Thanks to @dorian-lee for bringing this up.
There is an ongoing issue with rewards. Let me tell you about it. Steem has several models for handing out rewards.
The most recent version of Steem (the one we built on) uses what's called the convergent linear reward curve.
Unfortunately, the convergent linear reward curve assumes history, and had no history at launch.
In hard fork one, an adjustment was made to the rewards curve. I think to a value, recent_claims. Which would kind of give it a history, but the adjustment was not large enough.
You may remember that we were supposed to launch two chains on July 4th 2020, blurt and zapata.
As it were, we launched and got it rolling, and not too long after realized that we had shipped with some rather significant bugs. One of those bugs was the reward pool bug and the other one was the resource credit bug. Both of them were related to the fact that Blurt had no history, it started from zero and so those two systems malfunctioned.
For resource credits, we are studying the issue. My personal point of view is that transaction fees are actually better. I think that transaction fees are better because there is no such thing as a free transaction. It simply does not exist.
There is a cost for storing things on the chain, and in our rushed patch solution, we actually allowed the people who bear that cost, the witnesses to set the fee level. So I am pretty comfortable with the solution for fees / resource credits.
However, for rewards it is still not fully solved. I think you may recall that now, rewards are a bit bigger than they were before. However, they are not as big as they should be.
There are two ways that we can address this, I think. I could be wrong.
- move to linear rewards
- readjust recent_claim
Here's hf1: https://gitlab.com/blurt/blurt/-/merge_requests/49?commit_id=ea40812450ce4d6705190a17c2d5d58af552d880
@baabeetaa did I sum that up about right?
보상 문제는 중요합니다.
빨리 정상화 되기를 바라며 이런 상황으로 다른 의혹이 없도록 투명하게 해주시기 바랍니다.
Sorry, are there suspicions?
If there are, could you please tell me what they are so that we can address them. We are running a very clean, very transparent chain.
If you have any questions at all we are happy to answer them.
큐레이션과 저자 보상이 정상적으로 분배가 안됩니다.
이 문제가 언제 해결되는가요?
버그라고 하는데 그렇다면 버그로 인한 미지급된 보상은 언제 지급 되나요?
Hard Fork 2
https://gitlab.com/blurt/blurt/-/milestones/2
HF2 is due 8/1/2020.
https://gitlab.com/blurt/blurt/-/issues/40
https://gitlab.com/blurt/blurt/-/issues/57
The most likely scenario here is that the reward pool is just going to pay out after the change. It does not seem easy to correct the fact that earlier posts have not received fair rewards.
I suggest the following:
Keep fees and keep RC's as a backup, witnesses could decide to have no fees at some point and you want so other fall-back system to kick in. Also RC's help to further limit things should a large stake account not really care about the fees cost.
Convergent Linear rewards are a tax on the poor, it was touted as a way to limit micro-spam and self-voting but we have fees for that. The fact is we are promoting a world of fairness so we should not have a system where big stake accounts are unaffected and small accounts have to struggle to get over the convergent threshold. Linear is fair for everyone and we have fees to reduce micro-spam.
RCs are being disabled in HF2 but will remain in the codebase forever.
Linear it is!
수수료와 선형보상은 좋은 생각입니다. 시스템은 단순해야 합니다. 복잡하다고 우월한 것이 아닙니다. 설명하기도 쉬워야 합니다. 지금의 스팀 보상곡선을설명할수 있는 사람은 100명도 되지 않을것입니다.
단순한 논리로 작동하도록 만드십시요.
선형보상으로 변경시 큐레이션 보상도 선형으로 작동할수 있도록 코드를 변경해야 하는 것을 잊지 마십시요.
Thank you so much for this valuable feedback.
one of my big issues with moving back to resource credits is that after reviewing the code and all available documentation, I cannot with any clarity tell you how the resource credit system works.
I can tell you the simplified version that was told to users, but unfortunately that's not how it actually works.
So, I will take that advice into account and move toward keeping the fees, in a way that is controlled by witnesses.
When it comes to curation compensation, how would you feel about reducing the compensation for curation to 10%?
What I mean is a post's author does the creative work. They generate the value. Curation does have value, but I think that in our current model, we are overvaluing it.
Before hard fork two, I will make a post detailing all proposed changes. We very much value feedback from the community, so when it comes up please do feel free to comment on it and suggest any changes that you would like.
In addition to that, we do code reviews. Usually, this just consists of @baabeetaa passing a merge request to me for code, or of myself passing a merge request to @megadrive for docs.
In discord, we have a channel, git.
In that channel, you can see all of the commits, merge requests, pipeline bills, and everything else in the blurt organization on GitLab.
seems to me like it might be a good idea to rig up an account that serves the same purpose and makes posts to Blurt every time something happens in our GitLab org.
There are too many chains to create original content but I am here curating and commenting just because I like supporting projects.
If my curation is 10% I wouldn’t like that...
We won't be changing curation rewards after all.
Thanks for your feedback!
I love this idea!!!
My suggestion to lower curation rewards was quickly replaced by better suggestions from @baabeetaa and @megadrive
It would need to be a time based (daily maybe?) digest for automatically posting, we push about 10-20 notifications through the discord channel daily.
I would like to learn more about how much my upvotes are worth. I see that they are worth 0 but I have 2800 blurt power. Seems like they should be worth something.
Currently such votes seem to be worth less than 0.01, not sure exactly how much.
There is work ongoing to remedy the problem, as it is a bug caused by the algorithm assuming history (which we don't have yet).
Presumably we would want to see a situation where 1000 blurt can tickle a reward by 0.1+
The current curve and its constituent constants clearly don't produce this effect. For the time being there is a few roving whales dropping some 0.1-1.0 sorta level upvotes.
There may be a second hard fork to tweak the parameters that calculate the rewards to bring them up to a more reasonable level. The rewards should be pretty nice once it's fixed, we reset the inflation rate to 10% at genesis so it's higher than the 8% or so in force on Steem and Hive chains.
They're worth something, but honestly I couldn't tell you exactly what they're worth, I apologize.