For your insight, what I'm going to talk about down here, is the actual card that is going to be made. New hardware support can be made after the card is shipped. The difference between this solution and a PC, is that with this new stand alone card, you can make all the wiring by yourself. If you want something extra, you just have to add passive wiring to make things work (or basic components in some cases). If you want something extra, you just plug in a cable to one of the headers, and you can make all your wiring on another passive card inside your cabinet. This makes it possible to make all the connections in your cabinet before you attach the JAMMA+ board.
But as this will be a standard JAMMA+ board. You will get support for most things without any wiring at all. The JAMMA connector makes the board self sufficient. But the card also has VGA, and RCA. If you want to use this board without any cabinet. You have to use a converter (that you buy or build yourself) to add joysticks and power to the board.
I haven't quite wrapped my head around how all this works so my questions about controller assignment and top speed may be out of left field. If it's all hardware run with no emulation what roms won't be supported? Is there a year benchmark where the hardware to run the game is just too beefy? Like Super Model 3 type stuff? I don't know how to properly ask this question, it came from your comment about having enough RAM to support all NeoGeo games, sorry for the confusion.
Well NEO GEO has cartridges with ROM. For flexibility this new board (that needs a name) will load all ROM files in to RAM instead. So the machine will need at least 128Mb of RAM. There will be no emulation of hardware in software. All old hardware will be emulated in real copies of the hardware. As this is the case, there is no reason to use a benchmark. It's only a question if the card have enough gates to fit all a copy of a certain hardware. And concerning processor speed, I think it only will be possible to implement processors up to 100Mhz-140Mhz max. That will be up to anything up to about year 1997-1999. But there can be a problem to make certain mass of logic fit. It's about spending about $150 to make it able to run everything up to this around 1998. But there won't be any support for these fancy stuff for a few years anyway. Yes, someone has to make all the logic designs that copies the hardware of the original games.
There are a lot of nice to have features that are starting to get integrated into the hobby such as auto rotating 4way to 8way joysticks that cue off the ROM being played, LED lit buttons and features that display what controls to be used and provide light shows, auto rotating monitors that match the original games screen orientation; however, the original games didn't have these features. If cramping the board is an issue I'd say that these features are top candidates for the cutting room floor. I think that a cool phase two project would be a supplemental board that supported this kind of stuff, but these features rely on information about the ROM that is found in MAME and use software to call the hardware required (either stepper motor controller, servo controller, or LED controller board).
Well with the design I'm thinking of now, it will be possible to support almost any controllers. The issue is only to make support for all games. The issue is only to reserve I/O pins on the headers for certain things. Most of that work will be done after the hardware is released, as you can upgrade the hardware via USB. It will be the people that have these new boards, that has to come with reasonable suggestions about what is going to be implemented on what pins. Then we can make these new upgrades via USB.
Like I said, vector would be nice but isn't necessary, if it gets cut because it would be a compromise I say that's fine.
Well support of vector displays is no problem, it can be handled by a USB-upgrade.
[...]
Support for two monitors would be great but anything more than that and I lean back towards a link solution. Kills clutter, add functionality, satisfies the what-if-ers.
Having 2.1 audio should be enough.
As for controller inputs I think that four groupings would be fine. P1-P4. All can share a ground or have a ground for each set. Either way, tho most here will prefer a shared ground for all. As for what types of connectors to use I like pins like on the Xin-Mo encoders but others prefer solder points or screw terminals.
As standard there will be 4 audio channels, and 2 monitor connections out of the box. The JAMMA connector, the VGA connector, and a RCA stereo pair will take care of this.
Anything else can get support by dedicating I/O pins on the headers. To make almost anything fit, I will put three 40pin headers on the board, that will give about 100 I/O pins, for spinners, light guns, trackballs, extra monitors, sound channels, analog inputs, servo control, led controls, force feedback, multi cabinet communication, and what ever. And all can be handled without fancy extra hardware that you cant build by yourself. It's just a USB-upgrade away.
But as you understand someone has to write all logic descriptions. And that is what I am intending to do. I can right away say that there will be support for at least 20 games minimum by the end of this year. But I will do my best to make support for over 100 in the beginning of next year. And I'm heading for 200 games to be ready by next summer. But until then it will be a lot of testing. One problem as I see it will be compile time. I will need some immense amount of commuter power to make this happen. I don't know for sure, but I think the compile time for one new game will be like an hour or more (the work done by the computer to make the design fit into the new JAMMA card).
So I think I will need lots lots lots of help testing games, and get feed back on what is not working, to make this happen.