There is absolutely nothing at all wrong with a hard fork that is *not contentious*. There are trivial fixes to the quadratic hashing problem, most of which limit the size of “non-witness” (NW) transactions, or aggregate NW transactions in a block. There is a hardfork-bits proposal that allows users to signal readiness for a hardfork *before* signaling is allowed by miners. Core has been at work building out the infrastructure necessary for a hard fork.
If Bitcoin Core were to support a non-contentious hard fork, with full user and miner support… and with appropriate protections to ensure the minority chain is extraordinarily non-viable, then a hard fork can be done without any threat to the network.
Bitcoin Unlimited does have a solution for quadratic hashing, but I’m not 100% sure it works (qmax). It hasn’t been tested well, and there is some issue as to whether the number of inputs is counted correctly? But more importantly, it has no user-mediated upgrade in place.
Some people use BU to rage against hard forks in general. Hard forks can be managed…. they are just difficult, and several intermediate versions of Bitcoin Core will need to be released to ensure that, in the future, non-upgraded nodes can enforce future, successful, hard forks by failing to relay and validate blocks in the future chains.