There should be more than one implementation of the bitcoin protocol.
Also, there might be legitimate use cases for "enterprise" style full nodes etc.
I could think of a thousand reasons why XT could be useful, but the most important ones will likely be the ones I can't imagine right now.
Exactly. What is wrong with the goal of decentralizing development across multiple competing implementations?
That's not a bad idea. I don't like the control the two competing teams of devs are exerting over what path Bitcoin takes. The more choices of implementations there are, the more choice he miners have over the path Bitcoin takes. If the miners chose to go with a new set of devs it could end this mess we are in.