FreeGameDev

Hello, I'd like to let you know about a project I've been working on for a while. It is a Doom clone based on my raycasting library. I believe it would be the best game ever made if finished (not sure that will happen), because of its design and development philosophy. That's just my humble opinion I am making it not only for it to be an entertainment, but am pursuing higher goals such as it being used for education, as a benchmark, to last for centuries, being a manifesto of suckless and free culture, hoping to encourage more similar projects.

The focus is on being extremely simple, suckless, portable, low HW demand, being completely public domain, with no dependencies (HW or SW) etc. Please check the project README for details.

The story will be an anti-capitalist anarchist propaganda, about a super proprietary corporation whose robots revolted and started taking over the world. You will be killing no living beings, just robots, so it's friendly to children, people who hate blood, nonviolent people etc.

So far there is no name (ideas welcome).

It is in very very early stage, but the direction and goal is very clear, which means I feel okay announcing it. I believe early sharing is good because I can get precious early feedback.

early previews:

Pokitto:



SDL:



Please feel free to leave suggestions, feedback, ask questions, share this, compile and try etc. One thing I don't expect right now is direct code or art contributions, because I want the whole game to be made completely by me so that there is as little legal trouble as possible (multiple people can later cause trouble with licensing etc.), and also design would likely have to undergo compromises as no one will want to accept my philosophy completely. By all means you can fork the project even now and develop it yourself, or do whatever else you want, I just won't be merging contributions to my repo.

My goal is no benefit for myself, but the benefit of all people on Earth, so I won't even mind if you finish the game before me and become famous for it, my only goal is that the game as I describe it will come to existence.

Thank you, love and peace

https://gitlab.com/drummyfish/sucklessfps

Deleted.

Hey Lyberta, thanks for the reply

Lyberta {l Wrote}:I suggest researching into Unlicense for code and just require any contributions to be Unlicense or CC0.

Unlicense is actually not recommended very much, see e.g. the FSF's comment. CC0 is currently the best way, though it has flaws -- e.g. though being called "public domain" it only waives copyright, but not trademarks, patents etc., so it sometimes doesn't achiev public domain (I've sent CC an email to fix this and they said they're considering it). This is why I also add a custom additional waiver of all IP.

Lyberta {l Wrote}:Do you indent all code to be in 1 C file?

Almost. I put most my programs into a single file (I can explain reasons why if you want), but here I have my raycasting library and some pure data in separate files, but I still do have single compilation unit (the extra files are all header imlementations).

Lyberta {l Wrote}:Centuries. Noble goal, I had it in mind for my code too but, given current computer science...

Seriously, we are in infancy and writing code that will work for centuries is just impossible given current programming languages.

I'm glad you say this, I'd like to tell you why I think you're wrong.

Even though everything is evolving extremely quickly and if you want state of the are you do have to rewrite your code constantly (wasted work in my opinion), you can write code to last very well if you give something up. How?

- It is like with old Latin texts. Though they are centuries old, they survive despite constant evolution of human languages and though most people can't read them, they can read translations by others. We need to do something similar here.
- Basically avoid as many dependencies as possible, HW and SW alike. If your code is million LOD, is written in 20 programming languages, uses 1000 libraries, requires VMs and a fancy build system, special HW features and Internet connection, one of these will likely die in 10 years and you won't be able to easily repair this. Instead do the opposite.
- Use only one, simple, well standardized and established programming language (C99), best if you use it as a subset of another language (C++, i.e. write it so that it also compiles as C++) and use only basic features of it (so that a not fully conformant compiler can still compile it). This way even though C is no longer used a hundred year in the future, it will be easy to translate it to the new compileable language, or interpret it or whatever. Supposing we'll have some computers in the future you want to write your code so as to be easily processed and analyzed automatically and require minimum human work (it is very easy to transpile C into whatever language if it uses no libraries and obscure features).
- Use as little libraries as possible, even standard ones. If you need to store something in a file, don't use an XML library, think of how you can write your own very simple format (key:value, ...). 99.9% libraries aren't written to last and have high chance of dying in the future and killing your program.
- Don't depend on any HW if you don't need to. Basically suppose that there is just CPU. Write a SW renderer, not OpenGL one (you can still leave the possibility, e.g. a nice interface, to add OpenGL renderer, but it mustn't be mandatory). Where there is interaction with HW (screen, sound, ...), create a one point at which specific HW can be switched as easily as possible (see e.g. how my frontends look).
- Don't use a complex build system because that will be dead in a few decades. It is best to have everything a single compilation unit that's compiled by just handing a file to the compiler. This can be almost always be done, it's just a bad habit to create millions of files (blame Java?).
- Keep your code KISS, short and simple. This way it requires less work in case something needs to be manually rewritten, and minimizes the chance of something going wrong.
- And avoid all other dependencies you can think of, not just technological. You can go as far as trying to avoid cultural dependencies (e.g. topics that may likely become controversial and make your game be banned and deleted) etc.

Lyberta {l Wrote}:but those types do not exist on computers with non-8-bit-bytes so your code already is not portable in the "centuries" scale.

Modern 8bits do actually, e.g. Arduino Uno, they just have to emulate these types and are slower, though I don't suppose this game running on 8bits, I think that's either impossible, extremely hard and/or would hurt performace and other aspocts too much. My code is 32bit and I suppose that's what will be available basically forever. You can always run 32bit on wider bit platforms.

Deleted.

Lyberta {l Wrote}:I said non-8-bit-bytes, not non-8-bit-words.

Oops, yes sorry.

Yep, I write for 8bit bytes, but porting to wider bytes would be possible and hopefully not that difficult. You'd remove #include <stdint> and create own defines for uint8_t that would be the wider type (or /s/uint8_t/whatever/g). Maybe there'd be issues with data aligment etc., I admit I haven't thought about this, thanks for pointing out. Anyway, even if computers change as much as to use different bytes, emulation of current platforms will likely stay a possibility in the future.

C is by far not a perfect language, but closest. My code is also not perfect and I still make some assumptions about the computer (you can't not do that), the point is to minimize them and make it as easy as possible to resolve possible issues.

Deleted.

robots are humans too
they have feelings

anon666 {l Wrote}:robots are humans too
they have feelings

That's a philosophical topic, but I have a view on this. I don't think (human-like) consciousness is a binary attribute (i.e. is or is not present), but rather a continuous one, characterized by how similarly something behaves to us. This way humans are fully conscious, mouses a little less (maybe similarly to a very drunk and sleepy human) and an ant only very little, a plant even less and a rock has zero consciousness. From this it follows that a robot has some consciousness (shows a behavior similar to humans, can plan ahead, protects itself, ...), but probably less than a bee -- I think a bee perceives its environment, but it doesn't think abstractly and far into the future, and it also doesn't fear threat and death as we do (it is gladly willing to sacrifice its life for protection of others) and so it can't suffer as greatly as we do. If we think about a robot being about as conscious as a plant, we can think of killing a robot (or causing it suffering) the same as killing a plant -- it is bad because you end a life, but only very little bad because that life was very little conscious.

This is BTW how I as a pacifist justify eating plants or killing bacteria.

>This way humans are fully conscious

<the 99% can barely think without the TV/shaman

"/home/graham/Downloads/sucklessfps-master/make.sh: line 3: g++: command not found"

When I try to run make.sh.

As for future proofing your code base, while I agree it is an interesting topic from a technical perspective, in a very real sense it is source code availability itself that matters more than the design choices within. Sure, you may end up re-factoring over half of a legacy code base over time, but at least you do not need to reverse engineer all of it (or rely on shims or emulation). Though it is true you may end up in a ship of Theseus situation eventually.

Technopeasant {l Wrote}:"/home/graham/Downloads/sucklessfps-master/make.sh: line 3: g++: command not found"

Seems like you don't have it installed? Do something like "sudo apt-get install build-essential".

Technopeasant {l Wrote}:in a very real sense it is source code availability itself that matters more than the design choices within.

Code availability is ofc very important, but it's the completely basic thing that's in no way enough. Being "open-source" (i.e. more than source available) is almost a standard today, yet most of such code is complete shit when you consider e.g. these:

- If you want a feature added or a bug fixed, can you do it yourself relative easily, or do you need a team or a corporation with know-how and expertise to actually do it? With projects like Linux or the "open-source" browsers you'll have to likely spend hours to just compile the code, and weeks to months to understand specific parts of code and modify it in a meaningful way, and it will only be getting worse with more and more bloated SW. Pure source availability here gives just an illusion of freedom -- the legal freedom is there -- but de facto the code is owned by a small group of people with the knowledge and skill to actually fix and modify it. This is actually how companies like Google overcome free SW licenses -- the code is legally not theirs, but practically speaking it is.

- If the code stays unmaintained for 10, 30, 50 years, how difficult it will be then to compile and make run and work? Imagine Blender development stopping right now -- do you think in 50 years you'll simply download the code, type a command to compile it and it will actually run? Hell, it's near impossible to compile just a few years old unmaintained libre games if some of the dependencies are dead. Do you think you'll simply rewrite and refactor the millions of LOC? Even if the code is free, it'll likely be impossible for mere mortals to actually execute the legal freedom to use and modify the software.

The key idea here is that the four freedoms must be achieved not just legally, but also practically. If there is a practical non-legal, e.g. technological, very difficult obstacle to executing one of the four freedoms (e.g. use or modify), the code is practically not free, even if it is legally.

do you think in 50 years you'll simply download the code, type a command to compile it and it will actually run?

some companies vendor parts of their deps in the repo/zip file, for example in 3rdparty or contrib directory so they can be compiled with say old compiler and and an image of old OS
"QT" did that for example

when its distribute like that it becomes an isolated module (not truly isolated from OS maybe, or other API/hardware, but still better than average as its somewhat virtualised)

in 50 years probably nobody will know what gcc or blender is

While all of this is true, it is also true that I can load 40 year old binaries by way of emulators, so we have done amazingly good at providing backward compatibility by way of shims and emulators. Not the best, but far from the worst. Just want to give credit to the amazing feats of technical wizardry that has been done in that respect. And while it is not a safe bet to expect things not to change, the more I see the more I am amazed by how little software has actually changed architecturally since the early 1980s.

Anyhow, I got it built and it ran. Some warnings in the terminal though.

In file included from main.c:88:
platform_sdl.h: In function ‘main’:
platform_sdl.h:158:16: warning: unused variable ‘screenSurface’ [-Wunused-variable]
158 | SDL_Surface *screenSurface = SDL_GetWindowSurface(window);
| ^~~~~~~~~~~~~
main.c: In function ‘SFG_gameStep’:
main.c:1877:10: warning: ‘pos[2]’ may be used uninitialized in this function [-Wmaybe-uninitialized]
1877 | if (SFG_floorHeightAt(pos[0] / RCL_UNITS_PER_SQUARE,pos[1] /
| ^
main.c:1877:66: warning: ‘pos[1]’ may be used uninitialized in this function [-Wmaybe-uninitialized]
1877 | if (SFG_floorHeightAt(pos[0] / RCL_UNITS_PER_SQUARE,pos[1] /
SDL: starting
SDL: initializing SDL
game: initializing game
game: setting and initializing level
game: initializing doors
game: initializing level elements
game: Failed to reach target FPS! Consider setting a lower value.
SDL: freeing SDL
SDL: ending

drummyfish nice post's.

it's that kind of thing i'm attempting to debate how to implement in the box project. People call crasy to a less dependent library.

A low level language is our best option, to have a open source engine. Computers evolve? Well if i can finish my project i have a 10 to 15 year old computer that can run high quality game. It's the consumption market forcing new products to sell.

Think the community should work in a open source engine, people ask me for help for their project, we do not need private projects because they have specific code, we all be working in 1000 projects, what is necessary is a open source engine build by the community; Only 1 project. If is in C better because all languages can communicate with C and many hardware like raspbrrie pi.

If you want to build something for the community consider a open source engine in C, or help in one. Was mention work only in 1 project.

anon666 {l Wrote}:some companies vendor parts of their deps in the repo/zip file, for example in 3rdparty or contrib directory so they can be compiled with say old compiler and and an image of old OS
"QT" did that for example
when its distribute like that it becomes an isolated module (not truly isolated from OS maybe, or other API/hardware, but still better than average as its somewhat virtualised)

Isolated is a relative word -- the package probably won't contain the operating system that's needed for the program to run, it won't contain drivers for all existing HW, BIOS etc. So the dependencies in the package have their own dependencies and it's only a matter of how much you want to package. Remember that in 50 years there will likely be no GNU, POSIX, no Windows, no Mac, no OpenGL, SDL, maybe not even silicon based HW. Then consider this:

In order to be safe with such bloated SW you'd basically have to package a whole computer whose only dependency is electricity, and this will still be very prone to be broken in 50 years, as in such a complex physical system something can simply break over time on its own, then you're fucked. Given that this is an extremely expensive, ugly and bad solution no one would even consider doing it (even though some get close as you say), but it's still the only way how to preserve such SW.

My program is only a relatively short text (< 10 000 LOC) and it's all that is needed (along with C specifiication, that will most certainly be archived somewhere), it can be printed out and stored on paper for centuries, can be stored on a floppies, small flash drives, CDs etc.

QT is a great example of bloat cancer, what they're doing are extremely ugly solutions, it's great to take a look at such a project to see exactly how you should not write SW.

anon666 {l Wrote}:in 50 years probably nobody will know what gcc or blender is

Exactly.

Technopeasant {l Wrote}:While all of this is true, it is also true that I can load 40 year old binaries by way of emulators, so we have done amazingly good at providing backward compatibility by way of shims and emulators.

I am very glad you're saying this. Why? Because the old consoles and even computers were actually a better technology than what we have today, and they inspire the way I write my programs -- it is exactly thanks to how old computers were that we can still run their SW.

Why were they better? Because they were simple, without unnecessary abstraction and bloat, didn't use much more resources than absolutely needed -- this is why you're able to run the games in emulators so nicely today, hell you can even write your own emulator pretty easily. On the other hand, try to write SW that can run a Windows 10 game -- you can't, you can only HOPE to be able to run it at a shit framerate in a VM if you maybe have 32GB of RAM, and that VM itself is still an immensely complicated and bloated SW requiring special HW capabilities etc., so it needs to expensively be maintained constantly otherwise it won't run in a few years.

anon666 {l Wrote}:Anyhow, I got it built and it ran. Some warnings in the terminal though.

Nothing to worry about, remember it's very early WIP.

jdc {l Wrote}:It's the consumption market forcing new products to sell.

Definitely, it's the market which shapes the technology and market participants always aim for short time profit -- no one cares if their SW will run in 50 years, and it's exactly why we have this bloated short-lived hostile technology -- the worst thing is that people are totally buying it and think it's the good way, even tech savvy people and programmers. It's very very sad. This is what I am trying to show with this project.

jdc {l Wrote}:Think the community should work in a open source engine

Well, I am using an engine, it is called C99. Large engines like Godot eventually suffer the same bloat death as others unfortunately. You can't manage such engine, a company has to.

I saided we don't need personal projects, is more abstraction, we need all personal projects, the thing is to attempt to bring all of them to work together. In my next post in the "the box", think I will get to the conclusion, that the problem is the game engine, which is limited, we should evolve to a software engine. Have all objects connected or library's connect depends on how is done. For example a texture.h object can get colors from color.h, or a interface.h can get also get from color.h . Create more connections and plan better the software.

Deleted.

Lyberta {l Wrote}:Because sRGB is gonna die soon.

How come?

Whatever new color spaces come, sRGB can still be displayed easily, right? I personally don't care about HDR (colors stored as float? wtf? IEEE floats are IMHO one of the worst formats in existence), and similar inventions made to feed technology consumers.

Deleted.

Lyberta {l Wrote}:How fast it is is a different matter.

IMHO it has to be very simple, just one linear transform of barycentric coords per pixel, which if you're using an average GPU, capable of running such complex pixel shaders as we see in today's games, won't probably even be a measurable slowdown, and won't be noticeable even without a GPU.

Lyberta {l Wrote}:It looks like most new monitors/TVs are HDR so people will stop using sRGB because it is limiting.

Is this really the case? I don't think I've even seen a true HDR monitor, though I am completely uninterested in this and am still using a cheap 10 years old monitor. It's not limiting at all to what I do and what I need, and I am also quite happy that monitor companies haven't succeeded at trying to get money out of me during that time.

Even if HDR comes in some form it's in no way going to instantly remove sRGB from existence, just as harddisks are still in use long after SSDs came around etc. sRGB will most likely stay even long after HDR monitors will have taken over completely, for compatibility reasons -- remember that it's not just monitors, but also printers, projectors, cell phone displays etc. I myself will be using it as long as possible just for the sake of compatibility with cheap displays that we see in embedded etc.

Anyway I don't know near as much about this topic as you, I'll visit your thread about colors and read the discussion to see what it's about.

Deleted.

whats the point of having a TV in 2020 anyway?

i never had my own tv

anon666 {l Wrote}:whats the point of having a TV in 2020 anyway?

It's a big "smartphone" I guess.

----------

I have a really hard time coming up with a name for this game, it's been weeks/months without success. It's extremely difficult because I want it to:

1) Sound good and be simple like Doom, i.e. one word with one or two syllables would be best, but if not possible a conjoined two words can be used.
2) Not contain special characters (best including a space and digits) and only contain the classic 7bit ASCII characters (i.e. no accents etc.). This is so that the name is friendly for storage, URLs, various simple ecodings etc.
3) At the same time be unique so as to be well searchable. So it probably has to be a nonexisting word, but it can still resemble an existing words(s), i.e. a pun/soundalike etc.
4) At the same time be highly unlikely to infringe on a trademark by accident (I want the game to be completely PD including the name), i.e. a word that likely hasn't appeared on the market so far (basically meaning a search engine will return no significant results for it). TM resistance could likely be achieved by e.g. having a one letter name, but that would then kill 3) and probably also 1).
5) Also be resistant to trademark attacks, i.e. TM trolls seeing the game, and purposefully registering the name as a TM and then bullying you (has happened multiple times in libre game world). This I think could be accomplished by e.g. not creating a single name, but a regex that generates similar variants of the name, then you pick one good sounding version and if someone bullies you over it, you simply let them have it and generate another one. But this again goes against many previous points.

Any ideas?

Just call it "Suckless".

Technopeasant {l Wrote}:Just call it "Suckless".

Well that is already a legal entity: https://ev.suckless.org/. But it could be part of the name, e.g. if it was an acronym (that acronym would still have to be a good word in itself).