There could be a solution to the time scale problem. It could be possible to have supplies,convoys and ground forces, as long as they were AI. A map interface to control them could be utilised and then it would play like a strategy game at that level.
So what's the fun in that you'll say. Well, instead of sailing the seas for a whole month or driving that tank for 2 days, have an option as to what unit to join in the ground/sea war. This way aircraft will still need to take off, fly to target, fight and land, but the ground forces will be AI with the ability to take control of one.
I know it's not very realistic in the sense of a purist sim fan, but then again we already have this in dogfight servers. People die all the time, respawn and the only thing that happens is that a player and plane "ticket" is deduced from their team's total count. I see no reason why it couldn't be made for ground units in the future, as long as time and resources permit.
I think it was hinted at that in BoB you would be able to join coops in progress and take the place of an AI plane. Well, that's exactly the same. You don't need all that time to get to the front. As soon as your little soldier dies, you choose to respawn as part of the same unit. Of course, units get depleted. But supply convoys and trains are moving to the front, so you can add a counter for reinforcements. For example, you want to spawn in a regiment/company/platoon that's suffered casualties and all the places are taken. You hover the mouse over the unit icon in the map (a map with filters to choose what units you want to see) and it tells you "next round of reinforcements: 25 soldiers in 10 minutes", in which case you can wait it out. Or it says, "100 soldiers in 3 hours" in which case you choose another unit because the one you picked has been effectively overrun. The campaign AI then retreats that unit if there are no human players left in it, or a player issues the retreat order from the map, or you could even retreat yourself as part of that unit with enemy armor and aircraft chasing you down (now that would be fun!).
The spawning system for aircraft would remain the same to encourage players to use ground units. If you could spawn in the plane of your choice 10 km from the battlefield as the infantryman can, then everybody will choose to use planes and nothing else. Also it would keep it realistic on the grand scale of things. For example, you might set up a CAP in the event that your guys need to fend off enemy CAS aircraft, or set up a flight of Jabos or bombed up P47s to loiter over your ground forces in case they need help. Something might happen and you get action, or it might not and you just fly back to base. Or maybe there's no friendly aircraft close and you need to race to the ground units and help them. Lot's of possibilities here that will make it interesting.
I think that combining a persistent battlefield that runs for a few weeks in a combined arms simulation will be doable in the next few years as far as technology is concerned. I used to play a subscription based online game set in space that had 10 times as much complexity. Of course it didn't take 2 weeks to travel from one end of the map to the other thanks to the space story setting (stargates, wormholes, etc) and it didn't need to model real world physics to the extent a flight sim does. That's why the only way to keep this realistic in a WWII setting without making it boring is to use AI ground units that the player can take over. They travel on their own, you take control of them when they are close to the action.
But then again, the bottleneck will lie mostly with our PCs trying to display a furball over a tank regiment that's attacking an infantry position, not the server so much. True, a server that needs to push positional data for so many objects will be taxed, but it's still a bunch of numbers (coordinates and vectors). The question is how feasible it will be for the client's PC to turn that positional data into a few hundred riflemen fighting on the ground with artillery, tank and air support without turning into a slideshow. And last but not least, the biggest limiting factor will be a developer's capability to pull this off against financial and time constraints.
But we can still dream, can't we?
