#mmo programming

[timetocode]

200 player test of nengi beta v0.2.0, with players flooding from the overworld instance to the underworld instance (through that cave).  Also wrote some view-scaling code that can handle anything between a phone and a 4k screen without distorting pixel art (a mixture of zooming and adding/removing more tiles around the edges). These screenshots are 2560x1440. This is all getting baked into a nengi demo and released soon.

Edit: looks like tumblr distorts these :/

[timetocode]

6 screenshots of the nengi beta 2 demo (unreleased, WIP)

I’m using Tiled Map Editor, a spritesheet from BrowserQuest, and Pixi.js. I’m not too good at actually making the maps yet, but I’m getting better.

It is a magical experience to go from the map editor, which is essentially a drawing program, to refreshing the browser and seeing my new additions to the world immediately. I could get used to this type of workflow... it is very content-centric, but it is wildly productive. I feel like I could make a very elaborate environment in just a few days. That said this isn’t my usual type of game with procedural terrain generation and destructible buildings. I’m sure I’ll eventually figure out how to hybridize these methods.

This is the first game intended to show a multi-instance world. A player will be able to walk around on the overworld (which is an instance), or walk into a cave and explore the underworld (which is another instance). The overworld and underworld will exist as two separate nengi instances and can be hosted in a very flexible ways. In the simplest case, the overworld and underworld can live together within the same server. But with nothing more than an ip:port:id these instances could be separated to two different machines. There could be one of each instance, or a thousand of each instance. There could be ten players, or there could be ten thousand players. The remaining limitation is how many players can fit in a single instance -- which depends heavily on what the game actually is and what hardware it is running on, but for nengi can be in the range of 50-300 players . So I hope some people wanna make some MMOs! And I’m talking about real MMOs, not just “we have a lot of players that exist in totally separate servers and can’t meet each other” -- nah, that’s a very loose meaning of mmo and was already easy to make in nengi beta 1. This time the players can actually walk/portal/teleport/travel-by-some-means from any part of the game world to any other part, even though the game world is potentially an aggregate of multiple servers.

I’d also like to mention that these instances can be created and associated in very dynamic or very static ways. We could just open one overworld instance and one underworld instance and say that the cave in the overworld goes to the underworld, and that the ladder in the underworld goes to the overworld. Or we could be fancy and create a server that goes to the cloud and turns on whichever instances we need and wires them up a bit abstractly. It could create an overworld instance.. and then when it starts to get full of players it could create an ‘overworld2′ instance and send players there instead. When a player in overworld2 walks into the cave to the underworld, perhaps the instance asks our fancy server where it should go. Maybe the the cave goes to underworld1, or underworld2, or underworld305, or somewhere totally illogical. It is entirely up to the developer how these things go together.

Edit: sources

http://www.pixijs.com/

http://www.mapeditor.org/

https://github.com/beeglebug/pixi-tiled

http://browserquest.mozilla.org/

https://github.com/mozilla/BrowserQuest

https://opengameart.org/content/browserquest-sprites-and-tiles

[timetocode]

Offering nengi.js walk-throughs

This is a FREE session where you tell me about your game idea (if you want) and I consult you on how to make it via nengi.js. These are VERY LIMITED, I’ll be doing 2 a week for a few weeks.. There can be actual screensharing and hands-on coding if you’d like, as we turn the nengi barebones template into the basic structure of your game (if applicable) or start to stub out something new. This prototype will use PIXI.js as a renderer, but I can explain how to change it to regular canvas or webgl. I can talk about current features and limitations, as well as the whole concept of how nengi works and gets its efficiency.

Requirements: 

* JavaScript experience and basic game dev experience (or fairly advanced game dev exp in another lang)

* strong desire to make a SIMPLISTIC multiplayer HTML5 game

* have a nengi license ( https://timetocode.com/create-account , 5% royalty on game revenue greater than $3000 per quarter, free for anything less )

Please note that there is no support for nengi. Beyond the very slim existing tutorials and this session, I can’t promise to consult or help in the future.

In return I’d like some feedback about nengi in general and my attempts at explaining it. It is also nice to meet fellow game devs! There’s no obligation to actually make or finish a game.

To apply please send an email to [email protected] with a small introduction including:

* your level of JavaScript experience

* your level of game development experience

* (optional) a simple multiplayer game idea that you’d like to work on

* (optional) a link to something you’ve made or attempted to make 

I can do these sessions some evenings on weekdays (pacific standard time) . I have discord, google hangouts, and skype. The latter two options are required for any screen sharing. 

https://timetocode.com/nengi

[timetocode]

A diagram of an infinite mmo world made with nengi.js instances. I’ll follow this up with a technical post that describes how it all works and in what challenges it might experience as it tries to handle thousands or hundreds of thousands of players.

[timetocode]

Analyzing a 100+ player game with nengi.Bot

I take a simple nengi demo game and show how to stress test it with bots. This is one of the less boring tutorials I’ve produced for nengi.. which is in open beta as of like 2 seconds ago. GL HF.

[timetocode]

Video of the 200 - 300 player stress tests

Here's the blog post about it, though the video covers more. http://timetocode.tumblr.com/post/63724469025/400-entities-1-player-300-bots-99-npcs-and-the

[timetocode]

tldw;

Overview of the ideas behind the mmo prototype. QuadTrees to decide which entities need to know about one another and reduce processing of crowded areas. Client-side prediction and entity interpolation to keep gameplay consistent despite the deliberately varying server perf.