I was going to bring this up on its own forum, but I decided to post to this one. Turning old games online requires one of 3 things. The most obvious answer is writing online code for that game. For systems not meant to be online, it means taking your legitimate 3DO games to a Pc or Mac, and augmenting webcode and loading it to a hard drive. 2 problems with that. 1 is it doesn't work with original hardware and controllers without even more tinkering, and 2, each game has to have individual netcode programmed for it.
The second is to play it on a generic web-enabled emulator. You don't necessarilly have to program individual games to work for it. But the controls on games with no specific webcode will feel a little skatey, or slow, or force you to act way more than react, meaning you follow the same strategy come hell or high water, and you have to anticipate your opponents adjustments, do your adjustments, and it's basically a giant, "do you own thing and hope opponents fall in line" type play. A guessing game.
Before I talk about solution number 3. I can explain why you need netcode, or get a skatey, guessy, or jittery game when playing something that isn't turn based online.
If you do a good ping test between Cleveland and Chicago, typical ping times are 40-80 ms. Cleveland and Chicago are 600 km apart from each other. Light travels at 300,00 km/s. If data traveled at light speed in a direct path, the 2-way ping time should be 4 ms. Data ravels close to the speed of light. The problem is that data doesn't not travel in a straight path. It goes back to the days of the nuclear football, where launch codes can go from Washington to Los Angeles if Kansas City gets nuked. The advantage is a roundabout network where no signal will die. The cost was anywhere from 50-200 ms of ping throughout a trip. Human reaction time is 100-200 if there are no accurate predictors, (like a pattern, knowing what to look for etc.) and a phone that can beat 200 ms is good for that kind of communication.
Then came video games, where actions and reactions have to be sent in at first 1/20 of a second and more recently 1/60 of a second. When video games first went online, there were many sync problems, because unless you're next door neighbors, the ping time of the network, 50-200 ms. is greater than the ping time of internal video game. 1 second / 20 frames = 50 ms / frame and 1 second / 60 frames = is 16 ms / frame. So therefore, due to poor ping times you cannot do a live joystick stream and have it get in on time to process before the next frame.
High bandwidths were used to compensate for poor ping times. Those use a server as a ref-bot and constantly process megs of data per second. Microsoft bragged that Xbox Live sends every piece of shrapnel individually. If a game can be thought of as an interactive record player, then the output is a function of the inputs and the random number generator. meaning if you can capture and replay joystick data and fix the random seed, then the exact same frame-for-frame game could be replayed, just by storing joystick maneuvers and random numbers (if used) played over real time.
So basically, I theorize that if you can a) send and receive data in half a frame's time-span (the other half frame would be for a confirmation signal) b) sync the random number data (should always be done, even if it isn't used), and c) have an independent OS which enslaves a unit like a 3DO for game playing, graphics and logic purposes, which can pause the game in case of a network brownout. and make sure the machines are synched up, then you can turn ANY game into an online game, no matter how great or obscure the game and/or system are, without writing individual net code, and should feel just like a local game.
Luckily, I know a connection which can connect 2 people close to the speed of light in as direct of a path that cell towers allow. IT's Sprint Direct Connect and Verizon's push-to-talk. Sprint told me that the "one person talks, everyone else listens" nature of direct connect is a function of the phone logic, NOT THE NETWORK. All you need is 33 bits/ms to do most games, but in theory, you can simultaneously listen to 12 simultaneous combined inbound and outbound streams (12 players is the Saturn limit. I don't know what the 3DO limit is. I heard it's 8.) so that any number of players can play on one machine and be connected to as many other machines to play an up-to-12-player game. and all you need is 600 kb/s if you limit streams to 33 bits/ms. (Most systems up to the PS2 will fit under 33 bits/ms/player). Plus if you live in a 4G area, you can do multiple Wiimotes, Kinect players, and Move controllers, and once the PS4, Xbox One and Wii U is done, those games can be retrofitted to work with this network, even when the game-specific network servers are cancelled.
The reason why cellular can do it is because Direct Connect/push-to-talk is all one network, either Sprint or Verizon, and doesn't have to negotiate between networks. Wired networks have partnerships with certain other networks, hence why a Cleveland to Chicago trip can go to Miami and Phoenix before it gets to its destination, and result in ridiculously high ping times. And I know a cellular ping tests rates 100-200 ms. The reason why that is is it does from tower to tower to onramp to the rest of the internet, then negotiates the same pingy internet, and THEN goes to the off-ramp to the cellular network then from tower to tower to your opponent. The advantage is you can contact anyone. The disadvantage is you get pingy connections. If every player is on the same network and the signal travels directly at light speed. then you can play ANY game with this procedure.
I got Anthony Gaccione of Sega Of America saying if I can show it working with an original Genesis cartridge, they'll license it. I got similar deals with Atari, Intellivision, and Colecovision. Nintendo said they don't comment on licensability until they see the actual prototype. Sony and Microsoft are concentrating on their new games. And Panasonic who either a) owns the 3DO after the company folded, or b) can get the rights from Trip Hawkens if they see a demand, said they'll review the submission by January 2015 or earlier. They just don't accept submissions until that date or sooner.
I don't know how much it would cost to produce a 100% accurate retro unit, but I was trying to sell 1000 units and charge $125 for a retro Genesis, but most people said that was too expensive. But I don;t want to undercharge and be forced to fund the rest if my guess is low, so my philosophy was charge high and refund any excess after a percentage for profits. It'd be better to be funded by either a company who has nothing to lose and lots to gain with retrogaming, or a retrogaming enthusiast company/hobbyist, fund $17,000 worth of research from Davison, or research it for less, and get a more concrete number for cost for a retro unit modified to work with the network.