What Can We Do

[ByteBit]

... it is to develop its successor.

If we really go for this, I think we should not base the new game on existing software. We should make a complete new project in a programming language everyone can understand (might go for C or C++) and work together in a huge community project. We can copy old player and grenade behaviour but we should not reside on all these old technologies were using, I think of rewriting our renderer completly and making a whole new client-side hacking proof protocol implementation.

But sooner or later we'll need our own forums and master server, so somebody will have to invest money in this.

Okay, let's start with design.

I'm here; ready to begin.

[Chameleon]

... it is to develop its successor.

OS, IB, VXW are successors of AoS... But they have been not good enough. OS is bad engine, IB lacks support of SW mode, VXW was based on Voxlap and was short-lived.
LeCom is working little by little on an SW+GLSL renderer (I talked with him last month, so idk how's progress). I guess we could merge IceBall and LeCom's renderer.

Okay, let's start with design.

I've already started (month idle, but I'm not abandoning it) with design, it's called "No Leaf Clover". Not as epic as google search results suggest, but good enough for testing ideas.

We should make a complete new project in a programming language everyone can understand (might go for C or C++) and work together in a huge community project.

You are literally talking about OpenSpades and a bit about IceBall.

[ByteBit]

You are literally talking about OpenSpades and a bit about IceBall.

This was not my intention. While OpenSpades is definitely written in C++, it was just a one-man project. I want us to talk together about implementation and how we want features to work.

I've already started (month idle, but I'm not abandoning it) with design, it's called "No Leaf Clover". Not as epic as google search results suggest, but good enough for testing ideas.

I wouldn't call that "design". It just a forked OS client with some new features and some performance improvements.
Also please stop advertising your project in every thread. We all know it already.

[Lincent]

If there is anything you guys could do it is to go work on Iceball.
No leaf clover is hopeless.

[Chameleon]

Also please stop advertising your project in every thread. We all know it already.

Ok, especially that it's been an idle month...

But one can never be sure how many people actually see that. I mean, I have 7500 views? OS introduction has 20000? And how many people just hit "Download The Game" button without looking anywhere else? Or just download from GitHub? So, making "N L C" appear on search engines might improve that.
On the other hand, it's just an unfinished compilation of ideas, not a very different game. So right now it wouldn't really deserve more than 5000.

Anyway, I wanted to point out that thinking of ideas and putting them into your fork is future-proofing them. Some guy with ideas in his head and fast code skills might see my ideas, look at my code and implement them himself into another game. Or if you talk about community-wide project, it is easier to copy written ideas from one project into another than writing those ideas yourself.
In practice, you'd take IB's engine, LeCom's SW renderer (WIP), ideas from VXW, ideas from NLC - and that's faster than a brand-new community-wide project.

@Lincent:
Oh yeah, OS/NLC are hopeless both engine and rendering-engine wise. NLC's more of a test-bed. A bit sad to abandon scripts that I worked on for 2 hours, but I might check out IB.

[bloodfox]

Honestly, I like the idea of bytebit that we should work on one game together. That would probably create some issues in itself like difference of opinions of where the game should go etc etc. Keep the game made by at least 3 to 4 guys maybe?

So, what are we waiting for, though? Let's all get into a chat, discuss the hell about what game we should make, get the more active members of this forum (like chameleon or lecom) to be head of making this game. Are we waiting for Lecom's renderer? My problem is the fact that we AREN'T doing anything. This is hellishly obvious but no one is gonna do something about it. I have a fair knowledge of C# which is forever growing and I'm ready to make something inspired by AoS.

[LeCom]

In your place I wouldn't rely on me writing that renderer, it's really hard and atm I'm stuck in a complex problem.
So here is what I suggest: first write a mediocre-ish placeholder renderer (the one from ByteBit's project seems simple and fast enough) and then an engine around it. The engine can have some kind of interface so we can just swap different renderers. This way we could actually support Voxlap until we have something better. Then, we'd need to think about the server software. Branch off of PySnip as I did for VXW? Write a new one?
I guess ENet would be the choice for networking

[ByteBit]

Let's all get into a chat

Why not have our own IRC room for this? I suggest #aosofthefuture.

Spoiler:

On QuakeNet of course

[longbyte1]

Let's all get into a chat

Why not have our own IRC room for this? I suggest #aosofthefuture.

Spoiler:

On QuakeNet of course

How about Slack? It's what the cool kids use nowadays, and it keeps a nice message history

No leaf clover is hopeless.

except 4chan lol

Anyway, here's our first question, and we'd better answer this one before we all go insane and the project goes under before we even begin: what language are we to use? We can go the pro route and pretend we're all good at C/C++. We can take the dirt path and do C#/.NET, although that might be risky given a lack of thorough compatibility with other platforms. We could do JavaScript, although I'd kinda be alone on that one, though the move to web might be a decisive factor against other voxel FPSes.

Second question is, are we going to start off with 0.75 compatibility or are we going to redesign the network protocol from scratch? (Obviously keep ENet since it's a very nice, portable UDP networking library.)

Third and final question for now: what are we taking, and what are we leaving behind in terms of past projects and their designs? I think for the most part they were coded excellently, but no matter how good your game is, you need hype to get people to play it. OS had hype, because the graphics were cool, even though it was quite buggy in its infancy; IB, while it was built strongly and based itself on a scripting environment, didn't have much hype going for it. If you really want to attract people, you could try adding two rendering engines: an OpenGL engine with beefy graphics and eye candy, and a raycasting engine for lower-end computers.

Naming is also critical: the best way to figure out great your name is is how well it fits in a conversation:
"Hey, you wanna play some Iceball with me?"
"Hey, get on my OS server!"
A name should be attractive; it should describe at minimum the theme of the game. "OpenSpades" does not sound as appealing as "Revenge of the Spades" or "Warblock".

[Lincent]

You guys always stress on making some shitty rip off of Ace of Spades, even tho its already been thrice.
The problem is getting people to play, not making another shittvwy ripoff.

[bloodfox]

I'm sorry, Lincent, am I hearing that you can code am making a contribution? No. Please leave.

Longbyte, I personally specialize in C# but I'm sure I'm the only one atm that actually is doing C#. We could go for C++ maybe but definitely not javascript. And, imo, please, let's not use javascript or make it a web based game. That would prevent me from playing a game like that. I think a bunch of people would agree with me on that.

Second, Warblock could work, I guess? Also, what's critical is also the picture for this game when you first click on the website. You need it to explain every awesome thing about this game.

P.S. Don't expect me to be on that IRC on weekends. Weekdays, yes.

[Chameleon]

I'm sorry, Lincent, am I hearing that you can code am making a contribution? No. Please leave.

Don't leave. I hear more constructive stuff coming from his mouth, and even though it's harsh - it's true (most of it).

Longbyte, I personally specialize in C# but I'm sure I'm the only one atm that actually is doing C#.

Every modern business-oriented programming course should touch C#, so you are not alone. I wouldn't recommend C#, because I hear that C++ or C can do better performance-wise.

Second, Warblock could work, I guess?

First rule of a voxel game - don't mention that it's blocky. Please.

I won't sit on IRC. I like slow, thought-out discussion that you can later look up (for cool ideas etc.).

Second question is, are we going to start off with 0.75 compatibility or are we going to redesign the network protocol from scratch?

Redesign, but ofc we should look at AoS protocol's both good solutions and bugs.

Third and final question for now: what are we taking, and what are we leaving behind?

I say that we forget eyecandy.
Take Iceball's strong engine, disable "toon" filter and add SW renderer.

About eyecandy...
Blur improves performance on SW (and still looks cool), but degrades performance on OpenGL.
Animated smoke particles look cool, but many people get big performance issues; so make cool, big explosions using cheap particles.
Dynamic light should be done a lot more cheaply - light up not parts of models/blocks separately, but whole models/blocks instead.
And add gradual shift from day to night, I believe that's all this game needs (that's all Voxlap lacked IMHO).
Oh, and my eye's not a camera, so forget lens flare.

[LeCom]

As language I suggest Dlang. It's not one of these hipster languages like lua or js, it's not owned by some shit company like C#, large code won't result in a mess like C and the language isn't a mess itself like C++. Being the official successor to C++, it's based on C/C++ and easy to learn for C/C++ programmers. It still has the speed of C.

Instead of saying it's blocky, you should say it's "retro style", which will actually attract people due to this retro style hype.

I never really understood how it's possible that smoke lags - alpha blits are cheap af for GPUs. One thing that has been neglected is gamma - fullscreen colour modulation that is "for free" even on PCs that are >15 years old. Good for blood effects and making everything more foggy.
Light effects only affecting parts of models are fine performance-wise, it can be optimized a lot for HD models.
Day and night is already simulated in AoS via fog. It can be improved via sky texture with sun or moon. Clouds would be really cool, they could be either drawn as a picture over the sky, or as real models or as 2D planes floating in a parallax style

[ByteBit]

Warblock. This name tells the user what the game is all about, but is it really only about blocks? AoS was always a game about shooting and modifiable terrain. War seems to fit great here, for the second part of our name I would go for a word that describes the ability to work with the environment. e.g. (trans-)form

This makes: Warform or FormWar

Also, I like the idea of using slack for discussion and we might use titanpad for brainstorming.

[Chameleon]

As language I suggest Dlang.

Well... I can't write my own engine, so I have no use for that really. But thanks anyway, I'll consider Dlang if I happen to write my own.

I never really understood how it's possible that smoke lags - alpha blits are cheap af for GPUs. One thing that has been neglected is gamma - fullscreen colour modulation that is "for free" even on PCs that are >15 years old. Good for blood effects and making everything more foggy.

I mean, OS has ~150-image long smoke animation and I've heard people say (plus my own experience) that it lags. Cheap, not animated smoke can be used in larger quantities so it looks good too. Gamma blood/dlight effect is nice, but over-do looks awful.

Light effects only affecting parts of models are fine performance-wise, it can be optimized a lot for HD models.

idk, maybe I wasn't clear enough.
You have origin of dlight, you have center of model, you calculate distance and apply same light intensity to all voxels on the model. If you have models that are not more than 2 map blocks in x/y/z axis, you'd end up with good enough results. Idk if that's already in OS SW dlight code.
Similar to map blocks, you'd take center of a map block and change the colour of whole block. If you play around with OpenGL dlights in OS, you can see that dlights are smooth. I aim for blocky dlights and I believe they'd have way better performance.

Day and night is already simulated in AoS via fog.

I know. But what I mean is make block colours themselves depend on the fog, even up close.
I've made this a while ago: http://buildandshoot.com/viewtopic.php? ... ilit=night

It can be improved via sky texture with sun or moon. Clouds would be really cool, they could be either drawn as a picture over the sky, or as real models or as 2D planes floating in a parallax style

If it's going to be possible to change direction of shadow (it is in OS), then proper daycycle would be insanely cool. Ofc keep shadows blocky and performance-over-quality.