Damn there are tons of posts. Was skimming over some earlier stuff about Unity / MMO etc.
Unity wouldn't be used for the server code, or it sure as hell shouldn't be. Reading the Unity3d site's scripting information, I see that it is
essentially built on top of Mono, an open source .NET platform. Since you can script with C#, Javascript or Boo, it'd be easy to design a half-decent client.
Dedicated server could be written in C++ or C# if you were looking at running it on Windows Server 2008 or 2012, or go with a straight C++ or Mono build and run it on Linux. The database performance becomes an issue when you start growing the total number of concurrent users. Eve online uses 64bit MS SQL Enterprise, iirc. A smaller scale MMO / Persistent game could get away with MySQL - but it lacks the advanced features of MS SQL and Oracle. I've heard some good things about NoSQL. You'd need a master server and a few helper applications, but the workhorse would be your zone servers (if the game was zoned).
Since the server would talk to the Unity client, I'd go with a windows dedicated server coded in C# with the networking as a separate library. There are many garbage collector issues with C# / Java but they all have work arounds and I've found C# to be a much more productive language for development than C++. The game I'm writing now uses C#, OpenTK (OpenGL for .NET) and
Lidgren Gen 3for the networking. Its stupid easy to get going but I've not seen any large scale tests. Almost all of my netcode is in a separate project so I can tweak it without having to recompile the dedicated server and client.
But enough about my shitty game.
It all comes down to getting a working prototype to test game mechanics and iterate, iterate and then more iteration. Get the gameplay, physics and netcode tight. Stress test the database performance. This can all be done while the art guys do their thing and slowly replace the generic models and textures in the prototype. Run the builds early on Windows and Linux and Mac boxes of all sizes and shapes and go from there.
Building a game is hard, and an MMO exponentially so. But if you have a dedicated team and understand what staying in scope means, it is doable with an indie team. I hope Brad and co use the existing assets for a solid prototype.
I remain skeptical but hope this or a game like it gets built.