Which system is the easiest (and hardest) to dev for?

If tgere was support for the 3d hardware in PS3 linux, I would say that starting there would be the best option. But as there is not, XNA is the easiest way to get started if you have done any PC directx development before. That or get a modded (or debug/devkit) PS2 and enter the PS2 dev scene.

 

360 is very simple which is a good thing for indie developers. Wii is technically just a Gamecube with a higher storage medium, online and a more complex controller. That explains why they are able to push the Wii way beyond the Gamecube's projected capabilities. PS3 from what I hear is a bitch to develop for. I have no experience so I can't confirm that. the GBa is one amazing piece of technology. Just check out the level of 3D capabilities this thing has.

http://www.youtube.com/watch?v=wVdOFjJ99ic&feature=related
http://www.youtube.com/watch?v=zr_DkfO_Csg&feature=related

 

If you want to take the *really* easy route, someone made a game maker type program for the NDS... obviously its much less powerful than using a dev kit but it could help get you into console programming: http://www.dsgamemaker.com/

If you want to take the "real" programming route, there is hardly a better place to start than with the gba. If you know C then I recommend HAM lib for getting started, as it has a huge tutorial and entire book dedicated to learning it. If you don't know C, then theres a version of basic (Dragon Basic) that was ported to the GBA... its dead now but you can still download the IDE and tutorials.

 

Um, no. Go read up on what hardware the Wii has inside it. It's similar but it's not "just a Gamecube".

 

I dont know how to code many things, But WIth the SDK and My XeDk I was able to make alot of cool small things very easily with all the samples that were provided.

So in my opinion thats the easiest system. I really dont have much experience with console programming tho.

 

gba is very easy, It took like 2 minuts to write and compile a hello world program

 

A system is easy when its easy when you write things in ASM, or in some high level language without any libraries whatsoever.

I'm a big fan of MD, the system basically programs itself, and is nicely designed. You can dedicate Z80 there as it runs totally independently and can access most of the outer world by itself, unlike SNES where you need tremendous amount of main CPU power...
Adding a 32X in action makes your life a bit more difficult due to the freakin address shift and that you cannot see all the ROM at once on 68K side. MCD is just prone to crashing, or so I've heard...

Next stop is Saturn for me, the sheer number of chips on its mobo makes me wanna mess with it, and I'm quite sure I will like it.

 

N64 is a royal pain in the butt when it comes to textures and audio. Otherwise it's great.

Say your game requires a textured sky sphere. Well, on the dreamcast and any later system, you just take one big 256x256 texture and map it to a sphere. Easy.

Not so on the N64. It has a whole 4096 BYTES of texture memory. This means that you can draw at most a 32x32 32-bit RGBA image, a 32x64 16-bit 5551 texture. There's also two paletted modes. Want to use a palette? Forget about that upper 2KB of memory. So you can only store a 32x64 image in paletted mode either.

You want to do mipmapping? Cut the memory in half again. 2kb for your texture and that's it.

It's so bad that I had to write a tool specifically just to create a skysphere display list, each rectangle of the sphere has its own texture.

This is one of the worst reasons why the N64 had graphics issues. That, and the fillrate and zbuffer accesses are slow as heck because of RDRAM latency. If you want it to go faster, you have to put your framebuffers and zbuffer in different 1MB RDRAM banks.

But, I love it anyway :dance:

 

I've decided to start off with XNA and an Xbox 360. It seems like a good way to learn C#. I'd also like to have a play with Dreamcast coding too.

 

Programming for the Xbox 360 using the XeDK Developers kit seems easier than the Xbox 1 SDK.

XNA is too easy. Then again it doesent give you access to the 360's full resources.

 

I geuss sorta off topic but I have a Xedk And use the sdk with it, but ive never used XNA. What is limited with XNA? like what arent you allowed to do with xna?

 

I don't think getting the 'hello world' output is a good guide to complexity of programming. The older consoles are simpler because they DO less. PSX was easier than PS2 and from what I have seen, the PS3 is even more of a pain.
If you use a shell, stick to C, don't use the system 100% (just %90) then maybe the latest machines are simpler than they can be for 'to-the-metal' assembler coders.
I would still argue that the Gameboy is the simplest... it does so very little...

 

In my humble opinion the Gameboy Advance and the first XBox are equally the easiest consoles "bang for buck", you can very quickly get something decent done with both, using some libraries and simple code.

In the hardest corner I would put the n64 and the PS3; while you can do something quickly with both, doing things properly is much more complicated and time consuming, having said that they are also amongst the platforms I like the most...

Jollyroger

 

I don't like the Super Nintendo. It feels like Nintendo misunderstood the idea of a 16-bit video game architecture. If they didn't hire the realism-whores at Sony to design their sound chip, and if they made a simpler hardware register design for their graphics chip, and focused more on raw power instead of complexity, it could have been a better and easier system to program.

 

I take it you've not coded for the NES, then? :icon_bigg Sound likes you want a Genesis.

The PC-Engine (or Hudson Entertainment System as it's also known) had a chance at becoming the next Nintendo console. Hudson was Nintendo's first 3rd party developer and was very close to Nintendo. They tried to license their new system to Nintendo, but the deal fell through, and NEC came along and pick it up. The SFC was in developement for a long time and you can see what looks like to be a last minute addon's to the sPPU's.

Still, the CPU being crippled by the WRAM delay cycles and the slow clock speed in general, the solution was easier to fix than say on the Genesis or PCE. However convoluted the sPPU layout was, the final output was amazing back then and no 'cart' addon of the other systems at the time was going to bring it up to the same level. At least graphics wise. PCE fans were enjoying red book audio before and after the release of the SFC.

Also, I don't see how is the SPC700 is adding anything overly complex to the system.

 

I didn't mean raw power entirely as in cpu power, but if instead of making 8 graphic modes, they made only one mode, they could make that one mode more powerful than the 8 modes, and have it do cool things like real scaling and rotation on sprites.

For the spc700, wouldn't it have been a lot less expensive if they just had the 65816 have access to the dsp instead of having a secondary cpu running it? They could've used that extra money to make the cpu a little bit faster and I don't really expect there to be a big difference in games if it had less ram.

EDIT: I personally don't beleive slowdown was entirely the cpu's fault, I beleive it had to do with bad programming too. It's that if it were clocked faster, bad programmers wouldn't be able to screw up games as easily.

 

DS ... from not knowing the system at all to finished game in less then 2 weeks parttime - probably the easiest system I ever coded for

 

Well, they could only do so much with so many cycles in a scanline and hblank, etc. Removing the extra modes isn't going to make the chip faster or free any spare cycles. That's why they're modes and not just features you can instantly turn on and aoff. All they really did was optimize what chip real state they for more modes. Besides, quite a few of those modes are useful.

I dunno. The unit wasn't designed specifically for Nintendo. From what I've read, Sony used variants in other setups - including its own Playstation. The Playstation one is a slightly cut down version of the one in the SNES, but with more channels. Anyway, if you removed the CPU then you'd have to interface the memory into the '816 bus which complicates things (sPPU stalls the sCPU's bus every so many scanlines for memory refresh cycle - that would be an audible artifact). Having a separate dedicated CPU is also better for other audio related reasons.

True. The '816 is no slouch when the code is optimized, but the slow clock speed is still somewhat of a hinderance. Though there were some examples of great coding in games.

 

I meant in a transistors count vs speed trade off. BTW, what physically makes chips faster and what physically makes faster chips cost more?


For the programmers and the slowness. I'm very surprised at how stupid people can be when programming things. I hate how programmers think optimizing code is so freaking hard. Optimizing is so freaking easy. All you have to do is program the game the using the easiest way, that's it. I hate how programmers slowly come up with stupid tricks that either barely work or makes the problem even worse and tell other programmers about the stupid trick they made up and everybody takes it and believe it's the best thing sliced bread and they're geniouses just becuase they can cram more stupid "advance" programming "trick" into their tiny little minds. While in all of my codes, every single opcode I write is an individual trick, only because I only use 1 opcode when it can do something in only 1 opcode. Other people ask others "Do you know a complicated trick of doing this opcode" and they give them a complicated trick of doing an opcode, then they think of how ingenious they are for using complex programming in the place of a simple opcode and cram it into their brain, while I know if I just use one basic opcode instead of a complicated "trick" to doing that opcode it will run a lot faster and still take up less memory. There is more to intellegance just knowing a lot of stuff. There is something called common sense that people need to be smart. Just knowing a bunch of crap doesn't make people smart, that makes them brainwashed and nothing more.

 

when you're done with doing something the simplest way possible, some people like me will want to get more out of what they've done, so real optimization begins...