Can you explain please?
My pleasure =p
Currently one army (person group whatever) needs >50% of all hp to attack the Hive network. Thet will vote top 17~30 witnesses. There is nothing that can be done.
To elect 4 to stop a hardfork, an army also needs > 50%. Alternatively they could stop a necessary hardfork until they 'negotiate'. This is also called extortion (proxy.token helped justin do it).
If we have 1t1v, >85% ensures 17 top 20 witnesses and a certain hard fork. Its safer.
However >20% ensures 4 top 20 witnesses and blocking a hardfork until the extortion is paid. Its now 2.5× easier to do an extortion attack and well within the means of potential alliances.
Sure doing nothing or extortion is less severe than a rogue hardfork, but they could also wait until they find a critical exploit that desperately needs patching before acting. We could experience this before and after every hardfork.
We could change the amount of consensus witnesses from 17/20 to 11~12/20 to address this issue. Or require 51~70/100. In anycase the entire equation needs to be rebalanced if the vote weights change to prevent a blockage attack.
I appreciate the substantive reply. I don't agree that blocking a HF is centralization of governance, however.
What I do agree on is that it appears that using stake alone to influence governance seems to be impossible to secure against both vectors for attack. There's far too many financial assets available to known censors for any mechanism based on financial assets alone to prevent censors from imposing greater stake on Hive.
I have become convinced that those that are advocating adding rep, or some other metric, to stake to effect governance are on the right track. I'd very much like to hear more discussion of means of preventing stake alone from being able to effect governance, because with that system in place, all that prevents those better funded censors from undertaking to destroy Hive as Steem has now been is the effort to do so.
Thanks!