It is good to be prepared for any eventuality. Just because we talk about a PoW hard fork doesn't mean we'll use it on a whim.
A good thing would be to have some script somewhere tracking blocks and transactions, so that if any big censorship re-orgs happen then it can be proved to the public.
Yes, we should only use this after an attack occurs, but need to test and prepare now. If anything showing we are prepared is likely to prevent an attack in the first place.
https://twitter.com/petertoddbtc/status/843450186416365569Instead of PoW change could it be possible to add a supplementary PoS layer which requires a block to be validated in some way by nodes prior to miners being able to build the next block, reducing the centralizing effect of propagation times?
A hybrid PoW / PoS version should be investigated as well, my only concern is that we should keep the PoW HF as simple as possible and adding a "secure PoS" algo into the mix will be sure to complicate the HF as no
secure PoS really exist and politically adding PoS to bitcoin will be much more difficult.
My advice it to first test and get running a testnet version of simple PoW Keccak with 0.14 and than investigate a Hybrid SHA256/Keccak proposal ... and after these are running on a testnet with individuals actively mining we can start to explore a PoW /PoS hybrid.