Sometimes it's easier to build your new stuff from scratch and figure out how to migrate all the old data to the new system than it is to gradually update the system piecemeal.

The technical debt - the engine, the net code, the table structure, etc. is enormous, and its rapidly becoming an insurmountable problem. I suspect this is what Yoshi P is referring to with a new ARR.

Blow the whole thing up, start over from scratch, and figure out how to port the old data after you rebuild it properly a third time.