For the purpose of this case study, I will focus on the user interface of the massive multiplayer online roleplaying game, World of Warcraft. WoW’s interface is particularly useful to use for  applying an ANT + Hypertext frankentheory because it is modifiable by players and is even encouraged by Blizzard. Although many interfaces are capable of being modded, not all game developers actually encourage the use of modifications to the user interface within their games. Blizzard also implements rules within the terms of service about modifications, particularly in regard to compensation. Players who create and distribute modifications cannot sell these modifications for in or out of game currency.  Having a visual reference for the user interface provides an image in order to demonstrate the narrative of play. I’ve chosen to look at my own interface because ANT privileges narrative, as well as emphasizes the actor’s ability to narrate their decisions. It also allows me to point out specifically what the hypertext present in the interface allows me as the player to do/learn.

I’ve chosen to focus on two specific interface set ups that I use based on two types of play that I engage in– raiding and pet collecting/battling.

Pet collecting and battling is focused on little non-playable companion pets that you can collect, catch, and battle against within the game. The style is similar to Pokemon, but it is all integrated into the larger world of the game and pets can be bought from NPCs (non-playable characters) or received as a reward for completing certain achievements in addition to the catch/battle system.

Typical interface I use for pet battling/collecting.

Raiding is when a group of players (10-25 players) enter into a raid (large dungeon) in order to defeat a certain number of bosses (very strong monsters). This usually requires strategy work between the players with each player having particular role(s) they must fulfill in order for the encounter to be successful. Most raiding groups are composed of guild members and they enter into the raid a few times a week at scheduled times.


Typical interface I use when raiding as a Moonkin (damage dealer).


Literature Review

Currently, most research that discusses MMOs (WoW and Everquest, in particular) in relation to networks is specifically focused on the social network–the socializing between players. This happens between players in a multitude of ways, such as through social networks created through guilds which also is connected to the interface (Williams et al). Williams et al were interested in how the interface impacts social interactions, however their focus in the study was on social capital and very little discussion is given to how the interface ended up contributing to their study. However, they do note that in addition to other features such as game mechanics and player’s choices, that the “governing computer codes were ultimately foundational rather than entirely imposing. By analogy, we find that playing WoW is as social as a team sport, which has its own rules…” (357). Shen looked at network patterns and social architecture within Everquest II, another popular MMO. Shen found that “network patterns revealed that the social architecture of the world was quite effective in shaping the structure of interactions, as the involvement in various social networks” (672). However, Shen also found that many players preferred solo play despite the the built-in mechanisms of the game that encouraged social play. The interface is clearly part of this as parts of the interface encourage social interaction between players such as the Looking for Group tab so that players can queue for dungeons where each player is placed randomly with other players to complete dungeons or raids and also the Social tab that allows players to see a friends list. These are only two examples from World of Warcraft, but there are many others. Targett et al’s article, “A Study of User Interface Modifications in World of Warcraft,” discusses a survey study conducted by the team of researchers in 2007 and again in 2011 that was interested in the World of Warcraft users that were creating and using user interface modifications or UI mods.  The goal was “to study the effect that user created interfaces have had on WoW and its community of users” (Targett et al).  In order to study these effects, the researchers created an online survey that specifically looked at four aspects of the UIM (user interface modification) community. These four aspects were “(R1) the backgrounds of its members, (R2) their attitudes towards modifications and the community itself, (R3) their use of UI modifications, (R4) the characteristics and motivations of users who create and share modifications” (Targett et al). Despite the focus on interfaces, Targett et al is not specifically interested in the interface as a network. One of the reasons for my focus on the interface as a network is because of this gap in the research in which the interface of World of Warcraft is examined as a network. (The other is that interfaces are just so interesting!).

Theories: ANT & Hypertext

I will be using Latour’s Actor-Network Theory and Joyce’s discussion of hypertext for this analysis. Actor–network theory (ANT) is an approach to social theory and research, that comes from sociology. Latour writes that for too long sociologists have been practicing a ‘social explanation’ discipline. Instead, he calls for a sociology of associations, which is focusing the tracing of associations, or “social does not designate a thing among other things, like a black sheep among other white sheep, but a type of connection between things that are not themselves social” (5, emphasis his).He believes ANT provides a way of tracing these associations in a robust way, a way that requires a researcher to continue to find more and more traces, ultimately that can never all be found. I chose to use Joyce’s discussion of hypertext because as he says that “hypertext readers not only choose the order of what they read but, in doing so, also alter its form by their choices” (19). While I don’t think he was originally considering the reading the hypertext of an interface of a game, I think his point here fits better within a game system than a typical text, even an online text that follows typical written conventions. The interface hypertext communicates player choices to the game and the game responds based on the player’s choices. It acts as a much less linear space than typical online texts like web articles that tend to mimic traditional print media.

While I am using both ANT and Hypertext theories for this case study, it is primarily an ANT reading, showing how hypertext, particularly in the case of this interface, helps to illuminate the connections that are taking place between the interface, information, and player. I’ll be doing an ANT reading of two different interface images. One will be from the pet battling interface I use and, another will be a significantly modified interface I use for raiding. I’ll be talking from my perspective as my interface, both the typical interface and then the modifications I use when raiding and why. The hypertext theory will help me discuss the action bar and the levels of readings of the interface that the player takes on simultaneously as part of play.

Pet Battling + ANT/Hypertext

The pink surrounds the default interface that WoW provides. 

I highlighted the default “bar” in World of Warcraft that players have access to from the beginning in order to discuss what players first encounter as well as how this the action bar changes depending on choices the player has made during character creation as well as how new spells are added as players level up. For instance, the left side of the bar are made up of druid damage spells–anytime I click the image which acts as hypertext, and it causes my character to perform an action. There are two layers of hypertext here; the first is that hovering over the image provides a tooltip that allows the player to read what the spell does. The second layer is the actual clicking which causes the spell to be used.

Tool tip that appears over the 3rd spell listed, Frostbolt, when a player hovers their mouse cursor over it, but does not click. 

The tool tip allowed me as a player to become familiar with my move set when I first began playing. I still use the tool tips when I create a new character that is a class I am not familiar with, since I typically play a druid. The spells become like nodes within the network that the actor has some agency over as the player can move the spells within the bar. The player can also expand the bar through the settings under  interface settings. This allows the player to places on the actionbar for more spells. In the picture of my interface, I have spells above the default bar as well as a bar on the right hand side.

Blizz Bar Menu
Game interface settings: ActionBars

The following is a full markup of what is taking place for me as the player in this particular screenshot.


The screenshot shows me preparing to engage a little spider in battle. Although the spider is not visible, I found him through looking at the top right hand corner of my interface. It is difficult to see in the screenshot, but there are green pawprints moving around the map. Hovering over these (hypertext!) shows me the name of the pet, “spectral spider”.  This allows me to more easily find the pet and click on it in game to engage it in battle. My next step is to look in  my bag for a specific item that gives my pets 10% more experience when battling that I acquired through an achievement.  In order to do this, I click the bag hyperlink which brings up the small box in green, or my backpack. I am able to hover over the items for more information about in case I don’t know. Clicking the hat makes it appear on my character and a “buff” appears in the top right hand corner of my screen, indicating my pets will now receive the buff. The part of the actionbar that the player has little agency over in terms of whether it appears on the bar or not is the game related hyperlinks section which brings up smaller interfaces for specific types of gameplay, although players still have some agency because they can choose to click on these or not.

Clicking on one of those game hyperlinks (the tiny little horse) brings up the interface box in the screenshot in orange: the pet and mount journal interface. This interface has tabs at the bottom, including “pets” that allows me to see all of my pets in one place as well as move them into battle spaces. It also allows me to check their spell list (and hover over it for tooltips). There is also a “rivals” tab, which is a modified part of the interface. It is from an addon created by another user that is helpful when battling new opponents. It allows the player to see what move sets rival NPCs have so that the player can make strategic decisions on which pets to use.

A lot of the nodes, although separated, have to be engaged by the player simultaneously, and the pattern for reading that the player has is likely different from other players. As players become more experienced, they will likely look at the tool tips less frequently, meaning that the content travels differently through the network because of this. The player’s choices also show how the network can grow or shrink, depending on the activities that the player chooses to engage in. For example, the pet battling I just described can be a social feature of the game, but for the most part functions as a single player activity. However, raiding, the next interface I will be looking at is quite different. My raiding interface allows me to collaborate with 9 other people as we work together in Dragon Soul (a raid/dungeon) in order to defeat Deathwing.

Raiding Interface: ANT/Hypertext


In a raiding situation, the network grows significantly because not only does the player (actor) exist within their own network of spells, bags, and item set, but players become part of a larger social player network as well. For example, most raid groups, like mine, operate within a guild setting.  This means that we have a social group that interacts, and while some raid members are also guildmates, not all of our guildmates raid with us. That’s limited to 10. We have to use the calendar feature to link our schedules so we can work together at a specific time and date.

There is also the network created among players inside the actual raid. The room is sealed off when the boss battle begins and players must coordinate, moving around the room to avoid damage and provide help to other players. Depending on the fight, this can become very chaotic. This is what is happening in the background of the screenshot and completely ignores the interface that players must also pay attention to as they carry out the boss fight together. All of these layers are important for how ANT defines the network–as the researcher we want to see as much of the tracing of connections taking place as possible. The interface provides the information to the player that allows them to make decisions about how to proceed throughout the fight. Because of the growth of the network, there are so many nodes, but perhaps of different sizes. I’d say that major nodes include the NPC Boss that the players are fighting as well as the 10 players.  But then each player basically has a smaller set of nodes that branch off from their larger node. I’m realizing this is actually very difficult to explain, so I’ve made visuals to explain instead.

Node Layer 1
Player characters interact with the boss as well as each other depending on their responsibilities. For instance, the tank interacts directly with the off-tank as they collaborate together to take hits from the boss.  Healers deal very little with the boss, but interact heavily with other characters through healing spells. 
While the player is working to collaborate within the network of actors present in the raid situation, the player also has to navigate the network of the interface that they must read in order to be a valuable member of the team. This includes juggling multiple parts of the interface (nodes) and recognizing where the nodes intersect.  For example: for a healer, health and debuff bars on the screen are important because healers watch these to know when to use their spells, while also keeping an eye on their own mana pool. 


All of the nodes are mediated through hypertext; I say this because the interface requires the player to use hypertext in order to interact with the objects of the game as well as with other players and the hypertext allows players to make strategic decisions for play. They can see tool tips to help them decide whether to use a spell at the moment or not, for instance. In the game, the hypertext is not necessarily added in an element that moves the reader to another page like in some articles online, but rather adds layering of information on top of what the player is already seeing. The hypertext allows multiple levels of reading to take place at the same time.

What is moving in the network is the results or content of player’s choices. This is most easily seen in the form of choosing to use certain spells, since the avatar makes specific movements with animations for different spells that are used. In terms of ANT, the actor (players) have a lot of agency as they create movement throughout play for their avatar as well as move other players, items, spells, and much more throughout gameplay.

The content that is traveling through the network changes depending on specific circumstances. In the raiding example, the content is constantly shifting as players react to boss moves, to the successes and failures of other players.  For example, for a healer, the health bars are constantly in motion moving up and down based on damage taken and healing received. The level of the bars means different things for different players.  For instance, low health on a ranged damage dealer that will not be taking damage can wait for heals. Low health on a tank taking damage almost constantly cannot, so healers must strategically decide who to heal and what spells to use to give the right amount of heals without stressing their mana pool, because running out of mana could mean the fight ends in failure.

IV. Conclusion

I think I underestimated how complicated doing an ANT reading would be. I realized that I had a lot of explaining to do–perhaps more than was really possible to do without a significantly longer piece in order to trace all of the connections, especially to explain them to readers who may not play games, or who may not be familiar with roleplaying game boss fights or dungeons. It creates a lot of layers of complexity that I don’t think I fully appreciated. I also didn’t realize until I really starting writing things out how many layers of networks that players are working through as they play, especially in more social types of gameplay like raiding.

Despite this, I do think I was able to show the tracing of connections happening within the network of the WoW interface and how this interface changes depending on the choices of the player and the type of gameplay that the player is engaged in. The diagrams I think also capture the complexity of the layering of networks that the interface has for the player, pulling the player/reader to look at multiple layers of information and to read and interpret a lot of information at once. Many of these things become easier as players continue to play and become familiar with how the network/interface functions, but it can be overwhelming.

Works Consulted

Blizzard Entertainment. World of Warcraft: Warlords of Draenor. Activision Blizzard, 2012. Online Video Game.

Joyce, Michael. Of Two Minds: Hypertext Pedagogy and Poetics. Ann Arbor: U of MI P, 1995.

Latour, Bruno. Reassembling the Social. Hampshire: Oxford University Press (2007). Print.

Shen, Cuihua. “Network patterns and social architecture in Massively Multiplayer Online Games: Mapping the social world of EverQuest II.” New Media & Society (2013). Web. 20 March 2016.

Targett, Sean, Victoria Verlysdonk, Howard J. Hamilton, and Daryl Hepting. “A Study of User Interface Modifications in World of Warcraft.” The International Journal of Computer Game Research. 12.2 (2012): n. pag. Web. 06 Feb. 2016.

Williams, Dmitri, et al. “From tree house to barracks the social life of guilds in world of warcraft.” Games and culture 1.4 (2006): 338-361. 20 March 2016.