Dear Esther or interactive poetry

Imagine the following scenery: Harsh rocks, that barely distinguish themselves from the dim grey sky, flank a withered and broken down building towering into the sky. A lighthouse of days long past. The wind is shrieking and waves are crashing on the barren shore you find yourself on. Somewhere in the distance a shimmering red light can be seen, pulsing through the mist that hides the rest of the coastal landscape from your gaze. The soft melody of a piano-tune echoes through the noise. It carries a somber voice with it, calmly reciting something that could be a letter: “Dear Esther, …”.

These are some of the first couple of impressions I got once starting the indie-computer game Dear Ester. The recently released game started out as a mod for the popular game Half Life 2 and I stumbled upon it about a couple of years ago in this article. The description the author gave back then somehow made me curious, but I didn’t own the original game needed to play the mod and at that moment didn’t intent to organize it just to give Esther a try.

However, after doing some research I discovered that Dear Esther became so successful that its creators were planning to relaunch it as an improved standalone application. So I made a note to check every now and then and discovered yesterday that its standalone version was released this month. Since I just had some time at my hand, and interactive story telling is somewhat of an interest of mine, I bought it right away to sate my curiosity.

Now what is Dear Esther all about? In fact, this not quite easy to explain — and intentionally so. It’s makers describe their creation as:

a ghost story, told using first-person gaming technologies. Rather than traditional game-play the focus here is on exploration, uncovering the mystery of the island, of who you are and why you are here. Fragments of story are randomly uncovered when exploring the various locations of the island, making every each journey a unique experience.

You find yourself in the role of an unnamed character in the middle of a small, barren island that is located somewhere in the Hebrides. Initially, you do not possess any information as for what brings you to this forlorn place, or what you are supposed to do here. Besides the island itself and the voice-overs in the form of monologues by the narrator, you are not provided with any information at all. No clearly defined goal, no dialogs or any other form of interaction with characters. No progress bar or any HUD giving you hints about your character or your mission. Nothing. Only the barren landscape and a voice.

But it is exactly this sense of being lost, of having lost something, that is one of the elements making this such an radically different experience — even more so, if you have played computer games before. The mechanisms that make Dear Esther work, are  precisely opposing the ones that are usually connected with the habitual gaming experience.

Traditionally some indicators for an highly immersive gaming experience might be a great amount of interactive possibilities with the constructed environment, and the possibility for freedom of exploration. In here both are radically stripped away to a bare minimum, so that one might see it as rivaling the barrenness of the rocky island itself: Our unnamed explorer isn’t capable of jumping, nor of running. He is not allowed to pull levers, push buttons or collect items to solve any riddles — nor does he have to.  There is nothing to win and nothing to lose. The distant goal of the aerial in the sky is always offers one a clear path to follow along the island. One can never get lost. Simply put: Dear Esther isn’t about choices, instead its secret ingredient is the unique atmosphere it conveys to tap into the players own imagination along the predestined path.

The fragments of plot provided by the narrator at certain time only offer glimpses at a coherent picture — like shattered pieces of a broken mirror. They talk about different characters, and also shed more light on Esther.

By taking away all but the most essential possibilities for action — locomotion and the shifting of the gaze — one is allowed to focus on the smallest of details provided by the environment and trying to close the gaps purposefully left by the narration.

For example: A pair of rusty ships wither away on the rocky shore line bearing cryptic writings on their hull. While one might see them as mainly opportunities for action in an ordinary computer game, e.g. spots to take cover from enemy fire, they become much more in Dear Esther: Like the carcasses of long gone giants they seem to bear witness to the secret of the island — to the purpose of one’s journey — and instinctively one struggles to find their testimony. How did these ships end up here to begin with? Are their crews still somewhere on the soil of this island? Did these same people put the cryptic writings on them or was it the protagonist himself? Perhaps their steel bodies represent merely manifestations of another aspect of his connection to Esther?

Interestingly enough, this distinction between the surreal and the real is only once made tangible: When the character’s path leads him underground into the subterranean caves beneath the island. Only here the grey and barren coastline gives rise to the bright and shimmering crystal caves with an endless array of colors (a twist that somehow instinctively reminds me of Tim Burton’s Corpse Bride with it’s dull world of the living and very lively afterlife). In combination with the eerie sound scape this gives these tunnels almost something sacred and awe inspiring.

The climax of this journey through the underworld certainly is a scene in which one falls into one of the blue lightened ponds of water just to drift away into an underwater dream scenery of a car crash site on an empty road. Again: There are no explicit evidences as for what happened. Just a sense of loss and sadness. And this is what constitutes for me the core of the whole experience: There were a lot of questions I asked myself on my own journey through the island. Whether or not they are actually solvable, I don’t know. I am certain however that it doesn’t matter, but it is the process of asking — the imagination and feelings involved in this process — which truly do.

Because of this, I wouldn’t agree with the people that praise Dear Esther as a highly innovative game. This is because in my opinion it is simply not a game at all. Just as much as a poem by is not a thriller: Their individual qualities are not comparable. They might use the same medium, but that is about the only thing they might have in common. To me, Dear Esther constitutes the first representative of interactive poetry I encountered. Just like poetry you need to be in a certain state of mind to be capable of enjoying it. It probably isn’t what I would turn to for letting of some steam, or something to play on a portable device inside an overcrowded train. If however one approaches it with an ability for imagination and the patience to let it wander freely on the empty beach of a forlorn island, the reward will be a unique experience and some moments of extraordinary beauty.


The process of designing interactions between humans and machines has been dominated for decades by the creation of complex patterns of user actions with simple input devices. This started to change in the recent time: Systems that allow the creation of interactive environments making use of the human body as a whole are spreading. Also, they are no longer detached systems of information processing, but aware of the wider context in which they are situated and include it by design. One of the major challenges in current interaction design is therefor to transform them from mere objects of curiosity into designs usable in everyday life situations.

The project IN-CYCLE is a contribution to that attempt by building and evaluating an prototypical setup for the documentation of human interactions involving the whole body.

It is based on multi-perspective slit-scan recordings rendered to three dimensional objects and combined with various sensor data.

It enhances the process of analyzing users’ interactions with gestural-based interfaces by providing a unique perspective on them, highlighting properties not visible by traditional means of documentation.


In his paper dealing with the Video Streamer (1999), Elliot highlights a couple of interesting criteria that have to be addressed when “rendering time”.

  • How to portray motion in a still image.
  • How to display video frames so they portray their own temporal characteristics.
  • How to associate a view of motion within the frame with a view of motion beyond the frame.
  • Different ways to effectively render different time scales and how to relate them.

They are hinting at a couple of problems that have to be addressed by IN-CYCLE as well.

The core problem is obviously already stated in the first one: How to transcend the limitations of the still image while at the same time being confined to it? There are various possible solution coming to mind. Some are elaborated in more detail below and are mainly an evaluation of previous attempts using similar technical ways of transforming video footage. Another attempt would be to look at how motion has been tried to depict in illustrations and other creative processes dealing with still images. Even if it is clear, that animation is rather interested with recreation of the illusion of motion through still images (so in a way quite the opposite of what I am trying to do here), there may be some interesting insight to be gained from it as well.

In order to truly highlight the development of an interaction process, motion needs to remain motion. I think no simple reduction of the interaction process to mere atomic events, will do justice to it’s emergent properties. A solution for this could somehow lie in Elliots fourth criteria. If the connection between the current limited amount of visualized data and the overall process could be maintained in an appropriate way

Other projects tried to tackle this issues in various ways before, so it seems to be a plausible first step to build some prototypes similar to these existing works and evaluate how well they fare with addressing these critical aspects of “rendering time”.

Related Work

salient Still from Teodosio&Bender(2004)
SALIENT STILLS (Laura Teodosio&Walter Bender)


videoStreamer screenshot
VIDEO STREAMER (Edward Elliot)


time crystal sculpture
TIME CRYSTALS (Tomas Walizky)

Experiencing interactivity – part1

After resigning from my job at the GUC due to my plans of pursuing a master’s degree, I finally have the chance to continue working on my own projects. I kept telling myself that I am doing just that for the last two years – and indeed everything I did was connected to my research interests in one way or another – but somehow I there was just never enough room to get beyond the I-will-start-now-or-latest-tomorrow-but-in-the-worst-case-the-day-after-tomorrow phase. Well, I’ve got the opportunity to change that now.

I have been thinking about for a while to follow up on my bachelor thesis about the phenomenology of interactivity. Hugely inspired by Dag Svanaes groundbreaking PhD work on interactivity, I asked myself a simple question in it (although, I still have problems formulating it in an equally simple way): What are the effects of interactivity on our experience while engaging with interactive systems?

I think it is easiest to explain it by using a small example.

What do you see here? Most probably you recognize what is depicted in the image above as symbols/icons in a desktop environment. However, you would not try to use them in the same manner as you would use them if this was a desktop, despite them looking the exact same way. In fact, I bet you are a bit confused right now because of the last sentence. The very thought of this comparison seems absurd to you: Of course you wouldn’t try to use them – it’s a screenshot!

However, if you stop for a second and give it a thought, you will understand what I’m aiming at. They look exactly the same as they do in any standard Windows7 desktop environment all over the planet. Yet the context of them being displayed in your browser makes you not perceive them as objects to interact with. If I would not have started to elaborate on them, most probably you would not even have regarded them as separate objects on top of a background, but just as another screenshot in a blog post. This is because the phenomenon in question is so much part of our existence that we usually don’t think about it.

What makes the graphics displayed in the context of a browser different from them being displayed in the context of a desktop environment is, that in the former you do know that your actions on them do not have any reaction attached to them, while in the latter case you learned that they do (provided you are familiar with the concept of operation systems making use of any kind of desktop metaphor at all). The relation, between the icon being displayed on the desktop and it being depicted in the screenshot is of the same kind as the one between an object and an image taken of it — Only one of them is the real thing. The possibility for action is what distinguishes it from a mere graphical element and makes it what it is.

Now this insight about an desktop icon alone may not seem to be terribly exciting, but it is an small example that in it’s triviality highlights an important point — one which has long been neglected in designing interfaces. The point is, that Interactivity is not only a result of changing-system-feedback of some kind, but an first-class entity of our experiences in dealing with artifacts/designs/applications/systems/whatever-might-be-an-appropiate-term-for-the-things-we-both-design-and-experience-using-computer-based technology. Usually we do not focus on this aspect of our experiences of interfaces, because they are so tightly connected to the visual sense and quasi-standardized structures of systems.

This goes beyond mere conventions in the sense of Norman, however. It is not an functional indicator that we infer from it by logical deduction or by knowing it already. Possibilities for actions are not some kind of third things we add to their appearance. What things ARE, is tightly connected to what we can do with things (or cannot do with them for that matter). Thus, it needs the context, the appearance and the possibility for actions to emerge. Previous experiences shape, saturate and even define future ones — A very important insight for interaction design.