Knytt Online World

I go with salmoneous, i hate rpg's it doesnt fit in knytt at all
Going with a knytt experiment world multiplayer would be great, finishing levels co-op and stuff like that

Knytt online seems very possible.
We need someone to host the initial router service so we can all be connected.
It does take up alot of space to hold so many users data and the game doesn't seem to  be getting done.

You'll need a dedicated server to support many members.

If you're a good programmer, you might be able to use a max of 4kb of data to transfer between the server and the player.

4kb seems alot, but I'll explain later why I say 4kb.

If you have 4kbyte of data, you have 4x8=32 kbit of data.

If you have a good ADSL conncetion, you usually have 1 Mbit upload. 1 Mbit is 1024kBit.

If you were to allow your entire upload stream to be used, you would be able to host: 1024/32 = 32 users. Unfortunatelly, you won't be able to get a constant 1024 upload stream. (yes, they promise you to get it, but usually you get between 600~800kbit. Lets assume you can have 800. You'd get 800/32 = 25 users max.

Now obviously, you'd not sacrefise your entire upload stream for this. Even if you have a good connection, it is unlikelly you'll give up that much for a game like this. Not to mention, that at that point you'll be going to put your pc online all the time, otherwise people cannot connect.

Seriously, unless you have a dedicated server (which costs alot of money) this is a no-go for a MMO.

Now, I mentioned earlier why you'd need atleast 4kb of data.

4kb of data is 4096 characters in one transmission.

Sending out data to the server is not a problem, but unluckelly, thats not the place where alot of information is being send. Its the receiving part that has the problem.

When you send your data, you'll be sending your location along with an identifier for your player. The server needs to know who you are so it can update your location in its map.

It will then send back what is happening around you. (sending every data would be a bandwidth killer anyway)

Say, there are 10 users in the same screen as you, the server would be sending you their locations and their names.

Depending on wether you're going to store the world on the server or on the client using updates would also be a role in this.

Lets assume you store the world on the server, the first thing you'd need to transmit, is the screen's layout. (so all the layers and where you can walk etc). This can happen every time you visit a world without getting who's on the screen yet. (this will most likelly also take up 4kb)

Next you'll need to store all the users that are on the screen with their names. You'll need a precise (pixel precise) location of the users along with their names.

So lets say the screen is 600 pixels wide and 240 pixels height, you'd need 2 bytes (256*256) for the width, and 1 byte (256) for the height followed by the name.

Now obviously someone would want to chat too, so if there's a message, we would need to send the message too.

Now, if you add functionality, like: the player is jumping etc, these actions should also be added to this list, so you can make it look more natural and less jagged on the client side. (otherwise, players will appear to warp through the screen rather than walk.

As you see, its not as easy as it looks.

I'd make just a server and client button, allowing users to join servers, or make their own. For the server a connection to an online central database that stores all the servers and their IP, and clients retrieve this list. This divides the amount of connections, everyone can show their level, are responsible for their level to keep online, and set a max join to not overload their upload.

As a developer there's no need to lift all the burdens on your own shoulders, pass it to your users if they want it all so much ^^.

If only this blasted mysql connection fromout MMF2 was working...

I think it's pretty silly to argue about how this "game" should be. If nobody can agree on anything about it, there's no point in thinking about it, because it would only work if people effectively collaborated on it (and I think this is ultimately not going to happen).
The only way I see it potentially working is if Nifflas (or someone else who is well-respected here and high in status) actually works on the idea and finishes it in a way that is accepted by at least half of the community here (100%, even 75% is probably not going to happen here no matter what). If that doesn't happen, people will probably just go with their own ideas of what makes a good online game and nothing will be agreed on. >_>
I, for example, think this whole idea is pointless. I much prefer playing games on my own where silly kids won't get in my way or spam nonsense at me.

And I think shit tastes better than candy. Sorry, I thought he was making a sarcastic joke too.

And that is rude (I think this violates at least three rules).

I tried to make that point, but people just submit their ideas along .
If anyone comes with a working version, no matter how it works, it'll be gladly accepted by everyone. I'm sure of that. Improvements can always be made on the way.

I once read the proposal of making a MMO Knytt without chat. I think that's a great idea. Only communicate through movements and animations. But as 4a states, if nobody works on it, nothing will happen, and all ideas will be in vain :p.

If you're going to use a website script to keep all the bits and pieces together, you'll get alot of performance issues. The only way to do this properly, is to make the game flashbased, so you can use it to directly connect to the mysql database. (or another form of webclient, like java, javascript, etc)

MMF2 has a database connector, but I have not yet got it working properly. Otherwise I'd do it via the mysql API in C. It's quite easy actually. Once you, as a client, have the ip to your desired server, for what I have read you can use MooGame to connect to this server. Once that works, rests a rebuild of whole Knytt Stories to implement multiplayer characters, which is the most hard thing, probably.


Questions arise like:

How do items impact on multiplayer... can multiple users have the same items? Or do they need to divide them? If player A walks through a red key block, can player B pass as well? Can players share items? It's easy to say yes to all of these questions, but is that really such a wise decision, gamedesign-wise and technical-wise?

Should chat be implemented? Or would that ruin the actual meaning of Knytt? To create a great experience, a moody environment without LOLs and RAWRs... Are sore animations perhaps a better way to play with eachother? Then again, many users may find the game too boring and cease playing, leaving the online worlds utterly empty (but however it turns out, I don't expect to have more than 10 players at the same time playing anyway).


The more you say "let's do the most awesome option", the longer one is busy to develop. Until your project became so big it will be abandoned. This is one of the main reasons why projects fail. Keeping it simple is very important, and it may prove to function much better than a game with all kind of jingles.

So, anyway, just to make a point: if you happen to be interested in following a course on gamedesign later in your life: try to think for every idea that comes up how much it is related to your actual goal. Is the trouble of implementing your feature really a benefit for the game, beside that it sounds awesome? Are there perhaps other, more realistic or simple ways of reaching the true goal that I want? Keep being skeptic about your ideas and always keep in mind that the most simple solution is often the best.

It might be interesting to have a setup like The Endless Forest's: players can communicate, but only non-verbally.

Will it be within a browser (Like Kingdom of lothing) or a .exe file?

Will it be within a browser (Like Kingdom of lothing) or a .exe file?

I'm hoping it will be in a browser because, if it were a .exe that changes, it would be a HUGE one! D:

I don't see how you could run a MMF game in a browser. (Although it's already possible to run GM games in a browser.) And the .exe itself wouldn't necessarily have to change. But, unless you want to play an extremely dumbed-down game, the game will have to be reasonably large. It might be possible to maybe have areas themselves just loaded from a server as you enter them, and kept in a cache or something. But I don't know that they'd be that big anyway.

I don't see how you could run a MMF game in a browser.

Not sure it it is the right thing, and not sure if it works, but after google...

http://www.clickteam.com/epicenter/ubbthreads.php?ubb=showflat&Number=89235

(Ctrl+F>Browser)

I don't see how you could run a MMF game in a browser.

If I remember correctly, you change the build type to an internet application and embed it in a webpage.

Ic.

In order to connect to a mysql database, you must have mysql installed locally. If you have a webserver with mysql installed on it, it is most likelly the case that a firewall blocks your connection to mysql because its not locally. You can talk with your webhosting to verify that, and they might be able to open a connection for your ip so you can develop.

My hostings, Nifflas' hosting and J's hosting, all block mysql connections if they're not locally unless you have an IP added to the whitelist of the firewall and in Nifflas' and my case, also in cPanel.

Ofcource, you could also compile the build, upload it, and then test it.