There has been a lot of talk over the years about witness voting and how some accounts have too much control over who is in the witness ranks due to their very high stakes. Some might not know what witnesses do but if you were here for HF20 two weeks ago and still haven't found out, that is on you.
Now, there has been talk of introducing flags for witnesses to reduce their power but that is going to lead to more problems and unnecessary complication. It will also mean that the supporters of the top 20 can then flag the following 10 and make a large airgap to buffer themselves.
There has been talk about degrading the votes which means that over a period of time a voter must vote again to renew their voting choices. This will be a good thing to remove dead accounts and this should also be applied to disabled witnesses. If a witness has been disabled for 6 months, should it still be able to hold a top 50 position?
This post was from 
In my view, there is already too much complication and misunderstanding with Steem but a simple approach to having a more community orientated witness selection is to reduce the number of votes. Five months ago I suggested 10.
I would now suggest 7.
The reason that 7 would be a suitable number is that it gives a 1/3 control of witnesses but not enough to have a 50% majority, and is a long way from a super majority. The concern voiced by 
Having less witness votes increases the risk of people just going for the most popular witnesses. Meaning: you,
are obviously someone who brings a lot of value and stability to Steem. So my vote for you is pretty much clear.
And the fact that I have 30 votes, means that I can also vote for some more experimental witnesses, who might not have proven themselves so much, but still need support.
source
But regardless of whether that's true or not, I think the important thing is to increase the security level of the system against an attacker who suddenly obtained a large stake (e.g. hacks an exchange that has a lot of steem). As far as I can tell, limiting it to 1/3 of the top 20 would dramatically increase the difficulty of such an attack (for example, all things being equal, the attacker would probably need twice as much stake to gain control of the chain). source
Chain security is a very important factor and if one account with large enough stake is able to choose the entire top 20, this means that there is risk for centralized governance and coercion. Not that there is coercion I know of, but on the new smoke.io platform one account controls 33% of the vested stake and since it is a Steem copy, the 30 votes means that 2 accounts control the entire top 20. That is a very large future risk.
Having 7 votes I think may increase the risk of people only voting for popular witnesses but, this doesn't have to be the case and if it is, it actually increases the larger staked players role in securing more of the top 20 as they are able to spread their large stakes (we will assume they understand witnesses) more wisely across the range and have it count. It means smaller players have more say and, larger players have more influence without having only a few control it all.
As a SP holder I want to have a choice to approve enough number of witnesses that protects security and reliability of the network so IMHO number of the voting slots needs to be at least 21. I would rather go for increasing 30 than decreasing it. source
The reason is that it means more of the top can be stabilized for security but, this is always a double edged sword since the promise of security is also the road to centralization. If a large stake player decided to buy in heavily and manipulate, there is very little anyone can do and they will be able to pad the top 20 and tuck them behind an insurmountable wall of security.
I think what we are trying to encourage here is security via decentralization and if the witness pool is able to be controlled by a small number of players, we run a very large risk of creating an identical system to that which we already have in the world. Where lobby and special interest groups are able to heavily influence the outcomes and directions and even those with significant stake or numbers can do nothing about it.
It is a relatively complex issue once we consider the ability to directly buy or influence witnesses. One doesn't even need to buy in directly to outvote the community, they just need to be able to convince a couple of large staked players to make a voting choice which likely reduces the cost dramatically. However, with less influencing votes, the number tat needs to be bribed in some way would increase making cost and complexity of organization much more difficult. A government, corporation or bank for example will have the capital to invest to buy in currently and control or heavily influence but if the witnesses and staked voters are more decentralized through reduction of influence, this reduces the likelihood of success.
This puts a lot more responsibility and power on the small to mid-range active accounts who in time, will hopefully be key players in the decentralization process. At the moment there is little impetus to even learn who and what the witnesses really do as there is not much any individual of even medium stake or a group of smaller staked players can do about it. Empower them to have more say over the governance and more will learn. This has knock on effects for projects, innovation and SMTs without risking the key positions at the top.
The move to 7 witness votes would likely leave much of the top 20 in similar positions as they have projects, services, business and developments that support their position but, it would create some dynamic play in the next positions where witnesses must more actively prove themselves to the community. Those that are able to do so are also likely to get a shift or two of larger stake accounts to bump the tail of the top 20 around. What this could mean is that while there are highly stable and professional witnesses in the 1-15ish slots, there is a process of succession planning if for whatever reason they choose to disable.
I also means that those in the top 20 are in somewhat riskier positions and must prove themselves often enough to maintain their stability. What we don't want to create in the witness selection is the idea that it is a job for life where complacency and group norms based on relationship considerations seep in to undermine the very decentralization and stability it offers that we seek.
What is fantastic about this community is that even though there are different perspectives and disagreements, conversations can still be had and those who are consistently involved are able to continually develop their views and actions. As I see it, the more we can sort out in this foundational stages of the platform, the healthier the platform can be in the future and decentralization and distribution of stake is part of the security mechanism and the value structure in Steem.
The more interest the blockchain gets, the more it becomes a target of malicious players. Having engaged, knowledgeable and staked users will go a long way to making sure that the witnesses are in the right positions while allowing for dynamic changes as attitudes and the needs from the blockchain change. The blockchain ecosystem is going to move heavily over the next few years and we have to be able to move with it at the very least. However, what we should be aiming to do is to lead some of the shifts. This requires environmental sensitivity, not complacency and security in what we already have.
The discussions and negotiations on governance should never stop here as there is no perfect system. We do know some things that don't work though and we can see the results of them in centralized state controls, stockmarkets and the majority of people in this world that struggle needlessly. If we can't do better as a community, who's task is it?
Taraz
[ a Steem original ]