Dwarf Fortress in 2013 | Hacker News Reader

Dwarf Fortress in 2013

gamasutra.com

117 pointsakg77 comments

> As for [porting to the] PS Vita, he notes that "if someone approached us, we're not giving the code to anybody for any reason, so it would have to be something we could compile ourselves."

Just imagine how good this game and community around it would be if it had always been open source. Just imagine how much money they could raise in a Kickstarter to simply open source everything.

Imagine how good it would be if two guys worked on it for 10 years straight, following their own vision and not compromising, making what they thought was cool and making a decent living the while. That's how good it is!

The interesting thing is that, when they speak, these guys sound like planners who would be happy to delegate some of the work involved in realizing their grand vision. That's obviously not how they're operating.

That's almost exactly the opposite of reality. They're not delegators. They don't want to be delegators. They want to make a game.

I think their attitude is slightly unusual, in the sense that you don't see a lot of people who put that much into design and seem to enjoy it but still demand to handle all the programming themselves.

> They don't want to be delegators. They want to make a game.

Delegation is part of engineering (or not, since these guys have chosen not to delegate). Or part of "making a game" if you prefer. It would be genuinely weird to take the attitude that the only people who "make" a game are the programmers.

> in the sense that you don't see a lot of people who put that much into design and seem to enjoy it but still demand to handle all the programming themselves.

Except in indie games all over the place, right? Zeboyd, Spiderweb Software, loads of small developers do exactly this. I do it. I don't want other programmers for something I can do myself because I don't need to share my project with them. It's mine. It's part of me. Bringing more people in lessens its meaning to me.

(I wouldn't contract anything at all if I could do music and art to the level I deem required.)

> Delegation is part of engineering

Whatever gave you the notion that this is engineering? They're not frigging engineers. They're artists.

How exactly is this incompatible with it being open source? Being open source does not mean that you have to accept contributions from others, nor is it incompatible with making a decent living especially when the game is free as in free beer anyways.

True, but they would have to deal with increased competition and market diversity due to anyone being able to fork their vision and putting their own spin on it.

This is (arguably) great for the consumer, but as a fellow creator I understand their reluctance to open source their baby.

I don't think it's about making a "great game", I think it's about making "their game", flaws and all.

Yep, they have said this countless times.

While I am absolutely an advocate of OSS and I do wish Tarn would go for it, the basic thing is that they dont want to and they love what they do, so deal with it, it's their game.

If they open source it they can still control things and follow their vision for their own main branch. But they can accept patches that have improvements for things that (i) are needed and (ii) do not necessarily go against their vision.

There are several great weaknesses of DF that are not the result of the maker's vision but of their limited time and resources.

For example, it is not Toady's vision for the game to get very slow even on powerful hardware when you have a big fortress. This is perhaps a negative blowback of Toady's very laudable vision to have each dwarf be independently simulated. However, an expert in parallel programming should be able to improve this a great deal. AFAIK, DF currently runs as a single thread, and many of its processes can be split off. This is, however, not easy and probably not something Toady is able to or wants to do.

It is also not Toady's vision for the game to be controlled by a mindblowing number of complex menus and obscure key combinations. It was probably not his vision for about half the game controls to be mouse enabled and half of them not. He has admitted himself that the interface needs a lot of work. But he is obviously not working on it, but adding new features like vampires, and Indiana Jones style careening mine carts (last time I checked). I am not saying he should not add those things -- it is a free game and he should work on what is more fun for him. But if he opensourced it perhaps someone else would fix the interface for him.

It is obvious the game needs outside help. People are hacking the game in memory in order to add functionality. Last time I played, you needed to run another program called "dwarf therapist" while playing in order to be able to manage your dwarf's skills and professions. Dwarf therapist would work by directly reading and manipulating dwarf fortress data in memory. Sorry but if the game worked well, nobody would take the time to hack DF just so he can write a helper program.

If they opensourced the game, the people that wrote dwarf therapist could just add that functionality directly into the game.

I say this as someone who loves DF, but has suffered plenty from its frustrations.

People mod every popular closed source game. That isn't a reason to open source it in itself.

Mods, and even total conversions are completely different from taking source code and building a derived game (which then becomes a direct competition to the original).

> If they open source it they can still control things and follow their vision for their own main branch. But they can accept patches that have improvements for things that (i) are needed and (ii) do not necessarily go against their vision.

Open-sourcing it dilutes their brand. It increases the likelihood that they can't fund it anymore because other people start blathering about their forks and distract from the game's actual mindshare.

> However, an expert in parallel programming should be able to improve this a great deal. AFAIK, DF currently runs as a single thread, and many of its processes can be split off

DF's simulation model is deterministic. Most aspects of it depend on prior aspects. It is a great case of Amdahl's law. Ripping that out and changing it would change the core of his game. I can't really see him wanting that if he doesn't understand it and, like Baughn's work on the graphics subsystem, I can't see him spending a lot of time to understand it. (Toady isn't a computer scientist and he isn't really a hacker as is commonly understood here. He learns what he needs to make his game and it pretty much ends there. This may sound critical but I mean it as a compliment.)

> But if he opensourced it perhaps someone else would fix the interface for him.

His supporters are content with the interface. It isn't preventing them from enjoying the game and it's not hindering their ability to fund continued development. You are suggesting that he potentially damage his revenue stream to fix problems that aren't really problems right now.

> His supporters are content with the interface.

I haven't sent him any money yet, so I'm not sure if I count as a supporter - but I am damned not content with the interface. If it weren't for Dwarf Therapist, the game would be nigh unplayable.

I've never used Dwarf Therapist (and I realize I'm in a minority there), but I also have given him money because the interface does not bother me. Enough people are doing likewise that I don't think the interface should be a priority until the game's there.

> DF's simulation model is deterministic. Most aspects of it depend on prior aspects.

I don't see what that has to do with using parallel programming to improve the performance of DF's simulation. If you are implying that threading implies nondeterministic behaviour in a simulation, you are mistaken. In fact, a system with such a discrete space model as DF would be quite easy to parallelize in a safe and deterministic way, with the proper expertise and forethought.

Sorry, you're right. Nomdetermimistic isn't the right word. "Heavily shared" might be better: it isn't perfect cellular automata, a lot of the game word cheats in a way that would necessitate really weird locking or sweeping changes of the game world. Couple that disruption with Toady not really understanding that sort of architecture and I can't see him ever being okay with other people doing it. He'll learn to do it when he absolutely has to.

The lack of perfect cellular automata isn't an obstacle either. One of the simplest and most robust way of doing parallellization of a system like DF avoids all such issues in a very simple way with no almost no synchronization needed.

Basically, when starting the game, spawn X threads and partition the world between them so that only one thread can write to any particular world block. Then, when calculating a new frame, make a read-only copy of the previous world frame. Any reads are done from that copy; any writes obviously occur only in the blocks assigned to a particular thread. Wait for all threads to complete, then render thread. No messaging, locking, event handling or whatever needed. Just care and forethought.

But you are right that Toady likely not possessing that kind of knowledge. That's the very reason why people would like to see others involved for things like optimization and UI design and the main point these people are trying to make; Toady seems either not able or willing to do it himself, and it could improve the state of the game tremendously.

They only need to provide a socket and a protocol. Then they don't have to waste time on the interface and it already gives plenty of community modding opportunity.

And in open-sourcing it, they dilute their vision for their game. No thanks. (And I am a donor. I don't care about being given their code.)

Open source games are rarely very good. There are exceptions (looking at you NetHack), but this is the one area where "give away software, sell support" does not seem to work. Then again, I wish more games were open sourced a la Quake after their sales start declining.

> I wish more games were open sourced a la Quake after their sales start declining.

That is where open source in games seems to excel: keeping old games alive and running on modern systems.

Closed source games are rarely very good.

Closed source games are the only ones that matter. Name one AAA production that's open source.

chess

Dungeon Crawl: Stone Soup is an open source game (and not open source in the sense of "you can look at our code but don't expect us to value your input", open source in the sense of "make pull requests!") and it's probably the best game I've ever played. I don't think I disagree with you, though: the reason it works so well for Stone Soup is because the game lends itself to personal extension (it's a roguelike, so you're dealing almost exclusively with game logic).

Thanks for the praise. crawl is the only major project I'm involved with, but I'm really pleased with how well development works. The team is totally open to external contributions. For example, some friends and I added a god back in 2009. Late last year we started getting high quality patches from a contributor who then became the most prolific member of the team.
One thing that holds crawl together is an overarching philosophy, particularly the Berlin Interpretation (http://roguebasin.roguelikedevelopment.org/index.php?title=B...).

Someone even made fan art for us: http://i.imgur.com/cGO39.png

You added Chei? I lost many a Naga because of you, until I got the hang of Temporal Distortion :)

I added most of the really bad features.

I so badly want to learn how to really play roguelikes like DCSS. I've never gotten past the wander-around-first-few-levels of NetHack and DCSS. There's so much there!

You should start with Brogue (https://sites.google.com/site/broguegame/). It's a roguelike that manages to be accessible without compromising the core elements of the genre. The interface is great and it actually looks quite beautiful with its lighting system. The design has been streamlined, simplified but not dumbed down.

Try FTL. It's the most modern reincarnation of the rogue-like genre and it's close to perfect in everything it does in that area.

I would suggest Moria. I think I was 10 or 11 when I played this. Nethack and DF are very hard, relatively.

http://roguebasin.roguelikedevelopment.org/index.php?title=M...

Dwarf Fortress is completely free (gratis) to play. The developers make money via donations, which I expect would only increase if the game were open source.

Quake is not open source. The engine is. To actually play "quake", you need to buy the assets.

Considering the quality of open replacement assets like OpenArena, I also don't think that open source games work particularly well.

Minecraft would be an obvious counterexample I think. It and its community are doing just fine.

I've being mostly uninterested in playing dwarf fortress until fortress mode sees improvement. However I've being watching its progress with gusto. Toady is focused on the simulation aspect of the game. The generated story. Right now, that has no value, presumably it will later.

If DF were to be open sourced, the first thing I would do is attempt to contribute to cleaning up fortress mode. I suspect many developers would do this as well. So as a long long long time observer of this game, I can understand the lament that it's not open source and I sympathize. These sore spots likely won't see improvement. But...

All the reasons I normally advocate for freely open source software just arn't here!

Toady is clearly not doing anything malicious with his program. He also has no intention of restricting his program in the future, anyone can use it. He's trying to make the best software he can, for the user. Dwarf fortress isn't doing dirty things behind the scenes, and Toady has more than earned my trust. Versions have being ported to multiple OS's, and it's clear he's going to work on it for a long time. Open source isn't needed here, it would only get in the way of what he's trying to achieve. I'm willing to wait until I'm a much older person to see the masterpiece he creates.

Once upon a time, Tarn Adams did an AMA on Reddit. I asked him [0]:

  Have you considered open sourcing more pieces of Dwarf Fortress so that the
  community can try to help with the nitty gritty engine bits while you can
  continue having sole ownership of the underlying direction?

  The pathfinding and multithreading optimization is a good example of
  something I'd love to help with given the opportunity. (Related to this
  question.)

  Is there any other way that the more technically-inclined can help?

Tarn's response:

  I'm just not a project manager. I tried sharing the pathfinding code with
  some people, and some progress was made, but unless my involvement is more
  hands-on, I don't think it'll work, and I didn't want to be involved in that
  kind of arrangement. On the other hand, the porting and graphics work with
  Baughn went pretty well, although now I can't work with the graphics code
  myself because I don't understand it. I'd hate to have that happen to more core
  pieces of the game. Baughn also had to suffer a lot of annoyances because I
  have no inclination toward things like version control.

"I'm just not a project manager." As someone who has several active open source projects, I can very much relate. I can easily see Tarn and his brother be completely overwhelmed by open source activity, paralyzed to make any kind of internally directed progress on it. Also, not everyone is good at maintaining this kind of thing, so it could cause more harm than benefit to the project.

I too wish the code was open source, but I can understand the reluctance and speculate that it may even be the correct decision until the project is completed or abandoned.

[0] http://www.reddit.com/r/IAmA/comments/1avszc/im_tarn_adams_o...

The thing is, they literally would not have to change a single thing after releasing the source code. They could refrain from looking at a single pull request. But perhaps their fear is that community additions may lead to a variously forked ecosystem which could dilute the "Dwarf Fortress" name thus dilute their grand vision.

Indeed. They have a captive audience right now which donates on average more than $4,000/mo for the work of Tarn and Zach.

Take the +12 month release cycles, there will surely be more active open source forks than what the Adams' can compete with, resulting not only in loss of their revenue stream but also loss of control of their vision. Perhaps this would be an altruistic benefit to the community, but probably a solid loss for the creators unless they're willing and succeed in diving in and leading the community of contributors as a whole.

Very risky.

Also from another interview, Tarn spoke about how he sees Dwarf Fortress as his life's work. This is the masterpiece that he would spend the rest of his life perfecting. Imagine a master painter like Picasso working on a piece for nearly a decade, being asked to let the community contribute to it so that it will be completed sooner.

Exactly, we should understand the fact that open sourcing is not always compatiable with a work of art. (Games are art now, right?)

If someone forked Dwarf Fortress and changed the key bindings to a sane default, I would switch to Midget Mountain or Little People Palace or whatever in a heart beat.

Someone tried making their own clone, with a better interface and gameplay in the direction they wanted it to go. It's hard to replicate the amount of effort that has gone into DF.

http://www.goblincamp.com/

Gnomoria seems to be coming along:

http://www.youtube.com/watch?v=E75mL6R42i0

Lots of updates, but it's not open source as far as I can tell.

There are a lot of "clones", but these are essentially DF lite. They mimic the basic gameplay of digging under ground and building some sort of fort. Read through the past 3 months of updates ( http://www.bay12games.com/dwarves/ ) and you will see how ludicrously complex & detailed the game has become. You can't explain some of these things in a feature checklist, but rather with paragraphs.

Sure, but look at what happened to Arc. There was excitement in the beginning, but since pg wasn't really taking patches (except bug fixes), the community forked the project and started working cool stuff on their own. Eventually, pg released new versions, which never incorporated the community fork's changes, so bit rot crept in, and people gradually lost interest and drifted away from the project. The project isn't dead exactly (is Lisp ever really dead?), but it isn't the hub of activity it was at the beginning.

I don't think dilution is main threat here. I think turning Dwarf Fortress into a viable open source community takes effort, and just letting the code into the wild won't necessarily accomplish much.

I have no inclination toward things like version control

This is the key right here! Without version control, there's absolutely no way he's going to be able to collaborate with a whole bunch of people on this thing. If he decides to dump the source code up on his site, it's basically read-only at that point.

Besides that, he derives his income from donations. He does not want his project getting forked and his community leaving him behind in the dust.

>Just imagine how good this game and community around it would be if it had always been open source. Just imagine how much money they could raise in a Kickstarter to simply open source everything.

I don't think he'd even need to even open source it, in terms of allowing anyone else to release it. If it was just possible to hook into it and make custom UI or Renderers, I think people would run with it. Some have tried but they are extreme hacks at this point.

I will continue to have no interest in Dwarf Fortress until it has an actual UI instead of whatever insane random combinations of controls the developer felt like adding on any particular day.

That's one of the reasons why Minecraft is so much bigger than DF.

These guys just don't get it. DF can be SOOOOO much better, yet they are working on adding features instead making the game better.

Here's the difference: DF is Toady's baby, his life work. Toady has the idea for it, and he does not want anyone else working on the code. Minecraft started out very rough around the edges, and only started polishing it once it had already hit it big and he could hire other people.

With a one-developer game, you don't spend time polishing things. You bounce around different areas, adding things as you wish. Letting yourself get bogged down in polishing details that aren't exciting makes you start looking for other things to do, losing interest in the project altogether.

Minecraft was developed by one person - Markus Persson (aka Notch), at least till 2011 when it already got insanely popular.

A perfect analogy! Toady's baby is legally blind, is very dumb, is very slow, speaks 4 languages in different contexts but none fluently, is very ugly, and is not fun for 99% of people who play with it.

There are cures for all of these problems, but he refuses to treat or allow anyone else to treat his baby.

It's fun enough for enough people, myself included, for him to live off the project. And that's enough for him.

For now. What happens when interest slowly peters off as the current fanbase gets bored or distracted with other areas of their lives?

...then he does some user-facing stuff to lure them back and add new ones?

This project isn't some attach-rate-obsessed no-value-add startup. As long as he can pay his rent, he doesn't need to make anyone happy but himself.

Ah yes. It's obviously they who don't get it. All they've done is create and turn into a career their pet project. Obviously they should just listen to all those that want to radically change what the game is rather than iterate on the design they (and many others) enjoy and see where it leads.

There is a reason DF has inspired so many others (including the creator of minecraft), and it was't the graphics.

Which do you think is more likely, that they don't get it, or you don't get them?

Come back in 11 years and see if Minecraft is still so much bigger than DF. There's a difference between UIs crafted for hooking new users and those designed for long term use by experts.

Basically: you're starting the equivalent of a "vim vs. IDE" flame war, and adding just about as much value.

DF is inefficient and inconsistent even with the hotkeys memorized. It's no vim.

I'm sure minecraft will be smaller by then, but that's because it's far less complex and development is basically nil at this point. Nothing to do with how the controls are implemented.

Actually, modded Minecraft sees an enormous amount of activity, particularly of late it seems. It's started cannibalizing my DF time, of all things--vanilla Minecraft is dull by comparison.

DF sees a little of this with dfhack and all the mods on the forums.

Yes, there is a bounty of mod activity, but it is rare for mods to outlast the source game by a decade.

I'm not sure Minecraft's all that relevant a comparison - it's little more than a glorified Lego set. It's easy to have an accessible UI when you don't have to interface with much more than "Point at something", "Hit something" and "Place something". You might as well compare it with FarmVille - after all, it's a hell of a lot "bigger".

> These guys just don't get it. DF can be SOOOOO much better,

Maybe. Or maybe it would lose its soul and become yet another pretty game with boring back stories.

I'm torn about that, but I can't help notice that the reason why the stories of your dwarves are so engrossing and so unbelievably epic and real is because 100% of Today's time is spent working on these stories instead of a pretty GUI.

The problem with DF GUI is not that it is not pretty, but that it is not effective at all. Additional colors, alpha channels (like looking at 2-3 depth cross-sections instead of one), no fixed font size, so smaller info screens, proper zooming (boy, how useful that would be), smoother scrolling to not get lost, etc.

I understand they do not want to invest time in these, but all they need is to open up a TCP port and have a protocol that accept commands and returns some game state. Then we could make all kind of open source clients, some fancy (2D, 3D, 4D, whatever) some crazy (multi-screen setups), web frontends or android thin-clients to cloud hosted DF, etc.

I have not played in quite some time, but if you are looking at playing: after you get the rough hang of things, make sure you look for mods. Specifically bug-fixes. There's a (small) community around the `dfhack` tool that fix a lot of game-breaking bugs. Masterwork [1] is a comprehensive mod that lets you turn off some of the "flavor" in exchange for a more performant game (about 25% or so). For example, the default game has dozens of types of leather. Do you really care if your dwarf's mitten is made of cow leather or bull leather?

[1] http://www.bay12forums.com/smf/index.php?topic=125633.0

I've said it before and I'll say it again. One of the central philosophical questions of our day will become whether or not it is ethical to even turn off dwarf fortress.

That was one of the taglines for SimCity 2000: "If this game were any more realistic, it'd be illegal to turn it off." (http://www.gamefaqs.com/pc/198648-simcity-2000/images/box-11...)

So that's why they dumbed down Simcity 5!

Sorry, couldn't help myself.

I am dangerously close to writing an MA thesis on just that question...

Prior discussion: https://news.ycombinator.com/item?id=5982459

Okay, new idea. What if he did a Kickstarter for $500k for additional high-quality devs to come on the project for a year and focus on 1) improving the UI and 2) improving performance. That's it. No promises of additional functionality, no open-sourcing to make Toady feel like he's given up his baby.

Given the cult following of the game and the people who would like to play it but are otherwise frustrated, I doubt they'd have trouble reaching the limit. The press they'd get just from announcing would give them a boost as well.

He doesn't want more developers. He is happy with the project as it is. It is his life's work, so to speak, and he is content with progress happening over the span of years.

I don't understand why this is such a hard concept for some people.

Every time I see threads like this, the comments are all about telling the developers what they should do. The developers make enough money to survive, seem perfectly happy, and are working on exactly what they want. They aren't in a situation where they owe the community something more than what they are already doing. Why are we trying to solve their problems?

They should not focus on interface at all, but provide a socket and a protocol to communicate with the engine. Then the community can build fancy clients, mods, whatever.

Reply on news.ycombinator.com