Game Development

I mean, we're done here I can't put up with your shit for even 2 more seconds, but I hope you will try harder to open your mind in the future when someone wants to help solve your problem.

Siddharth

sorry I need to talk with some other person regarding this

thanks for your time given to me

MickLH

When you fail to understand something, you have to give up on the logic you are using and try something else.

Siddharth

here in my code some problem exist in you are talking about some other things

MickLH

4:16 PM

You are the problem here, not anything in the code or in the design.

Siddharth

depend on person :)

MickLH

No. Objectively

No opinion involved. Whatsoever.

You've failed to consider or communicate the true requirements of your logical system.

Siddharth

I will discuss this with my other friends, I am not angry regarding you because you shared your opinion with me

MickLH

You've only thought deeply enough to guess at what you think you need, but have no analytic methodology.

Siddharth

so please don't consider anything wrong about me :)

MickLH

4:18 PM

I don't judge people, but I analyze problems.

Siddharth

it is just matter of understanding

MickLH

This problem is EXTREMELY straightforward, the ONLY issue here is your failure to communicate the necessary details I've asked for directly several times.

The algorithm is a very "1 + 1" math problem, but it's "1 + x" and you won't tell me "x"

Siddharth

now don't think about this more, just cool down :)

MickLH

You must be speaking to yourself?

I hope you're not "heated" about the way I am, I am not sorry that I am harsh

Siddharth

no no its okay, don't worry

really thanks for your time

user92578

4:24 PM

he'll probably just ping jon tomorrow about the same issue

Kevin van der Velden

The ignore button solves so much

MickLH

@KevinvanderVelden lol the real question is who should ignore who

I guess everyone could ignore everyone and then we can all have private conversations and never have to risk someone having an opposing view about anything!

Kevin van der Velden

It's not that

It's that he constantly does the same thing and doesn't learn

MickLH

I just thought it was funny, considering the extreme limit of using ignore to the maximum :P

It would be like multiplayer notepad with radar off

@KevinvanderVelden I kinda noticed this... My whole attitude shifted the moment I encountered that repulsion to actually doing analytic thought and problem solving.

Kevin van der Velden

I've stopped even trying to help him weeks ago

But now I've finally had enough of reading it

MickLH

4:30 PM

I was actually trying to set out to be friendly, I remember seeing his name on GDSE this morning and felt like I should reach out

Also to try and help fix my own racism because I've become severely racist after some clients that insisted on outsourcing tasks I defined for their project... to.. odesk

.....backfired

I'm seriously scouring for any evidence that I should not be racist.. and going out of my way to ignore the vast amounts of evidence the support it... and I'm really struggling

DH.

He probably doesn't understand everything we type, so he just answer randomly

MickLH

My strongest silver bullet against racism is only "All people are unique."

I seriously don't have a good real world example. It makes it a lot easier to see why so many people around me are racist, most of my neighbors are not as abstract and analytic as I am about things. They prefer real world examples

But I can reject at least a few million more bad examples, before statistical significance at least, so I'm safe for now

Expected value is looking pretty fucking bad though...

user92578

4:46 PM

I'm starting driving school tomorrow

user92578

Their awesome dynamic lesson managing website is awful

user92578

i have no clue if I'm supposed to do something

MickLH

@Tyyppi_77 Nice is this regular day to day driving, or defensive driving?

user92578

just the process to get me a drivers license

MickLH

Oh man so a basic class is required for a license at all? We really need that here lol

I don't know how the state of your average roads and drivers are, but I think a good defensive driving class is very worthwhile for anyone

Kevin van der Velden

4:56 PM

That's the good thing about requiring classes (and good tests) before allowing people to drive

You don't need a defensive driving class to not get killed

user92578

tomorrow is just theory

user92578

roadsigns and stuff like that

user92578

but it's a longish process getting a license here

user92578

and costly, ~2K for a piece of plastic

Sie

@MickLH Wait what happened? I missed another fire?

Ah I see now.

Jon

5:11 PM

don't fart on the fire.

Sie

I used to be a lot like that and still fight it off when I notice myself doing it. I always try to understand the code as best as I can so that I actually learn from it. I mean sure you can grab the code from SE/SO but really if you don't understand it you've learned nothing.

DH.

5:24 PM

Yeah, no shame in copy-pasting, but at least understand what the code is doing

Sie

5:41 PM

So I just finished working on a weapon swap script. However I don't want it swapping in the middle of an action (as the weapon swap is tied to the mouse wheel as are a few other critical controls). So I figure that I'm going to have a bool on each weapon script and have the weapon script check to make sure it is false before it swaps. Is that the best way to do this?

I'm assuming so but I'm not positive.

DH.

Just make a single bool that means that the character "is busy with some random action" to avoid moving, swaping weapons middle attack, etc.

Jon

6:08 PM

sounds like you are doing it right.

instead of having a bool in each one though, would be better to have it in the base class

Almo

still liking offworld trading company

not played against people yet thoguh

Jon

i'm playing broforce now

Almo

pretty good game (broforce)

Jon

6:26 PM

yah

who's your fav bro?

Indiana brones is shit

Sie

Happy mothers day Jon.

Almo

6:42 PM

@Jon probably Snake. I like the long range.

Jon

7:01 PM

you like my walking car?

user92578

Looks retarded. I like it

DH.

I wonder how a hill walking game can go wrong with this kind of movement

Would be fun

Darth_Wardy

7:19 PM

So I got my rendering code working and my networking code, now I've lost the will to carry on

MickLH

7:51 PM

@Darth_Wardy Apply 10ccs of pain directly to the face and threaten to do it again if you don't keep working

StackExchange

8:04 PM

0

Q: Are questions about games created using Scratch on-topic?

There is quite the library of appreciated games made using MIT's Scratch program. The program uses stackable blocks and pieces instead of raw code. I'm wondering whether discussion or asking for debugging help about development of games using Scratch would be on-topic here.

Darth_Wardy

8:58 PM

@MickLH meh ... i lack the motivation to even do that

this heat is doing my nut in

MickLH

9:09 PM

@Darth_Wardy bruh...

Darth_Wardy

I feel like fighting bugs in game engines is not how I want to spend my life

but really don't want to write my own

PearsonArtPhoto

LOL. Fighting bugs in game engines seems to be an inevitability.

Darth_Wardy

these days yeh

people apparently don't write well written engines these days

MickLH

@Darth_Wardy lol nice punchline

Darth_Wardy

that's a lot of eefort for a game idea I have in mind

MickLH

9:12 PM

I immediately thought writing your own engine is the source of bugs :P

Darth_Wardy

I just want to write the game not the whole dam stack

@MickLH yeh but at least that way when it all goes sideways you can just step through it

if its someone elses black box of code its not so easy

MickLH

Yeah but every single day... there is not ever a day where I don't have to fix something lol

I think writing the engine from scratch has cost me more time than I've spent on the actual game in total

Darth_Wardy

I don't mind fixing bugs in my code, I shouldn't have put them there, but crappy issues that should be trivial / not there at all is just cringe worthy

@MickLH yeh that's the problem, and I've tried so desperately to avoid it, but doing so has cost me more time than just writing a game engine lol

PearsonArtPhoto

Writing your own, using someone elses. It's bound to have bugs.

The advantage of writing your own is that you probably don't have bugs in the areas you care about.

MickLH

Lol idk what to say, I tried to hack my required features onto a few different engines and found that my use is too edge-casey and nothing I want is supported

Darth_Wardy

9:16 PM

well the engine i'm using atm is in beta, but the issue i currently have is that it doesn't have a networking layer and insists that my game use only .net portable assemblies

oh, shit, i just figured it out

I'll put IoC in the main project and inject a standard socket client in the windows build and worry about the others later

IoC FTW

MickLH

I actually finished the whole pile of refactor and bugfix crap... I've been working on that for so long, I have to sit down and think hard to even pick what direction to push for now.

Darth_Wardy

it'll break the engine rules a bit, and won't be cross platform, but if I want to implement the client on other platofrms I can just bolt in a different interface implementation for the network client later and have the IoC stuff return one of those instead

@MickLH well i finally got some core meshing code finished and networking sorted

also have a working world server

and i've got to the point where I am streaming my world data to the client pretty fast

MickLH

good stuff, I'm working on networking too atm

Darth_Wardy

scaleable ?

MickLH

yeah

Darth_Wardy

9:20 PM

you going for cloud / MMO type stuff or just basic?

I managed to write a simple socket wrapper that does message parsing, so I can stream the raw bytes and it raises events on messages being "completely received"

MickLH

I guess technically it's MMO style networking, I mean I have to solve all the MMO problems because of the edge cases, even though most of the time very few players are in the same area together

Darth_Wardy

I got some pretty good routing and its all async too so that's a reasonable start

need to figure out how to test those odd edge cases though

@MickLH sounds like a really large world like ED, or no mans sky

MickLH

Ironically the player should deal with mostly NPCs, and the NPCs present the greatest load on the system

But so far no roadblocks

I've been doing a "security first" implementation, so I'm still working out the handshake to be DDoS-proofed

The first pass defense is a proof-of-work system, where each client has to brute force search for a random integer that makes the packet checksum contain some specific bytes

Then if the checksum is missing the magic bytes, packet rejected. and of course if the checksum is wrong, packet rejected. This should let me control the rate that it's possible for connecting to start at, so if the server is overloaded I just require more proof-of-work in the handshake.

Ideally I'd like to figure out a way to let anonymous users join a game reliably even during an attempted DDoS

Darth_Wardy

@MickLH surely thats an infrastructural thing rather than a software thing (dealt with at the routing / OS level), or am i missing something?

MickLH

@Darth_Wardy My handshake process is pretty CPU heavy for the server, a client could crash the server just by starting handshakes as fast as possible. I can't rely on any external firewall because that might block legitimate users

Darth_Wardy

9:31 PM

hmm ok

I went for a hub and spoke model

if a spoke drops its no biggy, I can fire up more but the underlying hub will not be affected

MickLH

I went for a "master-manager-client-client" system, the idea is that it will be a hub and spoke system in practice

Darth_Wardy

detecting ddos attacks is not something I intend to ever do, my thinking was that I would ultimately just deploy to a cloud and have the cloud figure it out

might be a bit naive though

im probably wrong ... knowing my luck

MickLH

I don't know if it's necessary for most games

But when you connect to my server, both client and server participate in generating a session key which costs a big CPU hit

Before the client has even logged in, I mean

I got it down to milliseconds of processing, but that's pretty heavy still

That would imply less than 1000 connections per second required to crash the server with 100% CPU on network handshakes

It's not crazy to whip up a >1000 node botnet

Darth_Wardy

so are you passing that key in every message?

MickLH

Nah the key never gets attached, but both sides have to generate it at the beginning

Darth_Wardy

9:37 PM

@MickLH but thats for a single connection server though right ... "single spoke"?

in theory you'd have a load balancer sat over an array of them

MickLH

I could go with a load balancer, and move handshake & authentication to its own server cluster, I just wanted to avoid the complexity of external authentication

It's only generating the key the first time that's expensive, both sides have to generate the same key without ever sharing enough information for anyone else to know the key

Darth_Wardy

I've got a routing based system, each message comes in with a header, based on the header the routing engine makes a call

if the socket isn't authenticated it expects a login message

anything else results in a drop

the server will literally just say "fuck off" and drop the connection

I figured that was about as good as it gets

MickLH

I have a similar strategy in that I boot any clients that do anything weird and say "try again"

Darth_Wardy

the routing engine in the event of a login call might call a webservice or something, not sure yet exactly how I want it to work so I mocked a simple "list based implementation of the login handler"

MickLH

I went the polar opposite on the routing system though

Darth_Wardy

9:41 PM

just so i had something to handle the call

opposite?

how do you handle message routing?

MickLH

Yeah I ask client to route between themselves, and the servers only pass messages when the clients can't figure it out, or the game rules require the message to be processed by a neutral party

Playing with friends on LAN is a pretty important element of my game, so even in online play I take advantage of the faster P2P link instead of making a round trip

Darth_Wardy

hmmm, interesting

I did wonder about using p2p based udp packet routing and saving the server round trips but not sure, since all moves in my model come from the server there can't really be much cheating but it adds a lot of server load

MickLH

I have a system where Entities have an "owner" node, and once the server gives ownership to another node, it tries to forget about the object as much as possible

Darth_Wardy

haha, how strange, i did the same but the owner in my case is always a server node and never a client

MickLH

I let clients be owners in cases like someone walking through debris on the ground

As soon as you stir up the physics, the server instead of processing a physics tick just dumps all the object ownership on you

Your responsibility to compute physics for them, the server will accept your updates as long as they pass the most basic sanity check

Darth_Wardy

9:51 PM

Yeh i was wondering about that

weather to do physics sim on the servers or not

MickLH

More importantly, the server doesn't ask you for updates, but it does give you a key you can use to prove to other clients that you own the objects

Darth_Wardy

means needing gpu installs on servers though

@MickLH hmmm, there's an interesting thought, if i was to combine that with my model then I would only need to send the key once

I associate key auth info with the socket itself, once its there, there is no more need to send it again

standard handshake for me is "hello, im <auth data>, <message stream>"

essentially

MickLH

The per-message signatures work out gracefully as a more functional replacement for per-message checksum.

Take the debris example: The only client who cares about the debris owns it, the server can avoid any processing whatsoever, doesn't even need to load the models. If another client enters the area, they already know about the players in the area, and can request the debris info from the player who owns it. If the owner left the area then, they can pass ownership back to the nearest client. If no client is in range they can pass it back to the server, and send the server a state snapshot for it.

Without the signature, every client who saw the debris would need to ping the server to confirm the owner. In that case the server might as well own the pile since it has to lookup data for it and respond to packets constantly anyways.

Darth_Wardy

Yeh i deliberately didn't want that

since that would open a cheating opportunity in my game

i mean, does a each client own the rendered player object that the user is controlling in your case or does the server own all players?

having a client own the player means that if the client is compromised they do things like speed hack their own char

seen that in so many games before and it sucks

StackExchange

10:24 PM

0

Q: I have a specific question about [specific-question]

specific-question has no description and ironically isn't very specific. The 6 questions using the tag don't have any similarities. Also, if your question isn't specific, then it is off-topic or too broad. Does this make all the other questions on Meta Game Dev SE too broad? Any good post should...

discussion tags

MickLH

10:49 PM

@Darth_Wardy I went with a rule validation system

Darth_Wardy

I started trying to build a complex rule system then I eventually settled with basic event handling

I essentially only have a set number of operations that can take place, so I have a means to add a list of handlers for an operation to a concurrent dictionary

MickLH

I try to keep the rules simple. Debris has a low energy to destroy, so it can never reach too high a speed

Player also has a maximum speed, etc

Darth_Wardy

but i do have an interesting situation where the message in question needsto be passed between connection servers

I don't really want to be doing cross server "who's nearby" type queries so I rely on the server in question to know and pass on relevant information

so I added a means to say "I handle events in this area for this type of thing"

that then meant I could for example have multiple world nodes, and then have multiple connection servers hooked up to them

then have each world node tell each connection server that it was interested in all world events within an area

that allows me to break the the workload down quite heavily

but it gets really interesting when world nodes need to talk because somethings happening on a world node boundary, and there are players from various connection servers interested in that event

MickLH

I have similar problems at a different scale becasuse any empty space has to be assigned a cell before it can contain entities

Darth_Wardy

I feel like i did my bit for that question

@MickLH ah c i don't have that problem

MickLH

11:01 PM

It ends up working well for me because of other weird decisions in other places :P

Darth_Wardy

I can even define portions of my virtual world that don't connect nor ever have to connect in any physical way, I just have arbitrary lumps of 3d space

I'm using relative positioning

MickLH

I use relative positioning within a cell, but cells are absolutely positioned

Disconnected worlds can be implemented by putting an empty cell gap between them, and a box around them if you're paranoid

Darth_Wardy

i don't need the empty cell

you simply define what you can see, not what you can't see

its not perfect yet though, still working on the scaling of things

I still think I can speed up world loading

MickLH

I don't really have any use for disconnected worlds, so I compromised it for the reliable absolute positioning in the physics simulation

Darth_Wardy

streaming the chunks is getting faster with each iteration

it mostly makes sense in situations like in space

I can simply define a region of populated space and not have to worry about the non populated areas

in my case its floating islands so that's sort of like an asteroid field

MickLH

11:06 PM

I use a sparse allocator too

Darth_Wardy

but the asteroids could be 10's of km's across

so I could have a large land mass then nothing for 1km

MickLH

You don't need the whole grid, only the cells you use. Actually the asteroids seem like a perfect fit lol I'm using 1km^3 cells

I can scare you about my decision making: I made the physics engine represented by an entity :P

The simulation's origin is wherever the entity is, and all of its children become physics objects, wherever they are

Any time an object collides with something that's not physics animated, it pulls it into the simulation. When any islands fall asleep, they are removed from the simulation.

If one physics simulation is running too slow, it's supposed to split into two physics engine entities so they can run on separate processors, but my method for checking "too slow" is just seeing if the object count is too high, which is not really a good way to estimate physics load at all

Almo

11:23 PM

@ChrisMcFarland glad you're having fun in the clan

Darth_Wardy

@MickLH sounds evil

i've not gotten that far yet

MickLH

I have a friend working on 3D stuff, so I had to do a lot of the fun stuff first so he could play around with it. It was a very successful morale exercise.

Transcript for