To Build A #Dungeon: Interview With Sean Oxspring And Kieran Hicks

Earlier this week, I got the chance to host a Skype interview with Kieran Hicks and Sean Oxspring, developers of #Dungeon. Described by the two as a bullet hell-like blend of The Legend of Zelda and The Binding of Isaac (with a bit of Dark Souls thrown in there for good measure), #Dungeon equips players with the ability to generate their own levels using a very unique tool: Twitter. In other words, it's actual social integration.

I was quite intrigued by #Dungeon even before sitting down to chat. Now that the interview's finished, I actually can't wait for the game to launch. It's clear that the two are incredibly passionate about what they do, and that this is a project they definitely want to see through to the end. They're exactly the kind of developers the industry needs.

What first inspired you to create #Dungeon?

Sean: It was during summer last year, and I was at a barbeque flicking through Twitter. A thought came to me while I did - "what if you made a dungeon crawler using the 140 characters in a tweet?"

We decided third year that we'd do it as a project. Kieran took it on, and we quickly found that we needed it to be slightly more complex than that. Just using the characters in a Twitter message to generate a room is boring; it's a bit dull, so we came up with the idea of using it as a collaborative medium to make dungeons. After that, it sort of spiraled out of control, and we just put forward more and more ideas until we had this whole editing system put in place.

Kieran:  We really wanted to let the users create as much as they could, and we quickly realized that the more stuff we could give them, the better. The Lincoln University course we're developing it for - Computer Games Production - is actually discontinued now, switched over into a games computing course.

Sean: Yeah, it's more like programming instead of the design side. Kieran is one of the final people going through that one.

Kieran:  It's a bit worrying, because they're obviously disregarding a lot of the elements involved in the original course and in game development, but that's just the way the University needs to evolve right now.

So how does #Dungeon actually work? How do the Twitter codes generate the dungeons?

Kieran: The way we did it is that we split the tweet up into three sections. First, we've got the room co-ordinates, which determine where it'll be on the map. These start at 0,0, which is the entrance. Another separator after that determines what type of room it is. For now, we're just working with normal rooms, but we've got fire rooms and the like planned down the like. Next, there's the contents of the dungeon, which have two numbers - their position in the room and the type of asset, sort of like Dwarf Fortress. So a bat is a B, a dwarf would be a D, and so on. Finally, there are two hashtags at the end of the tweet: #HashTagDungeon and #Roomname. So, for example, #HashTagDungeon and #AwesomeDungeon.

Sean: Once a tweet is picked up by the Twitter API, it grabs it and passes it into the game.

Kieran: It doesn't update every second, because obviously that would waste a lot of data...

Sean: And Twitter would probably kill us, because that frequency would kill our servers. The game sort of updates every thirty seconds to every minute depending on how quickly a tweet's been passed by the Twitter API. Once a room's done, though, it's ready and waiting for players. At the moment, we've got it set up so that when people play the game, it'll generate from all the rooms that were in the platform when they started playing. Eventually, we're hoping to allow for real-time updates, meaning people can actually experience new room builds while they play.

We've got to be very efficient, though - Twitter's API is very strict based on how many calls you can do per minute, so we've got to build around that.

Continuing along the same line, how will Twitter be connected to #Dungeon? Will the game be playable without a Twitter account?

Sean: We won't ask people to log in, but they can upload their twitter handle so the game can pull in any dungeons they've made from that handle. What happens is when you actually want to tweet a dungeon from the game, you'll simply hit the tweet button, and it'll have the string for the dungeon there and ready to tweet.

Kieran:  There's going to be an option to not use Twitter at all, and simply play through randomly-generated dungeons.

Sean: We didn't want to limit our players too much. Realistically, the only people who really need to use Twitter are those who want to build their own dungeons. The game's completely playable offline.

How has building a game around Twitter influenced the development process?

Sean:  From the beginning, we set out hoping the game would expand on its own, and people would do things we didn't expect them to do.

Kieran: The second we realized we were using Twitter as the main source of dungeon building, we wanted to put as much social integration into the game as possible. We've got the ratings system, there are leaderboards for rooms, dungeons, and dungeon creators, there's a system that lets you leave messages on the's everywhere in the game.

Sean: It is the game.

Kieran: The community is what's going to make #Dungeon a success, so we want to foster that as well as we can. We've already been floored by the creativity of our players, even with the limited assets we gave them in our initial playtests. People are so much more creative than I am in how they make dungeons, they're coming up with ideas that I'd have never even thought of. That's great to see, and we're just kind of trying to keep it going.

Sean: We've seen some pretty cool stuff. For example, one designer took all the power-ups we've released so far and put them at the end of a long corridor filled with spikes. If you could get through them, you'd get all the power ups, but you'd end up with almost no health - by the time you got back from the room, you'd have about twenty health left. If you were skilled enough, you could still beat the dungeon; if you were a crap player, you'd get killed pretty quickly.

We're actually hoping some of the more active members of our community will set up a wiki. That way, those who don't have the stamina to die over and over again can go and read up on a creature and its abilities and stuff. We want people to be able to recognize the patterns of the enemy and get good enough to take them down

Kieran: We even toyed with the idea of setting it up so that people could actually make their own sprites, but we couldn't figure out how we'd moderate it.

Sean: We're also thinking that if people do things we like, we might name items after them.

There's bound to be at least a few trolls looking to ruin #Dungeon for everyone else.  How will you discourage them?

Sean: We've got different ways of dealing with that.  At the moment, the rooms that are closer to the spawn point are less difficult - we've got a difficulty thermometer thing. Close to the center you can only put, say, a couple rats or a couple bats in; you're not going to be able to throw twenty liches in the starting room. It depends how far away you are and how willing you are to challenge the player.

There's also the ratings system. If a room's too difficult, you'll get a lot of thumbs down and that's going to negatively impact your rating as a dungeon master. If it's interestingly difficult, then you might get a lot more thumbs up, as you've actually taken the time to think about how the room's going to work. For example, you could have a water mage that throws you into a big, spiky pit.

Along with the ratings system, we're incorporating achievements in the form of trophies. Whenever you accomplish something - like slaying a boss or having a dungeon voted to the top of the leaderboards - you'll get a new item for your trophy wall.

Is there some sort of underlying story or narrative to #Dungeon?

Kieran: We've got bosses that are at the end of each dungeon - once you clear a dungeon, a portal to a random boss spawns. We're focusing on environmental storytelling for the narrative - it's going to be there for players who want to seek it out, but we're not going to shove it in their faces. Sort of like in Dark Souls

Sean: I was hoping that we'd tweet out the occasional dungeon with some juicy lore in it sometimes, depending on how things done. Sort of like an evolving lore. So if someone makes a dungeon and they've used the messages tags to do story with - so "this is the story of this man who died and you have to follow his path and defeat him," - we could take the idea on-board and actually design bosses with that man's name.

What sort of plans do you have for your game's future?

Sean: We're hoping to get some of the stuff we've learned from the development process into some academic papers and get them out to different places. Best case, it'll springboard us into making more games as well. Ideally, we'd like to be able to afford rent, eat some food, and be able to make games as well. That'd be the nicest thing to do. If things get even better than that, that'd be cool too.

Kieran: In the short term, it'd be nice to exist.

Sean:  In terms of #Dungeon, we've got plenty of ideas for updates. It's something we're going to be focusing on in the long-term. For example, our first release will only have the basics: the editor, and as many creatures and monsters and things as we can put in. Eventually, we're hoping to do lots of little updates; we've got ideas for an Ancient Egypt-esque jungle expansion with snakes and mummies and the like, and maybe an underwater update.

We're also teasing the idea of eventually having gods, which would be a crazy idea. What we were planning on doing was having Twitter accounts for each god. Each one would find a dungeon somewhere, at random, and say "this dungeon I bless with a shrine to me at this location," and it would create a room with a shrine to them, and you could pray there and get buffs.

Kieran: A Halloween update, as well. We've got that planned out. We're also thinking of adding multiple floors in a later update.

Sean: At the moment, once you defeat enough rooms, you'll get a red demon crystal. Take that to a summoning circle, and you can go through to fight a miniboss. They can be defeated for loot and items. Once you've cleared the dungeon, you get the yellow crystal that brings you to the final boss of the dungeon.

Kieran: We've got the idea of maybe eventually having towers - different levels to a room, tweeting the same room atop of one another. That may eventually happen if we get the chance to include it in a future update.

I think one of the problems we're running into is scope creep. There's so much we want to put in, but we need to limit ourselves - we're a two man team, which sucks.

Sean: Which is why the first version of the game won't have all of these features in it. We're going to work on as many as we can before the early summer release. We'll be paying attention to what the community wants, as well.

Speaking of gods, would players be able to aspire to godhood?

Kieran: We've actually been thinking along those lines. The idea I was toying with is that maybe they could be a god for a single tweet, and we'd revoke it after that. We don't even have any idea what power we'd give to the player yet, but it could be interesting.

Sean: It could break everything if we give them too much power. We're trying to give them as much creativity as we can at the minute without making the game broken as hell

How has working on #Dungeon changed your day to day lives?

The Unitaur is but one element of their changed lives.The Unitaur is but one element of their changed lives.

Sean: For one, it's given us a lot of experience with stuff we're not exactly familiar with. Funnily enough, Kieran's doing a lot more coding than I am in spite of the fact that I'm technically supposed to be a coder - it's his project, so he's got to do most of the programming. As a result, I've been doing all the artwork and stuff.

When we started this project, I hadn't touched pixel art, really. I've come a long way from doing really crappy pixel art to begin with - our first player character looked absolutely awful - and things have gotten a lot better with practice. Hopefully I won't pull a Phil Fish and go over everything over and over again for years and years, but things we did to begin with look like crap compared to what we're doing now.

Kieran:  We've wanted to develop other games as well, but we've had to put a lot of our projects on the backburner since #Dungeon took off so much. All we know is that whatever we do, we definitely want to put a lot of social integration into it. Other players make games interesting. For us, at least.

Sean Oxspring is a recent graduate of Lincoln University's Games Computing Program. He is currently employed at the University teaching first and second years coding and game design, while doing light cloud-based contract work on the side.

Kieran Hicks is currently in his final year at Lincoln. Once he graduates, he plans to put together a Master's Thesis on social interaction in MMOs.