The definition of super linear is that for a given function f(x), it is super linear if for all k, there exists a an N such that f(x) > kx for all x > N. Essentially, given any linear curve, the function surpasses it at some point.
Back in HF 17, we proposed a curve that did not adhere to this definition. It was superlinear, to a point, and linear after a certain amount of time (this was done smoothly, not piece-wise). There were heated discussions over certain breakpoints of linearity. At how much SP does a vote behave on a certain part of the curve? Things of that nature. Many argued towards an overwhelming linearity to the curve. To the point where linear rewards made more sense. Yet others argued for breakpoints that caused the curve to look close to nlog(n). Even though it was better than n^2, it wasn't fair enough. The problem we run into is that no matter what super linear curve we propose, many will argue for more fairness. The limit as fairness approaches infinity is linear.
This is incredibly one-sided, including as a retrospective of HF17. Some will argue for more fairness, yes (and did), but some also argue for more resistance to abuse (and did). There are valid points on both sides, but to simply take the fairness argument, extrapolate to infinity, and implement it, is to completely disregard the arguments on the other side. On what basis?
Okay, fair enough, the platform can decide that "maximum fairness" is the right goal, but there's nothing that is a given about it, as some sort of mathematical extrapolation of agreed principles.
I agree things have changed. I was in favor of trying a linear curve reward. I tend to think any super linear would be better, up to a certain point obviously.
This is so for other people too who have been supporting trying the linear curve but now have changed their mind. I know @pfunk mentioned this in comment a couple days ago.
I've stated my perspective and reasoning pretty clearly in this comment.
I know that without the support of Steemit Inc such a change have a very low chance of happening if any at all. Making my reasoning to be known is the best I can do.
https://steemit.com/roadmap2018/@steemitblog/steemit-roadmap-2018-community-input-requested#@teamsteem/re-ned-re-teamsteem-re-steemitblog-steemit-roadmap-2018-community-input-requested-20171117t011341724z
So we sacrifice a coherent overall system and the actual social aspects of “social media” because some people complain about “unfairness?”
I hate to be the bearer of bad news then, but people who don’t earn as much as they’d like will always complain about “unfairness.” Since this platform has the money element included, those complaints are inescapable. The idea is to make the system actually work and work for the most people by a coherent set of rules/protocols, not to cater to those with absurd/mismanaged expectations and misguided “feelings.”
The largest problem with any of the previous rules was the result of the initial distribution - and I have yet to see anyone willing to address that. If this place was flawed from the beginning, as Ned suggested in another comment (below), then why pretend that any amount of code tinkering can make things “more fair?” All that happens is that a different set of problems emerge...but the origin has always been the same.
Overtime many of the root problems caused by initial over-engineering and other evils/flaws I won’t go into detail about have been mitigated, cleansed or eliminated - this place is a lot better than it was before our 19 hardforks but still it’s not perfect! It never will be and I love that. Though we will get to see lots and lots of imperfect solutions balance themselves out for the most possibly perfect combined solution through SMTs
I largely agree with @ned's points here, including that while it isn't perfect it is indeed a lot better. There's still some early over- (and especially overly-eager) engineering like the 30-minute reverse auction and lousy reputation system.
I really appreciate those precisions.
I just shared some further thoughts on the subject as a reply to Ned's comment.
I support the re-institution of a superlinear reward and I agree N2 might be unnecessarily too much with negative effects outweighing the positive ones.
I understand that it would take a majority vote by top witnesses and that it has been voted on in the past and that the sentiment might not have changed enough for this changed to be vote in at this point.
My comment was more about showing my support toward such a change.
By the way, I really enjoyed watching your presentation at SteemFest and I really enjoyed watching all the other Steeemit devs too. I already knew you guys were all very passionate about Steem and definitely hardworking but those presentation and the panel discussion and the thing I've read here and there about the team made it all the more succinct.