Why not an FB2 with Homebrew games!

[fsuinnc]

as a long time classic game fan I have a question.

 

Why don't we have a flashback style machine that comes with homebrew games built in?

it's my understanding that the FB2 is basically identical to the original 2600 (or at least close). I'm sure patents, etc. on the 2600 have expired since it was prior to the NES and NES clones are available cheap. It seems like someone should be able to produce a small, functioning copy of the 2600. slap a chip with as many homebrew games as we can find (getting proper permissions of course), add an sd slot and, of course, a real cart slot and presto, an awesome 2600 clone.

 

Maybe the market can't support something like this as far as getting it into Wal Mart but I am pretty sure the fan base would have great interest. I know I would.

 

It seems like all the major engineering work has been done (FB2) to make this happen. Frankly, some of the fan based devices that have been created over the years like cuttle carts, and controller interfaces, etc. make it seem to me like all of this should not only be possible, but not that difficult using mostly existing parts. (By that I mean not that difficult for the guys that do this kind of stuff cause it is all far beyond me.)

 

the only stumbling block I see is financing a project like thils. If any of you engineer types with knowledge of such things could chime in and give a realistic cost to get something like this off the ground I would love to hear your thoughts.

 

I guess all things considered, it might as well be 7800 compatible too.

 

If any homebrew authors want to chime in about whether or not you would have interest in allowing your game(s) to be included in something like this.( I'm not suggesting you give up any rights and if a product were developed perhaps, like to be sold in the AA store, I'm sure some small royalty could be paid, etc.

[bennybingo]

as a long time classic game fan I have a question.

 

Why don't we have a flashback style machine that comes with homebrew games built in?

 

Because the FB's sell almost exclusively based on nostalgia. The vast majority of those who purchase this machine, do so in order to play games from yesteryear, without the hassle of purchasing old hardware.

[kskunk]

the fan based devices that have been created over the years like cuttle carts, and controller interfaces, etc. make it seem to me like all of this should not only be possible, but not that difficult using mostly existing parts.

 

the only stumbling block I see is financing a project like thils. If any of you engineer types with knowledge of such things could chime in and give a realistic cost to get something like this off the ground I would love to hear your thoughts.

Sure, I'm an engineer type, that released a fan-based device for an Atari platform.

 

Fan-based devices are usually financed on the engineer's own credit card. No outside investor would be interested in a homebrew project - there's simply no profit possible on devices like these, for unit sales under 10,000.

 

The Flashback 2 was a brilliant product. As benny pointed out, it had an enormous potential market: 30 million 2600s were sold, and at least that many people fondly remember one of the games in the FB2. The FB2 sold 860,000 units. That's worth financing!

 

Besides us, most people aren't that nostalgic about the hardware or interested in homebrew. They're after that one game they played all summer in '81.

 

Your project needs to face the cost structure of hardware manufacturing. The FB2 sold for $30, so surely a "Homebrew 2" could sell for $30, right? No way. The first unit costs thousands of dollars, the price slowly drops from there, and they don't start costing $10 until you've made 50,000 of them.

 

Devices like mine, or Cuttle Cart, can sell a few hundred copies. A really successful homebrew device might sell a thousand. At those volumes, you will be lucky to manufacture them for $100 a unit. With high costs and low volumes, there is no room for profit. Most homebrew devices are basically charity. Except for commercial ventures like FB2, I don't think anybody earns more than McDonald's wages, and many lose money.

 

- KS

[+Gemintronic]

I'm pretty sure the same guy that is working on the FPGA Arcade is also working on a new 2600 unit. Unlike the FB2 batari BASIC games should run just fine. I'd definitely volunteer my homebrew games as a pack in.

[fsuinnc]

let me clarify a little. My understanding would be that it is now possible to create a 2600 compatible machine with existing, "off the shelf" parts. Perhaps this is incorrect.

 

it seems to me that the development of the FB2 is already done. Perhaps Atari owns it and the work that Curt Vendel and the rest did to create it is proprietor, in which case perhaps new development needed. However, if the FB2 was assembled in some Chinese factory and they have the ability to crank out a couple hundred or a couple of thousand of them I think there is at the least a market among hobbyist. I think China is well known for quick set up and turn around of "small" orders of things like this. In fact, I am kind of surprised that A) there is not already a 2600 compatible available on alibaba or newfrog.com and B) that the NES/Sega Genesis compatible don't have 2600 compatibility built in. It seems that at the very least, it would be worth testing the market when you can produce these items very cheaply.

 

My employer (not electronics so much as "adult products") orders many items from China and our many suppliers are constantly undercutting one another to get the business. so again, it seems that if there has already been production runs done, there could be again at far less than the original cost, even for a much smaller run.

 

Lastly, I have a lot of homebrew games. I LOVE the homebrews for my A2600, A7800, Magnavox O2 and intellivision. (I try and support the developers whenever I can). But more than that, I often show these games to friends and guests. While they are often surprised they are always amazed. For the most part, the homebrew games that have been coming out over the past few years maintain the simple ideas and play-ability of the classic games but look and sound fantastic and push the hardware to its limits. In short, they appeal to the desire for nostalgia why offering excellent (and new) gameplay. I think it would be very easy to think that someone who liked the original A2600 Asteroids as a kid would love to get a hold of Space-Rocks.

 

and finally, I am not trying to start an argument. I agree that a mass market project of this kind is probably out of reach as well as a bad investment. I do think that there are a lot of classic game fans that would like a machine like this and, while I am not rich, I could probably help finance a small project if the right person or people were involved and thought it was feasible.

[+SpiceWare]

While I would love a FB2 with homebrew games, I don't think it would sell very well as the homebrew market is quite small and masses who bought the Flashback series did so for nostalgia, not to play new games (I did buy one, but that was to get new joysticks ).

 

Space Rocks wouldn't work in a Flashback system. It's using an advanced cartridge with a 70 MHz ARM processor which runs all the game logic.

[fsuinnc]

Oh well. I suppose a Raspberry Pi some USB controllers and good old Stella will have to do for a cheap system that will play the homebrews.

[Mr SQL]

Oh well. I suppose a Raspberry Pi some USB controllers and good old Stella will have to do for a cheap system that will play the homebrews.

 

fsuinnc,

 

I think you have an awesome idea about a Flashback that includes homebrew games!

 

I've suggested as much to Atari in this story, Working for Atari in Contemporary Culture:

 

http://atariage.com/...porary-culture/

 

Atari has gotten better with each successive Flashback (other than moving to emulation) but is still experimenting with doing things like putting different versions of some games on the latest Flashback instead of the Atari 2600 versions and breaking the feel of their paddle games; Homebrews are a great idea and more marketable IMO than either broken games or games from a different system because they have that 2600 feel and play like genuine Atari games.

 

Of course an Atari 2600 with the Harmony cart is probably the best solution for playing all of the homebrews, and the classics

[+Gemintronic]

Android dongles fast enough to run Stella are getting into the $40 retail range. The only major cost is finding a suitable wireless joystick.

[+stephena]

Android dongles fast enough to run Stella are getting into the $40 retail range. The only major cost is finding a suitable wireless joystick.

 

I think the hardware has been fast enough for some time, even the lower-end stuff. There's a Stella port for the GP2X, which is only 233MHz IIRC.

 

Also, Stella can be further optimized for an 'end-user' system. I've been communicating with a graduate research project that is using Stella in an environment where it needs to run as many frames per second as possible. Due to some rearrangement of code, removing stuff that's not needed, etc, they managed to improve performance about 40% more than the current public releases (some of these changes will be integrated into the main codebase soon). Of course, much of what would need to be removed is what makes Stella a developers emulator, but as a standalone system comparable to the original console, the code can be made quite a bit faster.

[+SpiceWare]

Of course, much of what would need to be removed is what makes Stella a developers emulator, but as a standalone system comparable to the original console, the code can be made quite a bit faster.

 

Hmm, maybe it would make sense to have a preprocessor flag to control the inclusion of the integrated debugger code.

[+stephena]

Hmm, maybe it would make sense to have a preprocessor flag to control the inclusion of the integrated debugger code.

 

It's already present, as are options to disable some other functionality. The following is partial output from 'configure' in Linux:

Optional Features:

--enable-gl enable/disable OpenGL rendering support [enabled]

--disable-gl

--enable-windowed enable/disable windowed rendering modes [enabled]

--disable-windowed

--enable-sound enable/disable sound support [enabled]

--disable-sound

--enable-debugger enable/disable all debugger options [enabled]

--disable-debugger

--enable-joystick enable/disable joystick support [enabled]

--disable-joystick

--enable-cheats enable/disable cheatcode support [enabled]

--disable-cheats

--enable-thumb enable/disable Thumb ARM support [enabled]

--disable-thumb

--enable-shared build shared binary [enabled]

--enable-static build static binary (if possible) [disabled]

--disable-static

--enable-profile build binary with profiling info [disabled]

--disable-profile

--force-builtin-libpng force use of built-in libpng library [auto]

--force-builtin-zlib force use of built-in zlib library [auto]

 

Of course, you'd have to edit the Win32 or OSX project files to disable these features, but the concept is the same. The current released builds include all functionality.

 

However, this was only part of what I was talking about. There's also the possibility to remove the built-in ROM properties database (would decrease the exe size by ~500KB), which should probably also be controlled by a pre-processor flag. Or to remove support for certain bankswitch types that you don't need, since some of them impose an extra penalty on other parts of the emulation (work is ongoing to fix this). Other work to be done is more agressive inlining of certain frequently used methods, potentially caching the results of peek/poke methods, etc.

 

There are probably still some optimizations to be made, but TBH, the code now runs extremely well on 8+ year old computers and OpenGL 1.1 or so. I'm confident that the current ARM devices out there are a viable option for Stella.

 

EDIT: I just compiled the current SVN code as mentioned above, with the following resulting EXE files sizes:

 

Full compile, all features included (as in normal releases): 3230168 bytes

Removed debugger and cheatcode support, and deleted internal ROM database: 1368248 bytes

 

Removing unwanted bankswitch schemes, doing some more tuning, etc could probably save another 100-200K or so. Runtime-wise, removing the debugger is the biggest gain, since there's a per-instruction penalty.

[sqoon]

I just can't imagine most homebrew authors are anxious to sign up for such a project either. Obviously, they make money from each game sold via other means, so why volunteer their IP's for basically a profitless venture? Not that there might be a couple out there who would allow it, but I am guessing the majority would not.