Creating a New Game Engine

C62877b8c2b32507a25bb9ed51c0c099
0
dilenshah23 101 Nov 03, 2010 at 08:52

Guys what is the cost and whole process to create a game engine form scratch for development of a game like Assassin Creed and Prince of Persia and this engine should be compatible with PS3, Xbox 360 and ofcourse PC.

Please need all details thnk u in advance, its urgent.

Dilen.

9 Replies

Please log in or register to post a reply.

8596d335bd0a1964a74eff2dc1ee5b14
0
necroside 101 Nov 03, 2010 at 15:44

I guess few hundreds million, and at least 3 to 4 years with a huge team, my guess is at least 50 people, also you need game developers to test the engine and develop games and demos as it evolves, then when you’re done, start working on keeping it up to date and expend millions o bucks on maintenance and development.

I forgot, you and the entire team will need a lot of food and money to do this.

So here Comes my tip

Why don’t you license a engine o pick a free engine and modify it to your needs, you can learn how to create a game engine from scratch but this will take you a lot of time until you lower tour expectations.

BTW Google is your best friend don’t forget it.

Forgive my bad English.

B20d81438814b6ba7da7ff8eb502d039
0
Vilem_Otte 117 Nov 03, 2010 at 16:20

Okay, most people will point you to Scientific Ninja (especially on gamedev net), I don’t like the article and I don’t have same meaning on thing (if you wish to read it, then here - http://scientificninja.com/blog/write-games-not-engines )

I’ve been working on game engines for quite some time (although I’m more doing graphics cores, but well), so here are few advices (not all of them):
Do it in small steps - creating huge engine from scratch is impossible even in large companies - we made ours in 8 major steps
Don’t accept more people to cooperate on basics, do them on your own - you will learn a lot and also it will be easier to extend; you can work then in team or separately.
Dont focus on single thing (then you’ll most likely make just rendering or physics engine), do all
Make demos everytime you make something new (not just for eye candy, but also when somebody will want to work on your engine later it will be good tutorial)
And don’t forget to comment, comment and comment your code.

Of course I’d recommend you also to sit and write on paper what you want in engine and your specs (of course not write something specs can’t do) - like good reflections and OpenGL or Direct3D just not work; decide your technology, target hardware and such. We didn’t and we sometimes hit wall :( (although as for our project, some things must not be decided to progress - like target hardware).

8676d29610e6c98d6dd2d9c38528cd9c
0
alphadog 101 Nov 03, 2010 at 17:06

@dilenshah23

Guys what is the cost and whole process to create a game engine form scratch for development of a game like Assassin Creed and Prince of Persia and this engine should be compatible with PS3, Xbox 360 and ofcourse PC.

The whole thing from the ground up? Including a viable networking subsystem, a fully debugged physics engine, a solid renderer, a good audio subsystem, etc., all while making it cross-platform for PC and consoles?

About $1.50 and a couple of days, I’d think. :whistle:

Well, Unity is a good example of what you are looking to qualify: Unity has about 50 people now, and started in 2001, released v1.0 in 2005, v2.0 in 2007 and v3.0 this year. If you take a blended average TCO of $65000 in payroll (devs expensive, secretaries less so, so we assume an average closer to devs since they’ll make up the majority of positions, and add taxes, benefits, etc.), and assumed that payroll makes up half your expense mix at that size, then it’s currently $4-7mil/yr. Last year, they got $5.5mil in series A funding, which is usually used to carry a company for 6-12 months, so that matches up with my off-the-cuff calcs.

Ceee4d1295c32a0c1c08a9eae8c9459d
0
v71 105 Nov 04, 2010 at 15:53

I ‘ve been working on my egine on and off for almost 10 years, previosuly i started with software rendering and texturing done entirely on the cpu this was before the raise of ‘accellerator’ cards. So i think i can give you some type of advices .My advice is to stick with data structures, build robust containers for object, maps, particles and whatever, try to separate the proper rendering code in a way you can update or rewrite everything related to graphics pretty quickly, without turning the entire engine structure from the upsidedown when you will need to add something ( and you will )
Start with engine data strucutre , then graphics, leave alone the old opengl pipeline and start with shaders , when you will have a good and stable rendering engine, add in small steps physics , networking and music.
Don’t throw too much things in your ‘idea bucket’ , or the task might appear daunting to you and you could quickly become demotivated.
Write down a journal with your major goals, eventually as you start coding the goals will explode in many subgoals, try to resolve them once at time and when you are done, erase the solved goal.
Example
Main goal: write a mesh class
Sub goal: write vertex container
Sub-Sub goal: decide whenever use stl or write your own container classes
Sub-Sub-Sub-goal : write containers for vertex , texture, coordinates

.. and so on , when you are done with every sub-sub-sub goal, erase it
from your journal this is a psychological trick which helped me quite a lot.

You always have the option to use a ready made engine, but where is the fun ??
P.S. am i off -topic ???!??!

8676d29610e6c98d6dd2d9c38528cd9c
0
alphadog 101 Nov 04, 2010 at 16:37

@v71

You always have the option to use a ready made engine, but where is the fun ??

The fun is in making the game. Or, given time is a precious resource, making more games vs. one engine and less games.

And, before anyone beats me up, I’m not making a value judgment on either approach. Depends on each person’s desires and what they want out of life.

Eb3cfff1925eec543d8b53b9b53a32c7
0
Goldbott 101 Nov 10, 2010 at 02:55

Well everyones just about covered the subject quite well. If you would like to make games it’s best to learn a part of game development (3d modeling/animation, programming, ect.) to a decent point then join a team to learn more. At least thats my opinion.

Fd80f81596aa1cf809ceb1c2077e190b
0
rouncer 103 Nov 10, 2010 at 05:52

wolfenstien doungeon, stickmen for characters $2.50.

Actually if you payed someone to put that together in a day, charging $20 an hour, itll probably come to 20*10 hours… $200 even for something that simple. and extra for bug fixing, if it didnt come out perfect first time.

But of course, if you could code it yourself…

8676d29610e6c98d6dd2d9c38528cd9c
0
alphadog 101 Nov 10, 2010 at 17:05

@rouncer

wolfenstien doungeon, stickmen for characters $2.50

Is that making a game just so one can say “I made a game”, or do most people want to make a game other will actually be interested in playing?

Fd80f81596aa1cf809ceb1c2077e190b
0
rouncer 103 Nov 11, 2010 at 05:33

If it was your first project, I probably wouldnt recommend doing any more than that anyway, especially if it was 3d, too much to learn. If you were a bone n00b, then you wouldnt even get the device initialized on the first day, I can remember back to when I was first learning.

It still could be interesting. :) But yes, perhaps not the greatest engine on the planet.