We will continue to extend client / server functionality. It's very important. We will not radically change the architecture. Splitting the server from the program, for example.
We also won't change the database we use (ours) in the forseeable future. I think it's world class in performance.
That's fine with me as long as we're eventually able to get some of the features we need added to the Client/Server system.
The main reason I suggested possibly splitting the server from the client was for "financial" reasons. That's a business-side decision, and you certainly have a (much) better handle on that stuff than any of us do.
Personally, I'm not really "burning" for any of the really fancy stuff (user accounts, multiple libraries, and whatnot). I'd just like to see the "equivalence of features between a client copy of MC on a LAN and using the server itself". That should be top client/server-scope priority to me.
PS. I agree on the database system. If you've ever actually used a big DAM system based on an "open relational" database (like Oracle or SQL), you know why he says that.