Sort:  

DPoS cares about stake not people. We have no way of knowing how many accounts someone controls and which of those accounts controlled stake used to attack the chain.

I agree with the sentiment, but I don't think we can evaluate it from that perspective

It was understood for a long time that the cost of attacking a DPoS chain is your tokens. It would help to know how many proxying users are aware of that before considering the airdrop

It'd also be nice to know if that would be feasible on Hive - to fork out the attacking stake without "moving"

Basically - what is the risk of it happening again, and how well can Hive deal with it if an attack comes again

I think we can evaluate it from that perspective for this one reason: sentiment determines almost everything, and money is just a belief system. One of the core reasons people on Steem justify blocking people's funds is a tit-for-tat mentality. Remove any wrongdoing, let all the people have the airdrop, then Steem's entire argument goes to shreds, and they have nothing to stand on. If they dump Hive, then so be it. We will buy it.

I don't think anyone should worry about dumping. Totally with you - would be stoked for some cheap hive :)

The concern for me is more about the attacking stake. Will that stake continue its attack?

I agree with the sentiment stuff, but there's not much point in being chill if it means the chain gets killed off

if someone has great motivation for attacking, they will simply buy hive and attack it. What is your definition of "attack" exactly?

A single actor controlling consensus is a successful attack on a DPoS blockchain. Essentially, any time we have a single actor controlling multiple top 20 witness nodes, we have an attack in progress. Once it reaches 11 nodes, the attack succeeds (I might be wrong about the number, it might be 13)
Pretty classic DPoS scenario, but we have to talk about it with strange vocabulary for reasons I don't quite understand :)
Basically, we can't determine who controls what accounts, and stake is all that matters in voting. The expected action from the attack is to fork out the tokens used in the attack. This was described way back in the early graphene days. Before protoshares was even live. Things are getting really muddled because of the emotional responses, twists in vocabulary, and the expected, messy cleanup that comes after the fork.
I think the weakness we're seeing now is that those who unwittingly participated in the attack and understand what they should do next time are in a very weak position to get their stake back. They don't start with any stake, and they've been vilified

*edit*
and yes, someone could buy enough hive to attack, but they should understand very clearly now that the money will be wasted. I suspect the rebranding on this fork was due to some complications with Steemit, Inc and exchanges. In the future, I expect we'll be able to "fork in place" and remove the attacking stake

Oh, and my other point: people are everything too. Dpos is just a system for people. Dpos is just a tool.

Think of it this way:

John has 8 accounts

2 of them have 5HP and voted to support the attacking actor
3 of them have 200HP and voted to support the attacking actor
1 has 200,000HP and voted to support the attacking actor
2 have 150,000HP and voted to try and remove the attacking actor

If we think about this situation in terms of stake, it's pretty clear: remove the attacking stake

How would we address John's actions as a person?