How long does it take to make a game?
i'm answering my own question here. For people new to game development: longer than you think!
i was just cruising around a forum where some folks were daydreaming about what they'd create with XNA GSE, and giggled a little when i read "Grand Theft Auto" and "FPS":
http://forums.gametrailers.com/showthread.php?t=32477
The Independant Game Festival has a great site if you want to check out the cream of the indie crop. It's especially valuable because it records how long each game took to complete:
http://www.igf.com/02finalists.html
You might check out some screenshots and think "three years? But that game looks like ***!" Yes, my friends. Yes. :)
Start small. Build complexity from simplicity, but know when enough is enough. It's better to have one small finished game under your belt than a hundred unplayable and over-ambitious folders full of code.
- MrSock
[1176 byte] By [
MrSock] at [2007-12-23]
Considering that it takes teams of dozens several years to create most games, indie teams normally would take several times that long, given that most are only a handful of people and don't work on them full time. This is probably why most indie projects fail - people lose interest when progress isn't seen all the time and there's no incentive (like getting paid) to keep them going.
My advice to indie developers is to start small and not get discouraged. And you don't have to make a complete game to be a good XNA Game Studio Express developer. You can make really good GameComponents that are useful to other developers, instead of a full-blown game -- am I right?
Yes, start small. Though GTA etc is generally out of the reach of small indie developers you can still write fun casual type games.
The Hexic demo I did for the XNA team at GDC (all managed code) was less than 3 weeks of coding and less art time (I'm not sure of the exact art hours). The game was not 100% a complete experience we did the intro screen and one of the play modes, but there were no level ups or game variations and several liberties were taken with the actual gameplay for simplicity. But if you watch the video you will see the parts converted are very similar to the actual Hexic HD game on the 360 http://www.thezbuffer.com/articles/376.aspx
As they mentioned in the GameFest key note video SpaceWars was 3 weeks of 3 people. Since I did the actual coding I can confirm the story. There was about 3 man weeks of art, slightly over 3 man weeks of code (though that includes updating the code base to a changing API and learning the new APIs with little documentation), plus some time with a game designer and a sound guy (I think maybe a week or so of sound?). Its a 100% complete game with intro screens, levels, scoring, weapon upgrades and the retro vector mode. From a dev point of view the only thing left to do is the perf, tuning and polishing but what exists (and will be shipped with the beta - see Shawn's reply) is a fully playable game done in 3 weeks. Its a long way from grand theft auto or half life but it shows what can be done.
Why not start by trying to make Pong using the X-Box 360 wired controller. XInput is easy to work with and Pong has physics, sound, and graphics.
Glen
http://www.zenfar.com
Thats a wrong assumption... you can in no way target the casual game market with XNA. You should check out http://forums.indiegamer.com/ - there are lots of discussion about the hardware platforms (especially graphics cards). The indy game market is mostly DX7/8 style coding for legacy non-shader hardware. If you have a lucky day, you target intel on-board stuff, and if you have a really good day it even has four texture stages. A big problem with XNA is the removal of the fixed function pipeline... so you don't got a chance to do any fallback path for older hardware... always remember, there are more intel-on-board accelerators out there, than ATi + nVidia + 3DLabs has sold together.
The next thing is deployment. So you point your customer to the .NET 2.0 install, then you need DX9, and then at the last point you install your game and the XNA libraries. I expect 50MB Pong clones that need shader model 2.0... Micorsoft, you can do better.
My interest in XNA has gone to zero between the past few reads here... no fixed function pipeline... no access to the native interfaces... no nice integration into VS2005. Will stay with my DX8 engine for any productive work... works fine everywhere and has fallbacks for fixed function. Tools in C# via native interop, voila.
XNA will be a nice toy on my harddisc to play around with... nothing more, nothing less.
