Any fork with a low hashrate would be subject to what Luke-jr did with to Coiledcoin.
Hi Richy, what if the lesser fork adopts a new proof of work algorithm that obsoletes the existing mining hardware?
I believe when Luke-jr did what he did, we were not on ASICs yet. Starting from scratch is rolling the dice. There are a lot of powered down (or being used on altcoins) GPUs out there. Who controls them?
ASICs will be produced for the new POW in a couple of weeks to months tops in any case. Or would be if it wasn't going to be effectively dead.
From my memory of the asic race it would take 6 months at least, probably closer to a year, but that's not important. What I'm considering is that it would be
necessary to change PoW in the case of a hard fork simply to avoid the risk of attack from the existing sha256 mining hardware..
On the less popular chain, blocks would grind to a halt under the existing difficulty. If we assume 10 percent remain then you would expect 1 block every 100 minutes, at least until the next difficulty adjustment, which would be a very long time because difficulty is adjusted every 2016 blocks. If you reduce the difficulty artificially to allow blocks to be mined and transactions to happen in a timely fashion then It becomes trivial for any mining pool to quickly switch over and mine a bunch of your blocks and then dump the coins...
I think, in the situation where one would want to cling to a minority chain after a hard fork, an algorithm change is
essential.
Also, there is no incentive to design or build asics to run the new algorithm unless it is insanely profitable, as was the case when the first bitcoin asics came on the scene.