interviews with makers

interviews with people who build great things. new posts every monday.

Michael Townsend, A Dark Room

Michael Townsend is the man behind Doublespeak Games, which recently put out A Dark Room, an ultra-minimalist browser-based game. (It’s a lot of fun and you should probably play it.)

There’s surprisingly little background information about you available, besides the fact that you’re from Ottawa. Can you talk a bit about yourself — where you grew up, education, interests?

I grew up in Ottawa, more or less, and spent a lot of my formative years hunched in front of the family 386, coding little distractions in QBasic. I distinctly remember copying the bundled version of Snake, line by line, to see what made it work — I like to think that makes me like Hunter Thompson.

I spent four years at the University of Toronto getting my degree in Computer Science, then spent a year traveling. Now I sling code for a living, and write games in my spare time. When I’m not playing games, of course. Or drinking coffee. Those other ones are probably more common, actually.


Moving to A Dark Room: it’s definitely a unique game, both in terms of mechanics and aesthetic. Can you talk about where you drew your inspiration from, both in terms of the decision to make it and in what directions to go?

The immediate inspiration came from aniwey’s Candy Box. About ten minutes into it, I knew that there was huge potential for the genre. Exploration as an aesthetic tends to be used spatially, with the player wandering around vast worlds and uncovering interesting things along the way. This is a great mechanic, and I’ve sunk more hours than I care to admit into Bethesda’s epics.

Less common these days is exploration of the mechanics of the game itself. This was a huge part of PC gaming in the early days, though not necessarily intentionally, and I miss it. Contextual controls and ever-present on-screen hints are excellent for removing the barrier to interaction and allowing the player to easily express themselves in-game, but what’s lost is a certain sense of the game as a mystery to be solved.

Once I’d decided to riff on Candy Box, I threw together a quick mockup of the interface. The vast expanse of white on screen screamed “post-apocalypse” to me, and so I ran with that. The rest just sort of came together as I built. I’m terrible at pre-production. I shamelessly stole the resource management stuff from games like The Settlers and Anno. I adore juggling complex economies, which probably makes me beyond saving.


You talk about the older PC games that also encouraged exploration of the mechanics — in my mind I’m thinking of stuff like the Monkey Island and Ultima where there’s a particular emphasis on experimentation. Are there any specific older series or games that you’re fond of?

I’d go back even farther to parser-based adventure games like the early Sierra titles. I’m particularly fond of Space Quest, myself. In those games, you weren’t even sure of which verbs the game would understand, or which objects could actually be interacted with. This is totally frustrating game design that would never happen these days, but it led to all sorts of fun side-effects. Developers would hide easter eggs and jokes behind many of the non-functional commands, encouraging experimentation and creativity from the player. Any game with a complex enough interface will give you this kind of experience, though, if you refuse to read the manual…


You obviously opted for the ultra-sparse narrative style and stuck to it — did you find it was easier to write this way, or more difficult?

Oh, definitely easier. Writing content is approximately 100 times less fun that writing code, so the sparse narrative helped me maintain interest. The game was actually in danger of not getting finished once all the mechanics were in place. All that was left was writing the encounters and dungeon paths, and I desperately did not want to do it. Once I fell into the writing style, it flowed pretty easily.


You open sourced ADR last week, to the delight of many (myself included)! I know you’re working on internationalization, but what other plans do you have for it? Was it an easy decision to open-source it?

Personally, I’m basically finished with ADR. My involvement with the project now is limited to merging pull requests and chatting with other developers. I know at least one of them is working on expanding the scope a little, and I’m totally looking forward to playing it. The decision to open-source actually came long before the code went up on GitHub. Very early in development, I decided against minifying the Javascript. The page takes a little longer to load, but if even one person pulls up the source and learns something, it’s worth it. I naively assumed that would be enough, but I got a few calls to put the source out under an open-source license, and so I did.


Speaking of plans — what are your plans for Doublespeak Games? Currently, the only content from the site is pertaining to ADR; do you plan on working on other titles?

I’ve been writing games under the moniker of doublespeak for over three years now, and ADR is the first thing I’ve actually finished. That should give you a little insight into my work habits. There are always game ideas rolling around in my head, and I start building a few of them. My problem is ambition. It’s really hard for me to keep the scope reasonable for a one-man development shop. I’ve got a couple ideas right now that I think are promising, and I’m going to pick one and start hammering out some design soon. I’ve promised myself a shiny new ultrabook if I can actually finish a professional grade design document before diving in to the code. Whatever it ends up being, though, it’ll be quite different from ADR.


Can you expand upon [the issues of ambition and scope] a little bit, if you don’t mind?

My imagination doesn’t care about timelines or my own personal sanity. Often, I’ll come up with an idea that I think is awesome and start hacking at it right away. Two months later, with only a tiny fraction of the game finished, I’ll lose interest and the project will die. This has happened more times than I care to admit. With A Dark Room, I specifically tried to keep the scope small enough for me to finish before getting distracted. Turns out that takes about a month. My original plan actually called for another section of the game after leaving the planet, but I felt myself starting to drift away and so decided to cut it short rather than leave it unfinished.


Are you surprised by the response to ADR? Were you expecting it to be as popular as it turned out?

I knew when I started that idle games had pretty large viral potential, so when the hits started coming in I wasn’t terribly shocked. Then I got linked on JayIsGames and my traffic went up by an order of magnitude. That felt pretty great but, as I watched the hits start slowly trailing off, I was convinced that that was the end of it. That cycle has repeated itself three times now, and I’m shocked every time. It honestly never gets old. About 160,000 people from all over the world have played a game I wrote here in Ottawa, and that’s pretty amazing. I love the internet.


That’s definitely an awesome number, and I can only imagine the accompanying feeling. Have you read any of the feedback about the game? Is there anything, either from a structural or actual mechanics perspective that you wish you would have done?

Oh yeah, I read feedback all the time. In the past few weeks I’ve learned that I’m actually incredibly vain. The majority of the feedback is really good — people really like the expanding scope and the sparse narrative. I love reading through forum threads where communities are playing through the game together and figuring things out collectively. Of all the complaints (except for the people who dismiss the game out of hand as boring), most people hated the ending. Some people wanted more exposition, which I will never give them, but some just felt like the loop invalidated all of their work.

One of my GitHub contributors pointed out that the spirit of ADR actually calls for space at the conclusion for reflection, which is denied by the loop. I totally agree, and the next version will end differently. There’s definitely things I wished I would have done. I don’t think I’m alone in saying that a work is always released before it’s “done.”

The problem is that “done” is unattainable. It’s constantly on the horizon, just a few more features away. At some point you just have to cut it loose, or you’ll go crazy. I definitely wanted more content in general: more random events, more paths through the map locations, more variety in map generation…


The idea of releasing something before it’s “done” definitely makes sense — there’s the idea of the Minimum Viable Product, and I think the idea of a Minimum Viable Game is even cooler. How did you know when to say to yourself, “okay, this is ready for the world”?

When I got sick of working on it, to be honest.

When production switches from being fun to being work, I know I’m in danger of abandoning the whole thing. I hit that point with A Dark Room and knew that I had to wrap it up quickly, or not at all. That point came while I was writing all the paths through the towns and cities, but before I’d built the space segment. I bribed myself with some fancy visual effects at the end to get all the writing finished, and then closed the book.

I played it for a few more days to iron out most of the bugs (but not all… There were quite a few fixes made in the first week after release), and then posted it. I was actually really worried that there wasn’t enough content to keep it from getting too repetitive, since I’d written only about half of what I’d planned.

  • 15 July 2013
  • 4