo. Trabalho Futuro Como referimos, o conjunto de heurísticas aqui apresentadas é ainda reduzido do ponto de vista das situações de jogo que permite de cobrir. Nesta fase, serviu essencialmente para suportar a investigação, incluindo através de experimentação. Como trabalho futuro, temos intenções de actualizar este conjunto de heurísticas, e fazer a respetiva avaliação de comportamento, nomeadamente envolvendo utilizadores. Também estamos particularmente interessados em estudar o impacto desta proposta como ferramenta de prototipagem rápida, em adição, ou como alternativa, à sua utilização para implementação. A utilização em cenários de prototipagem rápida poderia beneficiar bastante da construção de uma interface gráfica que permitisse a alteração em tempo real de alguns parâmetros relacionados com as heurísticas, assim como o visionamento de alguns dados estatísticos relacionados com a composição em curso. Porém, devido ao facto do módulo, actualmente, apenas conseguir ter um contexto activo de cada vez, seria extremamente valioso prototipar uma nova abordagem que guardasse contextos em lista de espera, ou inclusive que existisse suporte para mais que um contexto activo simultaneamente. Finalmente, seria também interessante avaliar o impacto da criação de um novo canal de comunicação entre o módulo de composição e o jogo. Desta forma, a lógica de jogo deixaria de ser surda, e poderíamos testar abordagens nas quais o processo de composição influenciaria o desenrolar da lógica do jogo. Conclusões Neste artigo apresentámos uma proposta para o problema composição dinâmica de som em videojogos, começando com a caracterização de uma das maiores dificuldades do sound design para jogos: a natureza dinâmica do meio. Durante um jogo, as acções do jogador podem originar eventos e situações imprevisíveis, dificilmente cobertas na totalidade pelo sound designer no processo de design a priori. Apesar de serem reconhecidas as potencialidades das ferramentas de Middleware para lidar com este problema, elas são tipicamente demasiado caras e complexas, para constituírem uma solução viável para pequenos criadores de jogos. Defendemos também uma abordagem holística para o sound design para jogos, fundada na Acoustic Ecology, 62

63 que propicia uma apreciação global do relacionamentos entre entidades da soundscape, que também se estendendo ao jogador. O conceito de soundscape saudável foi aqui usado para demonstrar um entendimento da diferença entre uma composição que mantém o seu valor comunicacional, e uma mera sobreposição de fontes que simplesmente se tornaram activas num determinado momento. Assim, uma das contribuições do trabalho apresentado neste artigo é a interpretação de conceitos associados à teoria de Acoustic Ecology, para o contexto do sound design para jogos. Esta transcrição de conceitos e princípios reflecte-se não só em aspectos específicos, como a escolha e definição de heurísticas, mas também nos princípios de alto nível que guiam a arquitectura da solução, e a metodologia de composição. Também valorizámos a procura de soluções acessíveis a todos os tipos de criadores de jogos, incluindo por via da simplificação da implementação de explorações de som em jogos. O desenho e implementação de uma API de especificação de soundscapes é outra das contribuições do trabalho, que vai nesse sentido. É a especificação feita através desta API que permite ao módulo de composição tirar partido das heurísticas, ao permitir uma avaliação das características dos elementos presentes na soundscape. Princípios como acessibilidade e legibilidade foram alguns dos aspectos que guiaram a sua especificação. Apesar de, mesmo assim, exigir algum conhecimento de programação, através desta API pretendemos fazer uso de uma nomenclatura e sintaxe simples, fazendo uso de termos e conceitos próximos da área de sound design. Desta forma, estão criadas condições que permitam que, ao olhar para implementações de som feitas com a API que criámos, possa ser possível esboçar uma ideia das intenções do designer que executou a implementação. Ainda, resulta deste trabalho uma contribuição em forma de lista de técnicas de composição de soundscapes. O objectivo maior destas técnicas é serem instrumentais para a manutenção da saúde da soundscape. A sua especificação foi informada por uma análise a certas prácticas comuns e validadas, usadas actualmente em sound design para jogos, assim como em princípios da teoria de Acoustic Ecology. Estas técnicas, embora ainda careçam de avaliação formal, são uma primeira abordagem a um conjunto de heurísticas que pode ser melhorada e aumentada, procurando dar uma resposta cada vez mais eficaz à imprevisibilidade dos cenários de jogo. Finalmente, e como contribuição principal deste estudo, apresentámos um módulo que executa a composição dinâmica em run-time. Ao servir-se das técnicas de composição embutidas, o módulo reduz a necessidade de prever e codificar a imensidão de respostas sonoras a todas as situações de jogo que possam surgir. A responsabilidade de moderação da soundscape é, assim, transferida do código de jogo, para o módulo de composição, que o executa autonomamente, e em tempo real. Argumentamos que o trabalho desenvolvido tem potencial para servir como base para futuras experiências, nomeadamente com o objectivo de melhor avaliar as heurísticas, e o processo de composição. Consideramos que existe potencial para que uma abordagem deste género, e que com o necessário refinamento e trabalho de investigação adicional, poderá vir a constituir realmente uma 63

64 alternativa em termos de implementação de som em jogos. Referências [1] Alves, V. and Roque, L A Deck for Sound Design in Games - Enhancements based on a Design Exercise. In Proc. of ACE 2011, Lisboa, Portugal. [2] Alves, V. & Roque, L SoundInGames.com Sound Design in Games. [3] Brandon, A Audio Middleware Mix: Professional Audio and Music Production, io_middleware/. [4] Brandon, A Audio Middleware, part 2 Mix: Professional Audio and Music Production, io_middleware_part/. [5] Brandon, A Audio Middleware, part 3 Mix: Professional Audio and Music Production, io_middleware_part_2/. [6] Chan, S.H., Natkin, S., Tiger, G. and Topol, A Extensible Sound Description in COLLADA: A Unique File for a Rich Sound Design. In Proc. of the Advances in Computer Entertainment, 2012 Springer. [7] Collins, K Game Sound: An Introduction to the History, Theory, and Practice of Video Game Music and Sound Design. MIT Press. [8] Ekman, I Psychologically Motivated Techniques for Emotional Sound in Computer Games. In Proc. of the 3rd AudioMostly Conference, Piteå, Sweden, ACM, [9] Hevner, A. R. and Chatterjee, S Design research in information systems. New York; London: Springer. [10] Kastbauer, D Audio Implementation Greats #2: Audio Toolsets [Part2] Designing Sound, [11] Microsoft, n.d. XACT Overview, Microsoft Developer Network, [12] Microsoft, XNA Developer Center Microsoft, [13] Murch, W Dense Clarity Clear Density, The Transom Review, transom.org, [14] Open Sound Control, Introduction to OSC Open Sound Control, [15] Peck, N Beyond the library: Applying filmpostproduction techniques to game sound design - lecture. In GDC 2001, San Jose, CA, March. [16] Peerdeman, P Sound and Music in Games, Vrije Universiteit, Amsterdam. [17] Schafer, R.M The Soundscape: Our Sonic Environment and the Tuning of the World. Inner Traditions/Bear. [18] Truax, B Acoustic Communication. Ablex. [19] Vaishnavi, V. and Kuechler, W Design Science Research in Information Systems. [20] Went, K., Huiberts, S. and Tol, R. V., Game Audio Lab - An Architectural Framework for Nonlinear Audio in Games. In Audio Engineering Society Conference: 35th International Conference: Audio for Games, London, United Kingdom, AES. 64

65 CROWDSOURCING Y VIDEOJUEGOS Los Jugadores como Productores de Conocimiento ISRAEL V. MÁRQUEZ Universidad Complutense de Madrid Av. Séneca, Madrid, España isravmarquez@gmail.com Resumen El crowdsourcing es un tipo de actividad participativa online que consiste en confiar o externalizar a una comunidad o grupo indeterminado de personas (masa) una tarea o tareas que normalmente se organizaban dentro de una organización y que se encargaba de realizar un empleado. Este tipo de práctica colaborativa en línea se ha aplicado a diferentes manifestaciones políticas y socioculturales, llegando también al mundo de los videojuegos. Este trabajo es un intento de explorar qué implicaciones tiene la vinculación entre crowdsourcing y videojuegos en la actual Sociedad de la Información y el Conocimiento mediante el análisis de algunos ejemplos. Palabras clave Crowdsourcing; videojuegos; participación; colaboración; gamificación. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Abstract Crowdsourcing is an online participatory activity where a community or an undefined group of people (crowd) perform a task or tasks normally performed by employees in a company or institution. This type of online collaborative practice has been applied to different political and socio-cultural manifestations, also 65

66 arriving to the gaming world. This paper is an attempt to explore the link between crowdsourcing and video games in the current Information and Knowledge Society by analyzing some examples. Author Keywords Crowdsourcing; Games; Participation; Collaboration; Gamification. ACM Classification Keywords Games 1. Introducción Como sabemos, la web 2.0 es una evolución de la red de Internet que ha dado lugar a un nuevo paradigma de la comunicación basado en aspectos como la participación y la colaboración. Las nuevas aplicaciones de la web 2.0 convierten Internet en un complejo sistema socio-técnico-cultural cuya novedad respecto a la web anterior es su mayor énfasis en el usuario y en el denominado software social, es decir, en las múltiples relaciones entre los miembros que tienen lugar en la red. Así, ya se trate de blogs, videoblogs, podcast, videopodcast, wikis o redes sociales, la esencia común a todos ellos es la relación, millones de personas participando en dinámicas periódicas de creación de contenidos, intercambio de información y construcción de nuevos espacios de comunicación y relación en la red. Como siempre ocurre cuando un nuevo fenómeno hace su aparición, junto al término Web 2.0 han surgido otros que pretenden designar la nueva realidad pero bajo expresiones diferentes. Por ejemplo, Tapscott y Williams [16] prefieren el término wikinomics al de web 2.0, haciendo hincapié en el papel esencial de la colaboración y el intercambio que las nuevas aplicaciones estimulan (en especial los wikis, esas herramientas abiertas y sencillas de trabajo colaborativo en línea simbolizadas por la Wikipedia). Para otros autores, como Henry Jenkins [11, 12], la noción central para describir la nueva realidad es la de la inteligencia colectiva. Para otros, como el autor y redactor jefe de la revista Wired, Chris Anderson, el momento actual se caracterizaría por lo que él denomina la larga cola [1], expresión que pone de relieve la aparición de nuevos modelos económicos y nuevas oportunidades de creación de valores que la nueva web permite, lo que da lugar a una economía de la diversidad y de la abundancia. Finalmente, otros como Pisani y Piotet [13] prefieren hablar de la emergencia de una nueva dinámica relacional que choca con la mecánica institucional tradicional, y con la herencia intelectual y social sobre la que ésta se construyó, para destacar el conjunto de movimientos no controlados, no lineales, con múltiples vertientes y causados por la participación de todos que caracteriza la web de hoy y que está afectando al establecimiento de relaciones entre personas, grupos y datos. Sea cual sea la expresión que prefiramos (conscientes, sin embargo, de que ha sido el término web 2.0 el que ha hecho fortuna y ha dado la vuelta al mundo), lo importante es que todas ellas son producto de un nuevo enfoque a la hora de pensar y entender la sociedad de la información y del conocimiento, un nuevo paradigma basado en la participación y la colaboración y que genera una verdadera poética de la relación, en expresión de Édouard Glissant [7]. 66

67 2. Qué es el crowdsourcing? Uno de los fenómenos donde mejor pueden apreciarse todos estos aspectos es en el movimiento conocido como crowdsourcing. Como veíamos con el concepto de larga cola de Anderson [1], la nueva web da lugar al surgimiento de nuevos modelos económicos basados en la abundancia y la diversidad. Asimismo, Tim O Reilly, la persona que popularizó el término Web 2.0, decía que una de las lecciones esenciales de la web 2.0 es que los usuarios añaden valor. Por lo tanto, la participación, que es el núcleo de esa dinámica o poética de la relación que anima la web de hoy, también da lugar a un verdadero modelo económico cuya manifestación más exitosa hasta la fecha es el crowdsourcing, o traducido al castellano, la externalización de la producción de masas. El término se inspira en dos expresiones conocidas en el universo de la web 2.0: el outsourcing o externalización, y el wisdom of crowds o sabiduría de las masas, y básicamente consiste en confiar o externalizar a una comunidad o grupo indeterminado de personas (masa) una tarea o tareas que normalmente se organizaban dentro de una organización y que se encargaba de realizar un empleado o contratista. El término fue popularizado por Jeff Howe en un artículo de 2006 para la revista Wired [10]. Howe pensaba que gracias a la interconexión de millones de personas a través de Internet y a que las diferencias entre profesionales y aficionados ha disminuido con los años, podía ser factible encargar trabajos no a pequeños grupos de expertos, sino presentar un problema o proyecto como una convocatoria abierta al mundo digital y dejar que la masa (crowd) encontrara una solución. En realidad, el crowdsourcing tiene sus orígenes en el esfuerzo colectivo de los pioneros de Internet, muy especialmente en el movimiento de programas de código abierto que empezó en 1983 con el proyecto de colaboración masiva GNU iniciado por Richard Stallman y en cuyo manifiesto se podía leer cómo una de sus motivaciones a la hora de realizar este proyecto fue volver al espíritu de cooperación que prevaleció en los tiempos iniciales de la comunidad de usuarios de computadoras. Dicho manifiesto fue publicado en marzo de 1985 en la revista Dr. Dobb's Journal of Software Tools como una explicación y definición de las metas del Proyecto GNU para llamar a la participación y obtener apoyo, es decir, lo que caracteriza actualmente al crowdsourcing. El lanzamiento en 1992 de Linux, el primer sistema operativo del proyecto GNU, y su posterior éxito mundial, demostró que la colaboración gratuita y movida por la pasión es uno de los aspectos más potentes de la cultura digital. Otros proyectos como el LazyWeb o el ESP Game de Luis von Ahn, también preceden la invención del término crowdsourcing desde hace varios años. Es más, aún podemos viajar más atrás en el tiempo y pensar la colaboración masiva como una idea mucho más antigua que cualquier tecnología digital: Qué son los buzones de sugerencias o los proyectos de investigación que pedían a miembros del público que mandaran sus opiniones, preguntas o sugerencias sino ejemplos de participación y colaboración masiva? [3]. Sin embargo, lo que sí ha hecho la tecnología digital ha sido incrementar exponencialmente el número, la facilidad y la velocidad de este tipo de colaboraciones hasta convertir la interacción masiva en una parte esencial de nuestra aldea global. El término crowdsourcing se ha hecho popular entre empresas, autores y periodistas como una forma abreviada de hacer referencia a la tendencia a impulsar 67

68 la colaboración masiva habilitada por las herramientas de la web 2.0 para lograr unos determinados objetivos. En la literatura científica existen actualmente más de 40 definiciones distintas del término, y distintos autores dan definiciones de crowdsourcing diferentes basándose en sus especialidades, lo que hace que se pierda la visión global del concepto. Estellés y González, tras estudiar estas más de 40 definiciones del término, proponen una nueva definición integradora: El crowdsourcing es un tipo de actividad participativa online en la que una persona, institución, organización sin ánimo de lucro o empresa, propone a un grupo de individuos mediante una convocatoria abierta flexible la realización libre y voluntaria de una tarea. La realización de la tarea, de complejidad y modularidad variable, y en la que la multitud debe participar aportando su trabajo, dinero, conocimiento y/o experiencia, siempre implica un beneficio mutuo. El usuario recibirá la satisfacción de una necesidad concreta, ya sea esta económica, de reconocimiento social, de autoestima, o de desarrollo de aptitudes personales, mientras que el crowdsourcer obtendrá y utilizará en su beneficio la aportación del usuario, cuya forma dependerá del tipo de actividad realizada [6]. 3. Ejemplos de crowdsourcing Uno de los primeros ejemplos de crowdsourcing fue el programa SETI (Search for Extra-Terrestrial Intelligence) de Berkeley, California, el cual se basaba en el análisis de enormes cantidades de datos en forma de señales electromagnéticas capturadas en distintos radiotelescopios. El análisis de estos datos era una tarea mucho más grande de lo que hasta los ordenadores más avanzados tecnológicamente podían asumir y así, en 1990, se lanzó el proyecto SETI@home, que se ejecuta sobre la plataforma Berkeley Open Infrastructure for Network Computing (BOINC) y el cual pedía a gente de todo el mundo que prestara la potencia de su ordenador para el análisis de tales datos, descargando un sencillo programa que funcionaba como salvapantallas. Este programa está siendo apoyado por millones de personas de todo el mundo mediante el uso de sus computadoras personales, que procesan la información capturada por el radiotelescopio de Arecibo, emplazado en Puerto Rico, y trasmiten los resultados a Berkeley cuando el ordenador está en desuso durante un rato. Se calcula que hay más de 5 millones de usuarios en más de 200 países que están participando actualmente en este programa, los cuales han contribuido con millones de horas de computadora [2]. SETI@home es un proyecto dirigido de manera central por una institución académica de élite. Pero también existen trabajos de colaboración mucho más ambiciosos y descentralizados. Entre los más famosos está la popular Wikipedia, una plataforma que nació en 2001 con el modesto objetivo de establecer una enciclopedia gratuita en la red que cualquier persona pudiera utilizar y corregir. A finales de 2010, Wikipedia ya era responsable de más de 100 horas de trabajo humano y en el proceso se ha convertido en la fuente general más exhaustiva de información, con una calidad muchas veces superior a la prevista. La famosa enciclopedia virtual se ha convertido en emblema del potencial colaborador de la web y de la nueva cultura virtual de la participación masiva. Otro ejemplo es el de MechanicalTurk (MTurk), un sitio creado por Amazon.com que permite encontrar la mano de obra necesaria para tareas que los ordenadores no pueden 68

69 hacer, tales como identificar elementos en una fotografía, describir un producto en pocas líneas, transcribir un podcast, etc. Un último ejemplo de entre los muchos que podríamos mencionar es el del sitio CrowdSpirit.org, que se define como una comunidad de crowdsourcing dentro de la cual todo el mundo puede participar en la concepción y fabricación de aparatos electrónicos baratos. La comunidad propone, vota, selecciona, financia, prueba y produce los aparatos, que están inspirados en los deseos y expectativas de los propios usuarios que los conciben. Todo esto hace que, como apunta David Weinberger, ahora todo el mundo es experto, como demuestran ejemplos paradigmáticos como el de la Wikipedia. El saber que se desprende de esta interacción masiva suele ser mejor que el que podríamos esperar de un solo individuo. El experto no desaparece, pero estamos asistiendo a una especie de negociación social del saber [13]. Así, el mayor valor del crowdsourcing podría ser que, al ampliarse la participación, la externalización permite aprovechar una gran cantidad de talentos, de nuevos expertos, generando mecanismos de inteligencia colectiva y transmisión de conocimientos con un potencial considerable en términos de invención, innovación, investigación y poder. En este último sentido, Henry Jenkins [11] considera que la idea de la "inteligencia colectiva" puede verse como una fuente alternativa de poder. Según Jenkins, actualmente estamos aprendiendo a usar ese poder mediante nuestras interacciones cotidianas en el seno de la cultura de la convergencia, principalmente a través de nuestra vida recreativa. Ahí están los casos bien documentados por el autor de cómo las comunidades de fans de programas como Survivor o American Idol, formadas en torno a unos intereses intelectuales y emocionales comunes, consiguieron desvelar mediante un arduo trabajo colectivo de búsqueda, recopilación y evaluación de la información los muchos secretos contenidos en ambas series televisivas antes de que se revelaran en antena. Para Jenkins, este trabajo en común, que denomina "destripe" (spoiling), es la inteligencia colectiva puesta en práctica, y la importancia que ve en este fenómeno le lleva a preguntarse, y nosotros con él, sobre las clases de información que podrían recopilar estos fans si su propósito fuese destripar ciertos gobiernos en lugar de cadenas. El crowdsourcing ya ha llevado a cabo este tipo de prácticas en casos como el del periódico The Guardian, que en 2009 utilizó una red de más de voluntarios para investigar las polémicas acusaciones de gastos de los parlamentarios británicos. Otro ejemplo sería el de la Library of Congress, que en 2010 utilizó el sitio de álbumes de fotos compartidos Flickr para llamar al público a identificar a los personajes que aparecían en una serie de fotos de la Guerra Civil [3]. En casos como estos podemos ver el potencial revolucionario del crowdsourcing, y de la inteligencia colectiva inherente a él, como instrumento político y fuente alternativa de poder, más allá de sus usos en el mundo de la cultura, la economía y el entretenimiento con los que comúnmente suele asociarse el movimiento. 4. Crowdsourcing y videojuegos El mundo digital ha dado lugar al surgimiento de una nueva cultura participativa que contrasta con nociones más antiguas del espectador mediático pasivo. Este carácter más activo y participativo del espectador actual responde también al nuevo paradigma relacional 69

70 que caracteriza la web 2.0. Así, de la web estática de los primeros años de la revolución informática, pensada básicamente como una herramienta para navegar y encontrar información y donde el internauta consumía contenidos en lugar de crearlos, hemos pasado a una web dinámica en la que el internauta adquiere un mayor protagonismo como usuario, un lugar en el que los instrumentos para colaborar, personalizar la experiencia, hacer vida social, compartir y crear están al alcance de todos. De ahí que la web 2.0 haya impulsado la idea del consumidor como protagonista, dando lugar a un nuevo tipo de espectadores caracterizados por su papel participativo y creativo y que exigen nuevas denominaciones (usuarios, webactores, prosumers, fans, etc.) que se aparten de las connotaciones pasivas y negativas que suele acompañar a la palabra espectador. Esto no quiere decir que los usuarios hayan tomado el lugar de los medios de comunicación de masas sino que, como puntualiza Henry Jenkins, se produce una mayor interacción entre ambos polos: Más que hablar de productores y consumidores mediáticos como si desempañasen roles separados, podríamos verlos hoy como participantes que interaccionan conforme a un nuevo conjunto de reglas que ninguno de nosotros comprende del todo. No todos los participantes son creados iguales. Las corporaciones, e incluso los individuos dentro de los medios corporativos, ejercen todavía un poder superior al de cualquier consumidor individual o incluso al del conjunto de consumidores. Y unos consumidores poseen mayores capacidades que otros para participar en esta cultura emergente [11]. En el mundo de los videojuegos todo esto se manifiesta en una mayor interacción con los propios diseñadores de juegos e incluso una mayor participación en los procesos de producción de los mismos. La propia industria de los videojuegos, consciente de que el mundo de los fans y otras comunidades de conocimientos desarrollan un sentimiento de afiliación y fidelidad a determinadas marcas-juegos, ha decidido ampliar la participación de los jugadores y fortalecer aún más ese sentimiento de afiliación, lo que asegura la longevidad de determinadas líneas de productos. Por ejemplo, LucasArts integró a potenciales jugadores de Star Wars en el equipo de diseño para el desarrollo de su juego en línea multijugador Star War Galaxies. Desde su inicio, Galaxies se ha caracterizado por la constante interacción entre jugadores y diseñadores, creándose una página web (starwargalaxies.station.sony.com) donde se iban colgando las ideas que se estaban manejando para conocer las reacciones de los fans y recoger sus comentarios y sugerencias de cara a incorporarlos en el producto final. Como señala uno de los productores de LucasArts, Haden Blackman, hubo algunos sistemas en los que, francamente, estábamos teniendo dificultades para encontrar una solución [8]. Por eso decidieron postear información en la página web y permitir que los jugadores aportaran sus propias soluciones: Han ofrecido alternativas y sin duda nos han persuadido en muchos casos para hacer cambios, como por ejemplo permitir que los Wookiees u otras especies fueran parte de la milicia imperial, que es básicamente un régimen masculino blanco, como puede apreciarse en las películas: Cuando empezamos a hablar con la comunidad acerca de esto, hubo una gran emoción y alboroto en cuanto a 70

71 permitir que los Wookiees estuvieran en el ejército imperial. Mucha gente sentía que eso rompería la ficción y arruinaría el juego. Así que llegar a un acuerdo sobre esto y encontrar una manera de resolverlo fue difícil. Pero trabajamos con la comunidad y llegamos a una especie de compromiso donde, sí, si eres una especie puedes unirte a los militares, pero será mucho más difícil para ti. Vas a tener que trabajar dos, tres o cuatro veces más, dependiendo de qué tipo de especie seas. Y eso parecía satisfacer a todos [8]. Kurt Squire [15] ve en el ejemplo de Star War Galaxies un caso de diseño participativo, ya que los usuarios habituales, normalmente excluidos del proceso de diseño, pueden aportar a la conversación su propia experiencia y pericia en el manejo de productos, y contribuir a la creación de productos más servibles y manejables lo cual, según Squire, beneficia a todo el mundo. A diferencia de la web oficial de la película, donde los trailers, imágenes y anuncios fluyen de arriba-abajo, desde el estudio hacia una audiencia entusiasmada e impaciente, el sitio web del juego sirve tanto para recoger opiniones y sugerencias de los jugadores como para lanzar detalles sobre el producto. En esos comentarios, ideas y sugerencias de los propios jugadores puede verse la lógica participativa del crowdsourcing, en el sentido de que se propone una solución colectiva a un problema planteado por el videojuego por parte de los mismos jugadores que lo jugarán, si bien la última palabra la siguen teniendo los desarrolladores del juego. La lógica cultural subyacente a tal proceso es, por tanto, doble: es tanto un proceso corporativo de arriba-abajo como un proceso de abajoarriba dirigido por los consumidores-jugadores. La convergencia corporativa de los medios coexiste con la convergencia popular de los consumidores, dotados de un nuevo poder digital en la configuración de la producción, distribución y recepción de los contenidos mediáticos [11, 12]. Uno de los casos más comentados con respecto a la unión entre crowdsourcing y videojuegos es el proyecto Digitalkoot, creado por la empresa finlandesa Microtask. Se trata del primer proyecto europeo que permite a los ciudadanos participar en la digitalización de documentos, concretamente aquellos relacionados con el patrimonio cultural finlandés y que se encuentran disponibles en la Biblioteca Nacional de Finlandia. Digitalkoot, que significa voluntarios digitales, combina el crowdsourcing y la gamificación (gamification) para convertir una tarea monótona y aburrida como es el digitalizar y catalogar archivos en algo divertido y colaborativo. Como señala Kai Ekholm, el director de la Biblioteca Nacional de Finlandia: Tenemos millones y millones de páginas de la historia y revistas de valor cultural, periódicos y revistas en línea. El desafío es eliminar los errores que se generan con el reconocimiento óptico de caracteres que omiten caracteres y dificultan las búsquedas. En este sentido, la corrección manual es necesaria para eliminar estos errores para que los textos sean de lectura mecánica, permitiendo a los académicos y archivistas buscar el material y la información que necesitan [9]. Todo ello se realiza a través de dos juegos en línea. El primero se llama Mole Hunt (Myyräjahti), y su funcionamiento es el siguiente: se muestran al jugador dos palabras diferentes y su objetivo es determinar rápidamente si son las mismas, descubriendo así palabras erróneas en los documentos guardados. El 71

72 segundo juego se llama Mole Bridge (Myyräsilta) y en él los jugadores deben deletrear correctamente las palabras que aparecen en la pantalla. Si acierta, el jugador ayudará a los topos a construir un puente sobre un río y evitará que caigan al vacio. De esta forma se verifica el reconocimiento óptico de caracteres y se asegura la exactitud del material digitalizado. Las palabras que los jugadores tienen que escribir proceden de millones de páginas de periódicos, revistas y otras publicaciones, las cuales han sido digitalizadas con tecnología de reconocimiento óptico. Cada vez que los jugadores completan un nivel están contribuyendo a algo en la vida real: eliminar los errores creados al digitalizar esos archivos y mejorar la visibilidad, accesibilidad y usabilidad de los mismos. Figura 1. Digitalkoot: Mole Bridge En un estudio empírico realizado por dos miembros de Microtask [4], se descubrió que el sitio recibió visitantes en los 51 días siguientes a su lanzamiento y que probaron al menos uno de los dos juegos. Estos usuarios donaron el equivalente a horas de trabajo efectivo y completaron 2,5 millones de tareas. A partir de estos resultados, se llegó a la conclusión de que el usuario típico de Digitalkoot gastó 9,3 minutos en los juegos y completó 118 tareas. Así, mediante un diseño ameno y una narrativa simple como la de unos topos que pretenden construir y cruzar un puente, Microtask ha logrado involucrar a un gran número de jugadores que aportan su tiempo y conocimiento a una causa real a través de un videojuego atractivo y sencillo. El creciente género de los juegos casuales (casual games) -entendidos como aquellos juegos con reglas simples, que se pueden disfrutar en periodos cortos y que no requieren habilidades especiales para jugar-, son ideales para la mayoría de los proyectos de crowdsourcing [14], ya que son fáciles de aprender, se basan en controles simples, y generan en el jugador un sentimiento de progreso rápido y una inmediata gratificación por sus acciones, lo cual aumenta su motivación y les anima a seguir jugando. Lo importante en el caso de juegos casuales vinculados a proyectos de crowdsourcing, como en Digitalkoot, es que siempre que jugamos colaboramos a una causa común y convertimos algo que de otra manera sería rutinario y aburrido en una diversión, en un juego con efectos en la vida real. Pero el ejemplo más conocido de crowdsourcing en el campo de los videojuegos es sin duda el de Foldit. Es éste un videojuego masivo en línea desarrollado por un grupo de científicos y diseñadores de juegos de la Universidad de Washington que simula la estructura de las proteínas. El juego está basado en el proyecto Rosetta@home, muy similar al proyecto SETI@home de la Universidad de Berkeley que vimos anteriormente y dependiente, como éste, de la plataforma Berkeley Open Infrastructure for Network Computing (BOINC). El juego, al igual que todos los proyectos BOINC, utiliza la 72

73 capacidad de procesamiento de los ordenadores de voluntarios para ejecutar cálculos en unidades de trabajo individuales, en este caso relacionados con el plegamiento de proteínas. De ahí pasó a ser interactivo, con los jugadores diseñando estas mismas estructuras y convirtiéndose por tanto en productores. Su propósito es encontrar las formas naturales de las proteínas que forman parte de los seres vivos. Así, jugadores de cualquier parte del planeta pueden plegar proteínas sin necesidad de conocimientos previos sobre la materia a través de la manipulación de modelos 3D de las mismas y con el objetivo de lograr estructuras funcionales, nuevas o más eficientes. Los jugadores colaboran y desarrollan estrategias mientras manipulan esos modelos de proteínas como si se tratara de un puzzle. Cuanto mejor plegada esté la proteína, más puntos obtendrá el jugador, cuyos datos son almacenados en un ranking on line de jugadores, por lo que también hay un elemento de competición en todo ello. Foldit cuenta con más de usuarios registrados desde su lanzamiento en 2008, y con más de activos por semana. Figura 2. Foldit En 2011, un grupo de jugadores de Foldit consiguieron descifrar en tan sólo 10 días la estructura cristalina de la proteasa retroviral M-PMV (Mason-Pfizer Monkey Virus, una enzima concreta causante del SIDA en los simios), algo que los científicos llevaban años intentando lograr. Y en 2012, otro grupo de jugadores logró mejorar en tan solo tres semanas una proteína en investigación hasta hacerla 18 veces más eficaces que la original: He trabajado dos años para mejorar esas enzimas, pero no lo conseguí, señala Justin Siegel, un investigador postdoctoral del equipo responsable del juego. Los jugadores de Foldit, en cambio, consiguieron un gran salto en su estructura especial, y todavía no sé cómo lo han logrado [5]. Ambos resultados han sido publicados en la revista de investigación Nature Structural & Molecular Biology, y los jugadores de Foldit responsables de tales hallazgos figuran como coautores junto con los propios científicos. El trabajo colaborativo de Foldit, una especie de crowdsourcing híbrido lúdico-científico, ha permitido que los jugadores hayan podido compartir ideas, soluciones, comentarios y sugerencias, compitiendo a la vez que colaboraban. Se trata de otro caso sonde la ludificación o gamificación es muy clara, ya que se busca diseñar y fomentar la participación de los usuarios en forma de juego. Foldit es sin duda un ejemplo maravilloso de cómo el crowdsourcing puede aplicarse a la solución de problemas globales (las enfermedades, en este caso) a través de los videojuegos, promoviendo resultados positivos para la comunidad científica y las sociedades por medio de la colaboración y la participación masiva de jugadores convertidos ahora en productores de información y conocimiento. Así, como señalan los propios responsables de Foldit: 73

74 Aunque recientemente se ha prestado mucha atención al potencial del crowdsourcing y los videojuegos, este es el primer caso del que somos conscientes en el que los jugadores han resuelto un viejo problema científico. Estos resultados indican el potencial de integrar los videojuegos en el proceso científico del mundo real: el ingenio de los jugadores es una fuerza formidable que, bien dirigido, puede utilizarse para resolver una amplia gama de problemas científicos [17]. 5. Conclusiones A lo largo de este artículo hemos explorado algunas ideas relacionadas con la web 2.0 y el crowdsourcing, para acabar analizando la influencia actual de este último fenómeno en el mundo de los videojuegos. Algunos casos de crowdsourcing, como el de Digitalkoot o Foldit, han mostrado el potencial de integrar a los jugadores en los procesos de producción de información y conocimiento, incluido el conocimiento científico. El caso de Foldit es un ejemplo reciente de cómo el crowdsourcing puede aplicarse a la solución de problemas globales a través de los videojuegos, promoviendo resultados positivos para la comunidad científica y las sociedades por medio de la colaboración y la participación masiva de jugadores, convertidos ahora en productores de conocimiento. Sin embargo, al tratarse de un fenómeno tan reciente y en pleno desarrollo deberemos estar pendientes de su evolución y de los nuevos proyectos que vayan surgiendo con el fin de determinar el verdadero alcance de estas nuevas prácticas de colaboración videolúdica. 6. Referencias [1] Anderson, C. The Longer Long Tail. How Endless Choice is Creating Unlimited Demand. London: Random House Business Books, [2] [3] Chatfield, T. 50 Digital Ideas You Really Need to Know. London: Quercus, [4] Chrons, O., y Sundell, S. Digitalkoot: Making Old Archives Accessible Using Crowdsourcing. AAAI Workshop (2011). [5] De Benito, E. Un videojuego permite diseñar una proteína más potente que la natural. El País, 23 enero [6] Estellés, E., y González, F. Towards an integrated crowdsourcing definition. Journal of Information Science (2012), vol. 38, no. 2, pp [7] Glissant, E. Introduction à une poétique du divers. París: Gallimard, [8] Herz, J.C. 50,000,000 Star Warriors Can't Be Wrong. Wired (2002), Issue [9] Hormiga Analítica, 8 de febrero de teca-nacional-de-finlandia.html [10] Howe, J. The Rise of Crowdsourcing. Wired (2006), Issue [11] Jenkins, H. Convergence Culture. La cultura de la convergencia de los medios de comunicación. Barcelona: Paidós, [12] Jenkins, H. Fans, blogueros y videojuegos. La cultura de la colaboración. Barcelona: Paidós, [13] Pisani, F., y Piotet, D. La alquimia de las multitudes. Cómo la web está cambiando el mundo. Barcelona: Paidós, [14] Ridge, M. Playing with difficult objects: game designs for crowdsourcing museum metadata. Dissertation. (MSc). City University, [15] Squire, K. Star War Galaxias: A case study in participatory design. Joystick 101 (20 july 2001). 74

75 [16] Tapscott, D., y Williams, A.D. Wikinomics. How Mass Collaboration Changes Everything. London: Atlantic Books, [17] VV.AA. Crystal structure of a monomeric retroviral protease solved by protein folding game players. Nature Structural & Molecular Biology (2011), 18 september 2011, pp

76 Our Ludic Sins A Survey on Belief in Videogame Axioms Rui Craveirinha Department of Informatics Engineering Faculty of Sciences and Technology, University of Coimbra Portugal rui.craveirinha@gmail.com Licinio Roque Department of Informatics Engineering Faculty of Sciences and Technology, University of Coimbra Portugal lir@dei.uc.pt Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Abstract Creativity has not been as pervasive an element in the videogame medium as would be desired. In this paper is proposed that one underlying cause resides in an imperfect study of the game design discipline, permeated by a number of prescriptive axioms on videogames form, value and expression that impair creative processes. To support the contention that these have become ubiquitous in the medium, reaching the community as a whole, a survey intended on testing the agreement of subjects to these misconceptions was realized. 95 subjects replied to 41 sentences in Likert scale responses, and 10 responses had high agreement. Results indicate that the majority of subjects sees videogames as both an art form and an extension of games and narratives, consisting necessarily of interaction, rules and challenges, and serving primarily as a form of intellectually stimulating entertainment. Conversely, there was no agreement to videogames being a qualitatively new medium, nor as a medium that need be emotionally expressive or mediator of discourse on real-life issues. Author Keywords Games, Art ACM Classification Keywords K.8.0 [Games]: General 76

77 General Terms Theory, Design Introduction What are videogames? Are they games? Art? What should they express? And how do we value them in an increasingly complex media landscape? Questions such as these are at the heart of a heated conversation that never seems to subside in the videogame community. The Game Design discipline has been the stage in which this debate occurs, with different authors from vastly different backgrounds proposing alternative answers to these questions. Recently we published a position paper [5] where we proposed that, despite the richness of perspectives, there are dominant currents of thought underlying analytical discourses on videogames. Our thesis is that both form, value and expression have been subject to constraining normative reasonings, that in an attempt to study videogame qualities, created a biased view of the medium that fails to encompass its more creative and heterodox currents. Thus, we claim that this perspective can constrain creative acts, for subscribing heavily to these conventions tends to guide production efforts to repeat the same formulas from which they were extracted in the first place. This analysis however, was strictly rhethoric, resulting of a review of several papers, design books and media articles. Therefore, there is no empirical evidence to even support that these axioms have been adopted, let alone that they have an effect in actual game design. Thus, to clarify which, if any, of the claimed normative paradigms have been subscribed by the community, a survey was carried out, testing subjects agreement with sentences that were representative of the normative assertions extracted in our previous work [5]. The paper will go on to provide a literature review briefly summarizing our thesis on this subject [5], so as to justify the survey. Then, the survey s setup will be detailed in section and its results presented in section. Finally, section will provide our own interpretation of what this results mean, followed by some final conclusions. Thesis Salen and Zimmerman acknowledged the dubious, infantile, unoriginal nature of games in stores [18], and Bogost bitterly recognized that our society tends to view videogames as a children s medium [2]. Chris Crawford believed videogames should aspire to be art, stating that computer games constitute an as-yet untapped art form...back in 1982 [6]. Meanwhile, videogames are still struggling to emerge from their arrested adolescence [17], remain focused on fantasy genres, monsters and trolls [8] and are emotionally shallow [10]. Even today, though their cultural relevance is beyond questioning [17, 2, 18] it is common to find videogame authors that present serious doubts, criticisms and ambiguities when qualifying their medium (see, as examples, [4, 10, 13, 1, 15]. While other causes are surely at work, we propose that this is, in part, a consequence of normative thinking. When studying existing videogames, all members of the community conjure videogames based on an abstraction, a prominent parcel of the medium they can perceive, analyze, catalogue. Then, a number of axioms, in the form of definitions, rules, tendencies and biases - or norms - are extracted as knowledge. This is what we are addressing as normative thinking. The problem with this process is that, in a creative medium, everything is in constant flux. Definitions, currents, movements, styles, forms, expressions, valuations, these are ever changing in time, place and context, as different trends gain or lose 77

78 favour from artists, scholars and audiences. Nowhere is this more evident than in the History of Art [11]: [the artist] does not follow any fixed rules. [...] some artists and critics in certain periods have tried to formulate laws of their art; but it always turned out that poor artists did not achieve anything when trying to apply these laws, while great masters could break them and yet achieve a new kind of harmony no one had thought of before. [...] The truth is that it is impossible to lay down rules of this kind because one can never know in advance what effect the artist may wish to achieve. The study of a medium by way of these norms is certainly useful when their limitations are properly understood and communicated. But when they are conveyed as absolute truths, creators become mentally constrained to repeat those same reasonings, by implementing artifacts of the future that sustain that perception of the past. This normative reasoning then becomes a constraining box that obscures creative approaches, by inducing a vicious cycle of reproduction. Creativity is the act of creation. And creation, by definition, presupposes the coming into existence of something that did not exist before...something new. Thus, every creative act is implicitly an act of disruption with the past. To avoid normative assertions pernitious bias, it is necessary to know which norms have become implicitly accepted by the community, so that they can then be framed in a way that facilitates the design of artifacts that are creative to the point of countering those. In [5], we dissert on a number of different norms which appear to have become accepted in the community s discourse, both in its academia and outside of it. For reasons of space, it is not possible to provide the theoretical rationale that supports their inclusion in this article s survey, so refer to the article for an in-depth look at their justification. Experimental Setup To test whether or not these preconceptions had been adopted by the community at large, a survey was carried out. Two samples of the population were incorporated in this study: one consisting of 41 Computer Engineering and Multimedia and Design students (inscribed in a Game Design and Study Course) that were queried in a class by paper questionanires, and another consisting of 54 subjects that were invited to participate through social networks via an internet form. The sample had an age average of 25.5 years, standard deviation of 5.47, and was comprised of 87 male and 8 female subjects, 79 were regular videogame players, 46 accessed regularly videogame related media, and 78 had read at least one article concerning the study of videogames. The survey consisted of Likert Scale with responses ranging from Strongly Agree (5) to Strongly disagree (1), with 41 sentences divided into 5 groups. Each group was designed to test normative assertions that seemed predominant in our review [5]. Two groups concern form, two refer to how value is assessed, and finally, one covering expression. Questions either follow normative assertions from our review (e.g. Videogames are games ), or present alternatives ( Videogames are something new which despite similarities cannot be accurately assigned to any of the above categories. ). The goal was to verify if there was a tendency for popular acceptance or rejection of normative assertions and other alternatives. Analysis was done by doing, for each question group, a first pass to verify which questions had average responses above or below the middle point in the Likert scale (avg < 2.5 avg > 3.5); when the average was very close to borderline values, the median was used as tiebreaker 78

79 79

80 criteria. Whenever this occurs, to validate that value, we test to see if the answer is statistically higher or lower than the remaining responses in its category, by means of Wilcoxon signed-rank test for related samples (a non-parametric test suited for ordinal scales) 1. Figure 1 shows all the questions, as well as responses average, median and standard deviation. Results The first two groups of answers (F1 F4 and F5 F16) concern form. The the initial four questions directly questions subjects on whether they consider videogames equivalent to other artifact classes (F1 F3) or a new class of its own (F4). As can be seen in figure 1, both F1 and F3 are clearly above the middle point, F2 is slightly above (with a 4 median), and F4 is in the lower bound of the middle point. Analysis with Wilcoxon signed-rank tests, with a Bonferroni correction applied for level of confidence of 95% (for 3 comparisons per question), results in a significance level set at p < (= ). F1 was statistically superior to F2 and F4 (p < 0.001), though not to F3 (p = 0.446); F3 was also superior to both F2 and F4 (p < 0.001); and it is also possible to ascertain that despite its mid-scale result, F4 was statistically inferior to all other answers (p < 0.001). The second form group questions subjects whether they agree that a series of formal qualities (associated with artifact classes other than videogames) have to necessarily be a part of videogames. F5 proposes interaction as an essential quality of videogames, F6 F10 propose formal elements typically mentioned in definitions of traditional games as essential. F11 proposes simulation as a core formal quality of videogames (following the simulational 1 For ease of space and added clarity, we will only present significance levels for each of these tests, and will not specify the Z-statistic. proposals by Frasca [8]), F13 does the same for story or narrative (following narratological authors such as [14, 16, 7]), and F12 for soundtrack and F14 for visual graphics. F15 proposes as essential all previous elements and F16 none of those. Only F5, F6 and F7 have high averages, and F16 has a very low average. Analysis with Wilcoxon signed-rank tests, with the Bonferroni correction, results in a significance level set at p < ). F5 agreement is significantly superior to all other responses with (p < 0.001). F6 agreement was significantly higher than F8 to F11 and F15 and F16 (p < 0.001), F12 and F13 with (p = 0.001), though not to F7 (p = 0.507) or F14 (p = 0.010). F7 was significantly higher than F8 to F16 (p < 0.001). Finally, F16 was significantly lower than all other group replies (p < 0.001). (= In what concerns the formal perception of the medium of videogames, F1 ( Videogames are games ) and F3 ( Videogames are digital art ) presented statistically higher replies than the remaining two answers in the first group. The sample also gave a high agreement to both F6 ( Videogames consist of rules ) and F7 ( Videogames consist of challenges ), further underlining the perception that videogames are games, as according to general reply, they need to have two fundamentally game qualities (these are mentioned in most traditional game definitions, such as [18, 12, 3, 19]). This means that this sample tends to strongly perceive videogames first and foremost as games. In respects to digital art, agreement in the firt group on form is equally high, but there is no further agreement in the second group that relates to traditional notions of art (see, for example, F12 F14). F2 ( Videogames are narratives ) was moderately positive, which means there was also slight concordance with videogames being stories 80

81 or narratives, though this effect pales in comparison with the games and art responses. Again, when it comes to stories being a fundamental quality of videogames, replies were generally neutral (F13). Besides this, F5 ( Videogames consist of interaction ) had the highest agreement in the second form group, higher than all the rest. Finally, F4 ( Videogames are something new ) presented the least agreed upon answer in the first group (and its absolute value is in the lower bound of the neutral band). The second group (E1-E9) is composed of 9 sentences that concern possible expression avenues for videogames, by providing normative assertions common in game design literature (that videogames should be easy and accessible forms of hedonic entertainment) while presenting eudaimonic alternatives, such as a strive for complexity and the mediating of emotion and serious subjects. With a Bonferroni correction applied, the significance level is set at p < (= ). Of the 9 replies, only 2 stand out as positively assessed by a majority of the population, E1 and E2. Both were significantly higher than all the remaining answers (p < 0.001), and E1 is statistically higher than E2 (p = 0.003). Value was addressed in two distinct groups (V1 V6 and V7 V16), the first querying subjects on which expressive qualities they think provide value to videogames and the second which formal structures. As to the first 6 questions, V1 and V6 stand out positively. Wilcoxon signed-rank tests, using a Bonferroni correction, result in a significance level set at p < (= ). V1 is statistically superior to V2, V3 and V4 (p < 0.001), though not to V5 (p = 0.555) or V6 (p = 0.065). Likewise for V6 in respect to V2, V3 and V4 (p < 0.001). For the second group of value, however, there was no answer out of the middle bounds of the Likert scale. Discussion The vast majority of sentences did not provoke either acceptance or rejection by this population s sample. Of the 41 questions, only 10 had average values diverging from the Likert scale s midpoint. This, in itself, is a positive global result, that signals that the population sample as a whole did not present a high number of normative biases in the covered areas of this study. It did however, present a tendency to subscribe to the more prevalent axioms in the literature [5]. The survey sample seems to be neither inclined or declined to perceive videogames as something new that does not fit with these other artifact classes. The danger in not seeing videogames as a fundamentally new medium has already been discussed in detail in [5]; this result merely adds further evidence of this problematic belief by the population at large. Videogames, despite whatever inheritances they might have from narrative mediums and traditional game forms, are a new medium. Analogously, just as to define film as equivalent to drama or literature or photography would be problematic, the same should be true for videogames. Subjects also subscribed to both the relationship with games and narratives, and this can can be interpreted as an acceptance of the dual nature of videogames as both a storytelling and ludic medium, in line with some attempts at a definition of videogames [20]. However, the equivalence of videogames to games was stronger than its equivalence to a narrative-form so, to some extent, in people s minds ludologists do seem to have won the definitional debate [9, 12]. The growing consensus on videogames being art presents 81

82 its own set of challenges. First and foremost, there is the issue of incompatibility between games and art, noted by distinct personalities as being historically distinct artifact classes with different qualities. As Brian Moriarty, LucasArts designer, states in his defense of Roger Ebert s view that videogames are not art: In preparing this lecture, I plowed through a 700-page anthology on Western art philosophy, including the writings of Plato, Aristotle, Plotinus, Augustine, Ficino, Kant, Schelling, Hegel, Schopenhauer, Shaftesbury, Croce, Nietzsche, Dewey and Heidegger. I also read a deadly-boring book on 20th century art definitions, including the writings of Weitz, Dickie and Danto. Nowhere in 25 centuries of philosophy did I find a single author who regarded games or sports as a form of art [15]. So, can videogames be art and games simultaneously, as both our sample seems to believe? If so, either there is a needed revision on what constitutes art, what constitutes a game or how videogames can mediate these seemingly different artifact classes. That videogames present interaction as a fundamental quality seems acceptable, given that there is an absolute consensus that interaction is at least, one of the the distinguishing qualities of videogames in respect to other audiovisual media such as cinema (this consensus is also visible in this survey, as this is the question with the lowest standard deviation in responses). More so, there is little prescriptive bias in terms of creativity by accepting that videogames have to be interactive (whilst the same does not hold when defining them in respect to challenges, which is already a specific type of interactivity, not necessarily shared by all types of videogame experiences). Interaction, however, also presents a challenge if accepted in conjunction with the notion that videogames are an art form (see, once again, [15]). As hypothesized in [5], the majority of responses denote a look unto videogames as a means for entertainment (E1) and fun (E2), more so than a means for artistic (E4) or emotional (E5) stimulation or a means for sprouting debate on real life issues (E8). Videogames expression is thus perceived as hedonic first and foremost. Though there seems to be distinction between the concepts associated with the words entertainment and fun, signaling perhaps a different connotation to the second (perhaps even a pejorative one). High responses of agreement to V1. A fun videogame is better than one that is not fun further give credence to this view, showing the perception of value primarily lies in an entertainment perspective. For reasons explored in [5], this perception severely constrains honorable aspirations of videogames to more creative and artistic forms of design and production. Despite subjects preference for videogames function as entertainment, V6. The more a videogame makes me think about new ideas the better it is, though lower in agreement than V1, does still gather high degrees of agreement in comparison with other replies. We interpret this as subjects craving intellectually stimulating forms of entertainment, though this is bound to mean very different things to each subject. In respect to formal elements of games related to value, the absence of any reply with significant agreement or disagreement goes against expectations. Neither questions referring to more artistic aspects of the videogame V9, V11, V13 or the other technological product terms extracted in the review seem to be perceived as general measures of a game quality. Given how media and marketing focus on these aspects as arguments of value, their negligible agreement on part of subjects is puzzling. Either the effect is group specific and not generalizable to the whole population, or subjects consciously do not agree with videogames being 82

83 judged by these structural aspects. Conclusions The survey confirmed that the videogame community agrees with part of the expected normative assertions. On a whole, they believe that videogames are games first and foremost, but also narratives and a form of digital art. That interaction, rules and challenges are fundamental qualities of the medium and that videogames should be entertaining and fun and that a fun videogame or one that makes them think is better than one that is not. This is how we perceive videogames. It must be stressed that, just because a large part of the sample agrees with certain axioms, does not mean all people do, and widely divergent responses were collected in this survey. The global pattern however, is there, and it is the pattern that can help explain why game design tends to be such a conservative discipline, in opposition, we would write, with artistic disciplines in general. More extensive data collection must be carried out with a more balanced sample and by trying to experimentally verify a causal nexus between beliefs in these norms and creative output. For now, we think these results are clear in one thing: videogames nature as creative medium has not been carefully assessed, studied and communicated in the past. For videogames to grow creatively, their traits must be cautiously framed, and there must be a greater effort from scholars, creators and the media to conscientize the community that videogames are an expressive medium, and that, like all media before it, that means that there are no absolute rules. To a proper artist, everything is permitted. So, videogames can be games, stories, art, can have rules and challenges, express enjoyment and entertainment, and can be judged for being fun and making you think. They can. But, like all art before them, they can be so much more. Acknowledgements This paper was developed and financed under PhD Scholarship number SFRH/BD/75196/2010, awarded by FCT - Fundao de Cincias e Tecnologia (Foundation of Sciences and Technology). References [1] Notgames. accessed July, [2] Bogost, I. Persuasive Games: The Expressive Power of Videogames. The MIT Press, Cambridge, Massachussets, [3] Brathwaite, B., and Schreiber, I. Challenges for game designers. Course Technology, [4] Costikyan, G. Death to the games industry. Escapist Magazine, accessed July, [5] Craveirinha, R., and Roque, L. Zero lecture in game design. In Proceedings of X Simpsio Brasileiro de Games e Entretenimento Digital (Salvador, Brazil, 2011). [6] Crawford, C. The Art of Computer Game Design. Washington State University, Vancouver, [7] Crawford, C. Interactive storytelling. In The Video Game Theory Reader, M. J. Wolf and B. Perron, Eds. Routledge, London, 2003, [8] Frasca, G. Videogames of the oppressed - videogames as a means for critical thinking and debate. Master s thesis, School of Literature, Communication and Culture of the Georgia Institute of Technology, [9] Frasca, G. Simulation versus narrative: Introduction to ludology. In The Video Game Theory Reader, M. J. Wolf and B. Perron, Eds. Routledge, London, 2003, [10] Freeman, D. Creating Emotion in Games: The Craft and Art of Emotioneering. New Riders Publishing, 83

84 2003, ch. Foreword by Will Wright. [11] Gombrich, E. The story of art. Phaidon, [12] Juul, J. Half Real: Video Games between Real Rules and Fictional Worlds. MIT Press, Cambridge, Massachussets, [13] Koster, R. A theory of fun for game design. Paraglyph Series. Paraglyph Press, [14] Laurel, B. Computers as Theatre. Addison Wesley, [15] Moriarty, B. An apology for roger ebert. In Game Developers Conference (2011). [16] Murray, J. H. Hamlet on the Holodeck - The Future of Narrative in Cyberspace. The Free Press, New York, [17] Poole, S. Trigger Happy. Arcade Publishing, New York, [18] Salen, K., and Zimmerman, E. Rules of Play: Game Design Fundamentals. The MIT Press, Cambridge, Massachussets, [19] Schell, J. The Art of Game Design: A book of lenses. Morgan Kaufmann, Aug [20] Tavinor, G. The art of videogames. New directions in aesthetics. Wiley-Blackwell,

85 Hábitos de Jogo em Dispositivos Móveis pelos Jovens Portugueses Ana Amélia A. Carvalho FPCE, Universidade de Coimbra R. do Colégio Novo Coimbra Inês Cardoso Araújo FPCE, Universidade de Coimbra R. do Colégio Novo Coimbra Resumo Os jogos para além de serem do agrado das crianças e dos jovens têm vindo a conquistar cada vez mais os adultos. Além disso, a importância crescente atribuída aos jogos no desenvolvimento de competências e de variadas aprendizagens tem levado à sua integração em contextos formais de aprendizagem. Num estudo que está em curso sobre os jogos e a aprendizagem, inquirimos os alunos portugueses, do 2º Ciclo do Ensino Básico ao Ensino Superior, para conhecer os seus hábitos e as suas preferências quando jogam em dispositivos móveis. Na sondagem que realizámos obtivemos 2060 respostas, sendo a amostra constituída por 1344 jogadores portugueses que jogam em dispositivos móveis. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Palavras-chave Jogos; Mobile Learning; Dispositivos Móveis. ACM Classification Keywords K.8.0 [Personal Computing]: General Games. 85

86 Introdução Os dispositivos móveis têm uma elevada aceitação pela população devido à sua portabilidade e crescente melhoria das suas funcionalidades, permitindo o acesso à informação a qualquer hora e em qualquer lugar. Estar conectado passou a ser uma exigência e uma necessidade na nossa sociedade. Com os dispositivos móveis os jogadores podem jogar em qualquer lugar e a qualquer hora, sós ou com outros online. Reconhecendo a possibilidade dos jogos desenvolverem competências e de envolverem os jogadores na aprendizagem, concebemos um projeto que alia dispositivos móveis, jogos e aprendizagem. Na fase inicial pretendemos identificar os hábitos de jogo e as preferências dos jogadores, alunos do 2º ciclo do ensino básico ao ensino superior, em dispositivos móveis. Numa fase posterior, pretende-se analisar os jogos que os alunos gostam de jogar, identificar os seus princípios de aprendizagem e criar atividades educativas com os mesmos princípios, dado que learning is most powerfull when it is driven by passion [10]. O jogo e a aprendizagem A importância do jogo na aprendizagem e como elemento de integração sociocultural foi analisado na obra Homo Ludens, de Johan Huizinga, na década de 40. Por sua vez, a evolução gráfica que os videojogos tiveram na última década tornaram-nos muito atrativos para jovens e adultos. Segundo Koster [6] o nosso cérebro é um devorador de padrões. Desde a nossa infância que adquirirmos os vários padrões que nos vão ser úteis no futuro através do exercício, da repetição e do jogo. Quando dominamos esse padrão, aborrecemo-nos e abandonamos essa atividade, já que não tem nada de novo. Isto ocorre pela própria natureza humana em querer preparar-se o mais possível para sobreviver. Se não traz nada de novo é porque já está aprendido e é necessário encontrar algo de novo para aprender, investindo as energias nesta nova atividade. É neste contexto que Koster define o jogo como: Games are puzzles to solve, just like everything else we encounter in life.( ) We learn the underlying patterns, grok them fully, and file them away so that they can be rerun as needed. The only real difference between games and reality is that the stakes are lower with games. [6]. Também James Paul Gee refere que Good video games are nothing but well-designed problem-solving spaces with copious feedback, good mentoring from the game s design and associated fan communities, and a win state [6]. É nesta última definição que encontramos os ingredientes que fazem dos jogos excelentes instrumentos de aprendizagem: resolução de problemas, o feedback imediato, uma boa orientação (tutorial) e a sensação de vitória quando se atinge algo. O jogo tem, por tudo isto, caraterísticas que os professores ambicionam para as suas aulas. Um estudo realizado nos Estados Unidos da América a 309 professores do ensino secundário pela We are Teachers, chegou à conclusão de que a grande maioria dos respondentes (81%) sentem os seus alunos mais 86

87 empenhados nas atividades onde são utilizados jogos [12]. Jogar em dispositivos móveis Já passaram 9 anos desde que Prensky (2004) escreveu: people, all over the world, are walking around with powerful computers in their pockets and purses. The fact is they often don t realize it, because they call it something else. But today s high-end cell phones have the computing power of a mid-1990 s PC.[9] Hoje essa diferença é mínima, um smartphone realiza as principais ações que necessitamos: consultar o , aceder à Internet, acompanhar as redes socias, ler um e-book, jogar, etc. Os dispositivos móveis estão presentes na vida da grande maioria dos nossos jovens. Os telemóveis e os smartphones acompanham-nos, tendo disponíveis inúmeras funcionalidades usadas com frequência pelos alunos como o acesso à Internet, SMS s, calendário e alertas [3]. No entanto outros dispositivos móveis estão a ter cada vez mais procura como os tablets que se prevê que até 2015 ultrapassem as vendas de computadores pessoais [1]. A tecnologia mobile começa a ganhar importância em contexto educativo estando já recomendada pela UNESCO aos seus parceiros [11]. É mais barata e mais fácil de transportar, permitindo a realização de uma vasta quantidade de tarefas que em muito enriquece a educação. Aliar os jogos a esta tecnologia que já está presente na vida dos jogadores poderá ser uma grande mais-valia para a qualidade da aprendizagem já que with games, learning is the drug. [6]. 1 Estudo No âmbito do projeto Dos Jogos às Actividades Interactivas para Mobile-Learning 1 foi desenvolvido um questionário para se averiguar hábitos de jogo e preferências dos jovens portugueses, relativamente aos jogos que jogam em dispositivos móveis. Este estudo abrange alunos do 2º Ciclo do Ensino Básico a alunos do Ensino Superior (Licenciatura e Mestrado). Pretende-se numa segunda fase analisar os jogos que os alunos dos diferentes níveis de ensino preferem para se identificar os princípios de aprendizagem que estão subjacentes e com base neles criar atividades educativas. METODOLOGIA Nesta fase inicial do projeto foi realizado um survey [2] e a técnica de recolha de dados usada foi o inquérito por questionário. O questionário "Jogar em dispositivos móveis é composto por quatro partes: Caraterização do aluno; Caraterização de hábitos de jogo; Preferências do jogador e Idealização de um jogo. O questionário depois de avaliado por peritos foi adaptado aos diferentes públicos-alvo, tendo sido criadas quatro versões, correspondendo cada uma aos quatro ciclos de ensino: 2º CEB, 3º CEB, Ensino Secundário e Ensino Superior. O questionário foi disponibilizado online, utilizando a ferramenta Formulário disponibilizada pelo Google Drive e divulgado o seu url junto de professores para que solicitassem a alunos o seu preenchimento. Projeto com a referência PTDC/CPE-CED/118337/2010 financiado pela FCT. 87

88 A divulgação foi feita através dos elementos que compõem o projeto, bem como através de grupos em redes sociais de docentes. Foi aprovado pela Direção Geral de Educação o que nos possibilitou a sua divulgação a nível nacional junto de escolas por . a idade normal de cada ciclo, mas em número reduzido. (Figura 4). Figura 1 - Distribuição das respostas por ciclo de ensino e por sexo. Quem não joga? É curioso verificar que em anos de maior exigência escolar se verifica (Figura 2) maior percentagem de não jogadores, como aos 15 anos (9º ano), 18 anos (12º ano) e 22 anos (mestrado). No entanto, a partir dos 30 anos a percentagem sobe para perto de 70%, talvez porque as we get older ( ) games are viewed as frovolity. [6] CARATERIZAÇÃO DA AMOSTRA Obtivemos até junho um total de 2060 respostas, sendo a maioria proveniente do questionário relativo ao ensino superior (n=1091; 53%), conforme Figura 1. No total das respostas constata-se que cerca de 35% [n= 716] dos respondentes não jogam, sendo 14% no 2º CEB, 30% no 3º CEB e no Ensino Secundário e 43% no Ensino Superior. Em todos os ciclos são os sujeitos do sexo feminino que menos jogam como podemos ver na Figura 3. Figura 4 - Frequência de respostas por idade e ciclo (n=1344) Em termos de composição da amostra por situação escolar verificamos através da Tabela 1 que é no 12º ano que existe menos respostas (3%), mas os anos restantes possuem algum equilíbrio. Figura 2 - Distribuição dos alunos que jogam e não jogam por idade. Figura 3 - Distribuição por sexo dos alunos que jogam e não jogam. Em termos de idades a amostra (n= 1344) é composta principalmente por alunos com idades medianas de cada ciclo de ensino, existindo alguns que ultrapassam 88

89 Figura 5 - Distribuição da amostra por sexo entre os vários ciclos de ensino Ano escolar n % 5º ano 77 5, º ano ,0 7º ano 114 8,5 8º ano ,1 9º ano º ano º ano º ano 36 3 Licenciatura Mestrado Total ,7 23,6 Tabela 1 - Frequência absoluta e relativa da amostra por escolaridade. Em termos de distribuição por sexo (Figura 5), verificamos que na nossa amostra predomina o sexo masculino no ensino básico e secundário, situação que se inverte no ensino superior. Apresentação e análise de resultados Equipamentos que utilizam para jogar Em média cada aluno utiliza 2,8 dispositivos móveis para jogar. No entanto se visualizarmos a média por ciclo de ensino, verificamos que esta vai diminuindo: 2ºCEB 3,4 dispositivos, 3ºCEB 3,2 dispositivos, ensino secundário 2,6 dispositivos e no ensino superior 2,5 dispositivos. Ao identificarmos os equipamentos que utilizam para jogar verificamos que o computador portátil é usado pela grande maioria dos alunos, no entanto, é visível o decréscimo ao longo dos vários anos de escolaridade (Figura 6). Outro dado a assinalar é a evolução que observamos no smartphone, que no 2º CEB (5º ano) é 15 o menos utilizado, passando para o segundo mais utilizado pelos alunos do ensino superior (mestrado). Figura 6 - Distribuição dos dispositivos móveis usados para jogar por ciclo de ensino. Ao analisarmos a Figura 7 verificamos que há diferentes preferências entre o sexo feminino e masculino. Em cada ciclo o sexo feminino utiliza mais o Telemóvel, o Smartphone (exceto no 2º ciclo), o Tablet (exceto no 2º ciclo) e a Nintendo 3DS (exceto no 2º ciclo) que o sexo masculino. De realçar que no 3º ciclo para o sexo feminino o telemóvel ultrapassa o computador portátil em termos de utilização para jogar. É na PSP que verificamos a maior diferença entre os sexos no 2º e 3º ciclos alcançando 76,2% e 62,6% no caso masculino (M) respetivamente e 34,9% e 47,4% no caso feminino (F), a partir do ensino secundário acaba por se aproximar, atingindo os 26,2% (M) e 19% (F) no ensino Superior. 89

90 Figura 7 - Distribuição dos dispositivos móveis usados para jogar por ciclo de ensino e por sexo. Hábitos de jogo TEMPO UTILIZADO PARA JOGAR POR SEMANA Em termos de média 2 global os respondentes referem utilizar 5,2h por semana para jogar. No entanto se analisarmos por ano escolar (Figura 8) verificamos que o tempo médio utilizado sofre alterações ao longo dos vários anos escolares. É de referir que é nos 8º e 11º anos que os alunos referem utilizar mais horas para jogar em média (8h e 6,9h respetivamente). Estes são anos intermédios de ciclo, talvez por isso os alunos se sintam mais livres para jogar, ou os encarregados de educação permitam a utilização dos dispositivos móveis por mais tempo por não ser um ano tão decisivo para o percurso escolar. Figura 8 - Distribuição das horas gastas em média por semana e por ano de escolaridade. Ao analisarmos o tempo médio de jogo por semana entre o sexo feminino e masculino por ciclo de ensino (Figura 9), verificamos que o sexo feminino declara que em média utiliza 3h por semana para jogar nos vários ciclos, enquanto o sexo masculino declara utilizar muito mais tempo e com grandes oscilações de ciclo para ciclo. Note-se que atinge as 9h no 3º ciclo e diminui até às 5h no ensino superior. 2 Para a sua obtenção foi calculada a média das respostas tendo sido atribuído a cada item os seguintes valores: menos de 1h = 1h; de 1h a 5h = 3h; de 6h a 10h = 8h;de 11h a 20h = 15h; Mais de 20h = 20h. 90

91 Figura 9 - Média de horas de jogo por semana distribuído por sexo e ciclo de ensino. NÍVEL DE DIFICULDADE DO JOGO Em termos de dificuldade de jogo destaca-se a indicação maioritária pelo nível moderado, na Figura 10. Esta clara preferência revela que a maioria joga jogos que se adequam às suas capacidades e que conseguem resolver sem grande dificuldade, sem serem demasiado fáceis nem difíceis, o que poderia desmotivar. Esta opção está de acordo com a teoria do fluxo proposta por Mihaly Csikszentmihalyi [4]. No entanto, é de assinalar o facto de que é no 2º ciclo que mais alunos responderam que o jogo é muito difícil. Esta situação é curiosa, talvez se explique pela ansiedade em crescer característica destas idades e que os leva a imitarem os mais velhos na escolha do jogo. Figura 10 - Distribuição do nível de dificuldade do jogo mais jogado por ciclo de ensino. Denotamos também que é o sexo masculino que arrisca mais em jogos que consideram difíceis e muito difíceis, em todos os ciclos (Figura 11). É de salientar que o inverso ocorre com o sexo feminino, preferindo jogos fáceis ou muito fáceis. Figura 11 - Distribuição do nível de dificuldade do jogo que mais jogam por sexo e por ciclo de ensino. 91

92 JOGAR SÓ OU COM OUTROS ONLINE Existindo hoje em dia uma panóplia enorme de jogos disponíveis, que beneficiam da colaboração entre jogadores, inquiriu-se se jogam sozinho ou com outros online. Constata-se (Figura 12) que os sujeitos do 2º CEB preferem jogar sozinhos (53,3%), no entanto nos outros ciclos a diferença entre os dois tipos esbate-se, até que no ensino superior aumenta significativamente esta preferência de jogar sozinho para 71,3%. Estes dados estão em consonância com as preferências indicadas em estudos de mercado para jogos mobile [5]. Os adultos preferem jogos que possam jogar durante 5 a 10 minutos e que depois possam retomar algo, sem que tenham de pensar mais no assunto. Os jogos online com outros obrigam a uma regularidade para a qual não têm muita disponibilidade e poderão ser mais exigentes cognitivamente, não dispondo os adultos de tempo disponível. No entanto se analisarmos a sua preferência por ano de escolaridade, verificamos oscilações entre o 3º CEB e o ensino secundário (Figura 13). Terão estas oscilações relação com os dramas habituais desta etapa da adolescência onde as amizades se criam e rompem com facilidade? Figura 13 - Distribuição da preferência por jogar só ou online por ano de escolaridade. Já se analisarmos por sexo esta distribuição verificamos que as respondentes preferem, seja qual for o ciclo de ensino, jogar sozinhas, enquanto que os do sexo masculino preferem jogar com outros online do 2º CEB ao ensino secundário, preferindo depois no ensino superior jogar sozinho (Figura 14). Figura 12 - Distribuição da preferência por jogar só ou com outros online por ciclo de ensino. 92