We have a brand new showcase interview for you. Daniel Marx from Osmotic Studios talked to us about Orwell: Keeping an eye on you, and the challenges of putting socially relevant topics into games.
Orwell: Keeping an eye on you lets you look through the eyes of Big Brother. You have to investigate the lives of citizens to find those responsible for a series of terror attacks. But, be warned, the information you supply will have consequences…
OSMOTIC: Hi, my name is Daniel Marx and I am the Game and Narrative Designer at Osmotic Studios. We’re a three person game development studio based in Hamburg, Germany. Our first game, Orwell, has been released as a five-episode weekly series in October and November 2016.
OSMOTIC: We use to describe the game as a “privacy invasion thriller”: You, as a player, have been recruited for the Orwell project, a massive surveillance machinery capable of putting websites, phone calls, chats, and even any phone or PC located or occurring in the fictitious country The Nation under observation. The whole game takes place inside the Orwell interface, resembling other operating systems.
There has been a recent terror attack on a public plaza in The Nation’s capital, Bonton, and as “investigator” it becomes your job to find out who planted that bomb and why. Suspects are found quickly, but it falls to you to look at their online data for potentially valuable pieces of information, which you can upload into a suspect’s personal profile so your counterpart, the “adviser”, can interpret those and act upon them. Or you could just ignore passing on certain data, because not everything you will find is right… or should be stored in Orwell in the first place. But be aware that based on whatever you may decide, the story will progress differently: Somebody might get arrested or walk free, or a following bombing may or may not be prevented.
OSMOTIC: I believe dealing with socially relevant issues is very healthy for games, both economically and as an art form or form of cultural expression. Game developers have been striving for decades for games to become accepted as a culturally relevant medium, and in order to get there we have to show or medium is capable of dealing with a large variety of topics relevant to our everyday life.
This does in no way mean games not dealing with these topics are irrelevant or should even seize to exist. Games are very well allowed to “just” be fun. They only can and should occasionally be more.
A reason for an increased number of games dealing with topics like these might be an improvement in accessibility to both tools and education for game development. A consequence is that more and more people altogether and with highly diverse backgrounds are breaking into the games industry, leading to a great range of different games.
OSMOTIC: It took quite some time for Orwell to evolve into the final game. As you suggest, it has been highly inspired by the Snowden revelations around the time we were thinking about founding a game company and would most likely not exist without them.
The idea we started from was the very rough concept of making a game entirely taking place on a PC desktop and exclusively being told through files, chats, websites and such. This game was lacking a topic though, but digital surveillance was a natural fit. It allowed to give the player lots of documents from various people to take a look at and tell stories about them. And, more importantly, allowed to raise complex moral dilemmas, which was a concept we were very fond of in games like Papers, Please (another huge influence for the game) and This War of Mine.
OSMOTIC: In the first instance we see Orwell as an entertainment product, but that doesn’t mean it shouldn’t have anything to say. Our ambition came attached to the topic: We wanted to reflect on the ever growing human conflict between being safe and being free, and what both of this concepts mean in the information age.
With surveillance a core problem often is that even if you’re aware of it happening, it is a concept too abstract to really get a grasp on possible consequences for yourself. It’s easy to say it doesn’t concern you because you have done nothing wrong. So if we were to make a game to tackle digital surveillance, at the very least we wanted to prompt players to make up their minds about their stance towards it and things that might happen to them because of it.
OSMOTIC: It’s easy to trip into preaching to your audience if you’re set on making a point, and I am not one hundred percent sure we have successfully avoided it throughout the game, but we definitely meant Orwell to be non-preachy.
The first thing to look out for is to not give a one-dimensional view on matter covered. When you’re taking on a subject chances are you have an opinion of your own on it since it is important enough for you to write about it. This means you will be biased, but that is fine as long as you are aware of it. You don’t have to be neutral, you just shouldn’t be one-sided. Take a step back and look at the issue as a whole. What defines it, why are there clashing sides? What are the opposing stances on the matter and why do people have different views? The resulting story or scenario will be much more interesting for it.
Secondly, never judge or even punish players explicitly for a choice you have presented to them. If you challenge them with a decision that is clearly right or wrong it’s a boring one to begin with. If you give them a choice not clearly right or wrong, don’t scold them if they pick the action you do not agree with or you wouldn’t have done. Instead confront them with believable consequences to their actions and let them pass the judgement over whether they did the right thing. Another gain of this approach is that your point will stick much more if players come to the realization you want them to have themselves. There’s a chance it won’t work out, but so is there with downright preaching, only that the risk is much higher for the latter.
OSMOTIC: The design of Orwell was a very long and highly iterative process. We started out with a concept pretty close to Papers, Please where you had to take a look at the digital documents of certain persons and decide whether a person might be dangerous or not – a simple binary decision. Having written that down it dawned on us this would not only be a Papers, Please clone, but just not cut the feeling of investigating people and giving away their data.
So we came up with the idea of making every sentence and picture in a document a piece of data that could be used to answer specific questions the government you were working for would be handing you. We built a prototype out of this concept and conducted a test session with some hand-picked testers who had never seen or heard about the game. The feedback from that was what players were now spending the majority of the play time searching for the right data to answer the questions. Even though there were multiple options of answering a question the players did not recognize them because they didn’t know whenever multiple options were around. Also the questions felt artificial and made the game feel “on rails” because you were effectively ticking off a to-do list. To sum up, the feeling of handing over data was now present, but the game was a game of searching when we intended it to be a game of tough choices.
In the next iteration we decided to cut the questions from the game altogether and went for a solution to let the player progress more freely. Instead of having any sentence and picture as a potential piece of information we reduced the number to a few per document and highlighted them for the players. This way we were able to eliminate the searching element completely and were able to juxtapose contradicting choices (now known as “conflicts”). Even more so, every information became a choice in itself whether to upload it or not. We implemented these changes, had players test them again and were very pleased when the game was received as much less linear and guided and people were making up their minds about the choices. From there on out the game was far from being finished, but the core mechanic was in place and we only kept changing details.
OSMOTIC: You cannot make a game that isn’t entertaining in some way, else it’s just going to be a bad game. And a bad game is sure to not transport any message at all.
To achieve being entertaining we put the story first: It had to have interesting characters, it needed to be tense, it should surprise players at every turn to keep them engaged. This is why we framed the story like a classical murder mystery (“whodunit”), with a closed cast of intriguing characters that all hold secrets. To our delight this went along perfectly with a scenario of spying on people and discovering clues, and building a picture of the whole story piece by piece. To make things more interesting and have something to investigate, we also made most of the characters know and have distinct relationships with each other, and share a common past inside and outside the activist group the player needs to uncover.
Still the story had to be constructed in such a way it would convey the meaning we wanted it to, putting the player into the conflict of safety versus freedom. In order to accomplish that, we pitted two fractions against each other able to carry that conflict: The restrictive and conservative government of The Nation which controls Orwell and a freedom-loving activist group called Thought, to which most of the suspects belong. And, without spoiling too much, it all needed to resolve in such a way the core dilemma of the game would come up with the player needing to take sides.
OSMOTIC: Before the release were pretty much on edge about bringing out a hot topic game in times of a troubling political climate. Personally I was worried people would hate our game because of its political nature, or that our game was being misinterpreted or misused for causes we would not want to support. I can’t say it never happened because it did (e.g. early on we were confronted by someone claiming we would make pro-surveillance propaganda), but it was only in such a minor way that left our previous concerns mostly ungrounded. The vast majority of the players liked the game for what it was, praising it for making them think about the topic, which was what we set out for all along.
OSMOTIC: Designing the narrative structure and the levels (episodes) was just as much an iterative process as was designing the overall game. In Orwell, narrative branches are often results not of a singular decision but rather a multitude of information the player has uploaded to Orwell.
As a result we needed to try out a lot, ideally in little time, and keep track of all pieces of information and the documents they could unlock visually. In order to do that efficiently we needed a tool to allow for clear level flow design combined with the ability to change and improve existing designs quickly. We also explicitly wanted to skip an extra scripting phase of implementing the level designs as playable in-game content, so the content of that tool had to convey customizable meaning and need to be exportable and readable within Unity, the engine we used.
We found all of these requirements met by articy:draft. As soon as we started using articy:draft it became an integral part of our development pipeline. As a team consisting of three persons all working in their own particular area of the game, programming, art, and game design, using articy:draft mostly fell to me for designing levels, structuring documents and writing, while Mel created document layouts in Photoshop, and Michael put it all together and made things work in Unity. This three-fold approach also means most of the time we do not have to work on the same files, which in turn means less coordination is necessary, resulting in less conflicts and downtimes.
OSMOTIC: We used separate tools for writing and planning level designs before, but we switched to articy:draft relatively early on during development, because the process of having to use purely visual graph editors for design, writing text-only documents with markups without having the possibility to implement a basic level of scripting within them or cross reference between documents and game flow appeared cumbersome.
OSMOTIC: As said in the previous questions, we needed a tool that offered great overview over level flows as well as the option to iterate on the design of these flows fast, and allow somewhere for the actual writing to be stored. articy:draft offered all of that separately from the engine and in one place, plus the option to get the flow logic and writing content into the game in no time. The ability to jump between our level flows and the actual documents referenced within them was an absolute benefit compared to a solution using another tool for each of that tasks.
Of great importance for us also is the feature to create custom templates for flow nodes and interpret them in-engine the way we need them to. Almost anything that happens in Orwell is a direct result of something the player did, and we use game flows to represent this logic of cause and effect where every flow node can take on the role of a condition to wait for, an action to perform, or both.
Here are two examples: A type of node called “New bookmark” will grant the player access to a new website in the Reader tool when activated (all incoming edges have been traversed, i.e. all previous nodes have been completed), while “Datachunk extracted” makes pieces of information appear in documents when reached and then waits for the player to extract (upload) them, upon which the node is completed.
A downside of this approach is that it goes contrary what was originally intended within articy:draft. The internal logic of articy:draft enjoins that multiple outgoing edges from a flow node mean a “branch”, i.e. a decision. This is not the case in our logic: Whenever there are multiple outgoing edges the game will follow all of its paths at once, until there are either no more outgoing edges or the level end has been reached on one of the paths. This makes features like Journeys to try out game flows in articy:draft obsolete for us.
OSMOTIC: A full-fledged localization integration would be extremely useful. Of course there are ways to implement localization within articy:draft already, but they come with some issues of their own, e.g. that additional language texts have to be stored in extra text field variables instead of being a native document text.
articy:draft 3 has been making some progress towards this by offering to store all localized strings in an excel sheet. A completely integrated solution where the game language can be switched and the texts edited in context at any time in articy:draft would still be something I would very much like to see.
OSMOTIC: One expensive task was to keep the production pipeline efficient. We needed to determine a lot of the story and the resulting game flow in order to be able to produce graphical assets, layouts and texts. But on the other hand you need to experiment with the game flow to find out what works and what doesn’t, only that in such a narrative heavy game it doesn’t make much sense without the content. The typical chicken-and-egg problem. You cannot simply “block out” the gameplay ahead of production as it could be done with spatial level design or make a paper prototype to try out the mechanic.
OSMOTIC: I don’t think we ever managed to solve this task properly. The way we went about it was to make main narrative beats for the overall story, distribute them among the chapters, create a rough game flow of a chapter according to the beats that needed to be covered, and then make a huge list of text and graphical assets needed for this chapter.
This meant two things: Since I was the one creating most of the story, the level design and the writing I had to work ahead the rest of the team to avoid downtimes on their ends, which didn’t always work out. It also meant that we needed to create images and layouts without having the final contexts.
With the upcoming second season titled Orwell: Ignorance is Strength we have been trying to improve and speed up the pipeline by moving major parts of the actual writing to Mel and outsource more of the graphics.
Thanks to Daniel Marx for taking the time to answer all of our questions.