Hacker News Reader

Comment by meddlepal | Hacker News Reader

I really wish DF was OSS. My great fear for this game is the developer dies and we lose an amazing piece of tech with him.

View on HN

Being a personal project, there's also a concern of losing control over one's art. Would Rembrandt invite a hundred paintbrushes upon his canvas?

Quoting a comment from a thread several months ago (https://news.ycombinator.com/item?id=11999393):

> Actually, he's done that in the past. He let some folks from the DF forums have part of the source code, specifically the part which uses OpenGL to draw the graphics. They were able to improve it so that it was much more efficient, and so that it could use TrueType fonts. The results were quite successful, in that the graphics now take up much less of the game's time, and TrueType fonts can make the game much more approachable.

> On the other hand, Tarn has said that he now no longer understands this part of the code. If he ever needs to modify it he'll either have to learn how it works from scratch, or farm the work out to the community again. Since he has no desire to become a project manager, he's unlikely to do the latter.

If Rembrandt could duplicate his canvases for free why not?

Why can't official Dwarf fortress exist alongside clones? Didn't seem to hurt the Quake engine and Carmack's engines were every bit the craftmanship as DF. Today you can run Quake on anything, it's been modded for VR support, new effects, etc. DF? We'll be lucky if it doesn't die and people hack it anyway.

You gotta respect people's desire to control their baby, doesn't mean it's productive, though. He'd literally lose nothing other than some notion of control by releasing it out in the open. He might even be able to find a novel open source business model if he tried (though that's not his desire or goal).

The last part about understanding does make sense; though if development was out in the open he'd have no problems finding people to explain it.

I know at least in some cases the reason for people not wanting to make something open source is that they don't want people to complain about their style of coding or to say that they are "obviously" bad programmers because they don't use design patterns or whatever the critic's favorite philosophy is.

I imagine a DF source release would spark a handful of Medium articles lambasting Tarn's coding practices.

And at least a few praising them as a refreshing new alternative to blah blah blah, or something.

The engine is the foundation upon which the content runs...but engines in and of themselves aren't content. In the case of most game engines, the two are well-separated.

In the case of Dwarf Fortress, I imagine the content and the engine are heavily coupled - if not one and the same. Separating the two will be an enormous task for Tarn, who likely has other priorities in DF development and is uninterested in the inevitable bikeshedding that will result should he open the whole project open up to others.

He's a mathematician, not a software engineer or project manager.

> You gotta respect people's desire to control their baby, doesn't mean it's unproductive, though.

While I agree, you'll probably find that his notions of "unproductive" will be wildly different than yours.

Depends, usually the content and engine become quite intertwined about 3 months before release as you try to squeeze every MB/cycle out of hardware.

Even then game design will dictate architectural choices(off the top of my head streaming, networking and platform all have huge engine implications and can't be easily retrofitted).

Quake was only open-sourced years after it was superseded by more advanced games. It was still an awesome thing, and I'm glad that ID seems to have made it a tradition.

What I like about Open Source software is that I can audit it to make sure it's doing what it says it's doing, and that I can modify it to make it work better on my system. I miss that when I'm playing a game.

But games are art, and art is a little different than non-art. I don't want to spoil the fun of a game by looking at the source code for it; I like that the author has some definite vision for the game. I don't even like to visit a forum for a game in order to report a bug, because I'm bound to read about other bugs in the game. I frequently enjoy games that are moddable, but if I start looking at mods too soon, then I also feel like the game is spoiled. Early-access games are also somewhat problematic, although I've been playing Factorio for years now, with no regrets.

I wish there were a level in-between. If games were half-way open source, so that the platform layer was open source while the game mechanics and artistic vision were not, then I would be happier about it.

  Quake was only open-sourced years after it was superseded by more advanced games. It was still an awesome thing, and I'm glad that ID seems to have made it a tradition.

Not since id Tech 4 (Doom 3/Quake 4), now it's proprietary (thanks ZeniMax).

I feel like a good compromise would be to stick the code base for each big release into some sort of vault. Not open to the public, but for there to be an agreement that should Tarn ever be done with it, it can be released. Basically give keys (or copies) to Tarn's brother as well as a few of the DF community members that Tarn trusts the most.

This way it stays just Tarn's work of art for as long as he wants it to stay that way, but if anything did happen it wouldn't be lost to society.

Why not just open-source and continue solitarily working on and releasing his personal fork of it?

As in, Dwarf Fortress: Tarn Edition, with only one contributor, and multiple forks from other people taking the game in different directions, as well as projects that just incorporate the best features from multiple forks into one version.

The minute that happens the more accessible community version with polished graphics and interface will become the dominant one. Only dilettantes will play the original.

The problem is he has zero motive to do so. His income is adequate for his needs and he apparently has no interest in fame or other reward. To him there's nothing to gain and everything to lose from open-sourcing it.

As a counterpoint, I think that perhaps DF is a project that benefits from having a single mind and a single vision directing the course of the game and making it happen.

That's not a counterpoint to releasing it as OSS, though. There are lots of OSS projects where creators don't require community input.

> There are lots of OSS projects where creators don't require community input.

Then what is the reason that you want to urge Tarn and Zach Adams to release DF as open source?

As stated by meddlepal above: "My great fear for this game is the developer dies and we lose an amazing piece of tech with him."

Provided the license is at least Q-Public or stricter.

Otherwise expect a Neo-Dwarf-Fortress in no time to put pressure on the original author and let the world know that he is actually a hindrance to DF development.

NetHack is a kind of famous example where the Dev Team doesn't use a public version control repository, traditionally didn't use a public bugtracker, doesn't use a public mailing list or IRC channel, and traditionally has made releases less than once a year. They also didn't promise to respond to proposed patches or feature requests, though they often acted on them if they thought they were useful or thematic.

Some aspects of this have been controversial and there's certainly the idea that it's let NetHack improvements be much too slow, but it's existed for many years this way, and it's completely open source.

Also the Free Software Foundation's own in-house parts of GNU were largely this way especially in previous decades. There was an official maintainer who would make periodic releases. If you had an idea for the maintainer, you could e-mail and propose it. The maintainer might say "Thanks" and then put your code in a future release, or not. It was relatively hierarchical and secretive in terms of the actual development process, without a particular emphasis on being collaborative. In this case I've generally thought it was because FSF's philosophy didn't emphasize collaborative development methodology, as opposed to people's right to change their individual copies of software, and people's right to fork projects. We often associate free and open source software with collaborative and transparent development, but there are occasional cases where it's definitely not that.

I'm not sure if you are aware, nethack is now hosted on github and commits are public https://github.com/NetHack/NetHack

The 'Net' part of the project Nethack is so named because it was developed over the internet and discussed on the internet (newsgroups). It's the first game to be developed by multiple contributors over the internet and released on the internet that I am aware of.

The 'Open' of OpenBSD is more towards your point: NetBSD was open source and made releases, but did not show public commit history. When Theo's private repository privileges were removed, he was unable to contribute any longer because his patches against the latest public source was too far behind the current private source, so they would be rejected, but he had no ability to bring it up to date. And so "Open" of OpenBSD is because the commits are public and open -- the first to share their VCS history publicly that I'm aware of. I believe they invented the first "anonymous cvs" system for this very purpose.

> I'm not sure if you are aware, nethack is now hosted on github and commits are public

Wow, quite the big change from how they used to work!

> The 'Open' of OpenBSD is more towards your point

That's a great example, thanks.

I think the community would be better served reverse-engineering it in this case.

With self-taught single-person projects like these, there tends to be an inordinate amount of crazy code that only the creator has the experience to understand.

Short of patching and having a reference for re-implementation, I doubt the community could do much with the source code.

For what it's worth Tarn apparently has plans for the source code to be released if he dies of natural causes.

Ditto. I know it's his creation and he has the right to do what he likes with it, but it's such a damn missed opportunity. A game as complex and detailed as DF cries out for a vibrant coding/modding community.

on the topic of modding, its incredible what people have done with simply -memory hacking- the game! literally editing memory to insert mods after the game has begun. it has a limited scope due to the inherent risks but some of the work is incredible

Why don't all these people who cry for DF being open source and claim that it would create such a vibrant coding/modding community start to create their own game (with a similar spirit as DF) under open source conditions of their choice?

Because they aren't capable. It's far easier to bitch and moan than actually do something,

I've always wanted to start one, only in side-view instead of top-down, but I'm too lazy and it would take a long time to come close to DF's depth.

I'm thinking of more of a Lemmings-style game with DF mechanics and terrain deformation..

Because you don't need to be a master chef to tell if a meal needs more salt. "Never criticize anything unless you could personally could do it better" is a ridiculous notion mainly proposed by oversensitive creators.

Also because it's annoying and wasteful to have to recreate years of effort from scratch because the people who've done it before won't share. Which is, y'know, the entire point of open source.

> Because you don't need to be a master chef to tell if a meal needs more salt. "Never criticize anything unless you could personally could do it better" is a ridiculous notion mainly proposed by oversensitive creators.

The people who want to add their own additions better understand what the source code does. So this argument does not apply here.

> Also because it's annoying and wasteful to have to recreate years of effort from scratch because the people who've done it before won't share.

The original creators also had to invest the time. Where do the open source fans take the justification from that if they want something similar, but under their preferred license, they should not have to invest a similar amount of time?

I'm not aware of any OSS DF alternatives, but it seems like every year or so I hear about new games hitting Steam that seem (to me) to have clearly been designed by someone thinking "Let's see if I can make DF but with better graphics and ergonomics." Naturally they don't approach DF's depth, given how long Toady's been iterating. I don't hold that against them, but it seems like DF has the market cornered on sheer number of internal knobs to twist. Here are the main ones, I think (there are many others):

2014: gnomoria - http://gnomoria.com/

2015: prison architect - https://www.introversion.co.uk/prisonarchitect/pc.html

2016: rimworld - https://rimworldgame.com/

2017: stonehearth (alpha?) - http://www.stonehearth.net/

None of these are open source, ofc, and my expectations for how long they receive updates is not high. Toady built his rep over decades.

Aside: on the topic of hilariously complex, long lived, more or less open source games written on top of ridiculously poorly architected code with a tiny rabid fanbase - Space Station 13 deserves a mention. I'm confident that I've seen no other game like SS13 in my life. I can't even start to explain it. Please check out this high-level review of it if you're curious: https://www.youtube.com/watch?v=nLAHBexJxrE

I am not so sure - I sometimes wonder if the problem with a "platform" that many people love and use, but very few can contribute to well, is that when completely open sourced it inevitably comes bogged down in politics and overcome by "deletionists" or similar - people who don't or can't contribute to the project but see themselves as guardians of it and hamper the efforts of others.

Of course there's wikipedia as an example, but in open source gaming I feel that Dungeon Crawl Stone Soup went the same way, with new releases tending to remove features in the aim of some ideological balancing "purity" regardless of it being actually fun or not.

Therefore perhaps Dwarf Fortress will fare better with only very few passionate people working on it with impressive productivity. If it went fully open, I bet before long there's be a mission statement, a "DF way", statement of principles to which nobody can ever live up to, and then before you know it nobody can be bothered to fight to get a patch accepted (and forking brings its own maintenance hassles).

I didn't read the wish for open-sourcing it as a wish for it to become a community project, or even just a project that accepts any outside contributions.

What do you want from the source code then?

Hit-by-a-bus insurance, basically.

I haven't played much DF, but from what I remember there are quite a few tools around it, which already hook relatively deep into the game, either by simulating keypresses (I think) or reading/writing game memory. Development of those probably would profit from easier access to game internals.

There is "academic" interest in "how was this built".

Even a project not normally accepting contributions can do this on special occasions if it wants (If I'm not mistaken, several features in Minecraft build upon mods made by fans. Of course, Minecraft is also closed-source, but it is relatively easy to get good decompilations since it is written in Java. You might argue that the existence of tools around DF proves that source access is not necessary for improvements to it?)

For other games, I've often had a question along the lines of "I wonder how gameplay would change if ...". Sometimes that's solvable with a debugger and/or disassembler, but code would make it a lot easier. (With commercial games there is also the aspect of "can I share hacks/experiments like this", which admittedly isn't directly related to source access or a game being fully Open Source, and likely not an issue in the case of DF)

Other posters mention the long term issue of game development surviving its creator, but of course that doesn't require source publication now.

In general, I'd say it would be wrong to pressure a game creator in any way to open source their work, but I understand why people like games being open.

I know DF is a hobby for Toady, and he's not interested in contributors to the core game because he likes doing everything himself, which is fine. But I'd like to see a situation like Minecraft (also closed-source, but even obfuscated Java is easy to reverse-engineer) where the creator(s) work on the core game and modders can add stuff if they like, without official support.

That's sort of already happening with DF, but modding anything more substantial than data and font files is a huge bitch, because the code is fully compiled.

Reply on news.ycombinator.com