This is a reply to the post of 
How value is created
I think that in order to understand how rewards should be allocated we first need to be sure about how steem obtains its value. Once we know the revenue stream, there is usually one logical way to distribute it.
Of course like everything in crypto steem is valued due to investors speculation on the future. We are not even the early adopters, we are still some crazy nerds. So one might think of a business model to cater to speculators and switch later to a more sustainable model. But i believe it is better to directly build the platform for its real use-case in the future.
So what creates the value of steem? It is people wanting to buy it, but for which reasons? Below I summarise four possible reasons someone might buy steem.
STEEM as an advertising platform
This is the number one business model of current social networks like facebook. Good user content attracts attention and that attention can be marketed. Currently advertisement does not exist on steem or it only exists in the first experimental projects. I have the feeling that many community members do not like this model very much.
However if this should be adopted, it is via a smart contract that will exchange steem for views controlling what you see. People then buy steem to be able to advertise and the content producer and advertisement consumer will gain the steem (no middle mans required). This might be implemented via a steem frontend, but I think steem is lacking some smart contract functionality.
In this model voting is not an essential part and the reward curve is a gimik that can be set any way we want. Probably voting should be replaced with likes that will determine advertisement output and these likes are independant from stake (so as sublinear as it can be).
STEEM as a decentralised network
Having censorship resistant open networks is very important and certainly one of the strengths of steem. It is the main reason I am here.
This business model is currently not realised anywhere, but steem has the unique potential to do it.
In this model RC are valuable. Holding steem gives you RC's and you need these to be part of the network.
But currently it seems unlikely people will buy steem for this reason. Big accounts already have more than enough. Most people building dapps that require a lot of RC are big accounts and still have more than enough RC.
On the other side of the spectrum steem is much more unfriendly. New users suffer and are only able to post or comment a few times a week. And when the make a tiny modification to their posts, they find out that that also costs big amounts of RC. This initial experience is not what will make them buy steem to be able to interact more. On the contrary it will probably deter most.
For this problem a simple fix exists. And that is not super-linear rewards but sub-linear RC's; for example a sqrt of your stake. New smaller accounts should be able to interact much more freely, but will eventually run out when they post a lot. They might be willing to pay a bit to be able to support their frequent posting.
On the other side big accounts wont automatically have enough steem to support any dapp they want to build and might be incentivised to buy some more steem. I think this idea of sublinear RC is really worth pursuing.
The logical voting curve in this model is not important and voting could be removed. Your stake pays the witnesses for maintaining the network and thats it.
Of topic: I have started a delegation service for redfish to help them with RC troubles. But I am myself only a small fish and this project is only just starting. link
STEEM as a token of influence
This business model is well known from many free to play games and has big potential for steem. People pay to be cool, to get special character skins, ...
Steem allows you to buy influence on the platform. You can upvote for bigger amounts, you can choose witnesses. That is a big incentive to buy for many.
In my opinion this is the number one reason people that are not speculators buy steem and this seems to be the currently realised business model.
Within this logic, I think linear rewards make the most sense. This provides the same incentive for every account to buy more, no matter if they are small or big. It does not require small accounts form pools and big accounts can split up if the want.
In this philosophy, you bought the steem to get influence over the reward pool and you essentially own your share of the pool, to be free to use it in any way you see fit. Also for self-voting. Many people however bought to gain social influence, so they will likely reward their friends, or others they want to impress. Self-voting is not a crime, it is just a strategy to avoid inflation when you do not want to actively use your stake.
STEEM as a quality content platform
I think here is where all the troubles begin.
As long as content is open for everyone to consume this does not create value in a monetary sense. It of course creates a value to society, but that is free and does not need to be paid. Steem could be filled with the most amazing posts and still these posts do not need to be highly voted. Steem could be valued low even though it is used a lot.
The real question is if it is possible to set up a game-theoretical system in which voting for good content works. I think this is not a trivial question at all, but so far we have failed. And while superlinear rewards might improve I dont think it solves the issue, it just requires new strategies to exploit the system. But even then, why would I want to buy steem in that system? I can just post and be rewarded never holding any steem myself.
Combining the ideas
STEEM does not have to choose one model. STEEM can be used for advertisement, while it is also supporting a decentralised network. Steem can be a token of influence that people buy to appear more important and to gain attention when up-voting others content. In that vision we would choose linear rewards and sublinear RC.
The only two ideas that cannot be combined without friction are the last two. A compromise is needed, where anything but linear hurts the idea of steem as a token of influence and it is not clear if we can even build a free and open system where human curated quality is rewarded.
Rejecting the last idea, voting on content is no longer seen as a certification of quality, but as an expression of influence. In the end I believe that in a lively social network, quality authors will get their fair share, be it from friends that support them with their stake on a voluntary donation basis, from advertising rewards or by selling exclusive content. We do not need to build a system based on a flawed business idea that claims that content is inherently valuable and where we need complicated methods to controll that value.