This is a proposed RFC for lowering fees:
Glad to read something that makes sense.
Edit: Oh yeah. Looking forwad to the future
Out of curiosity is there a need/ reason to have hard fixed amount? Could a dynamically increasing tiered amount or a precentage be implemented?
Both options would allow for lower priced use cases and quickly become costly for bad actors
We would honestly love to do this. It turns out the control theory algorithm for doing this in a way that doesn’t result in destructive behavior is non-trivial. You can imagine a few naive ways of doing this, like just checking the fee in the last block and whether all tx cleared, and if they didn’t, raise it, but that results in the possibility of someone dos’ing every other block. You could do it as a cascading window which is weighted more towards the present block, but that has interesting potential for abuse as well. Ultimately it’s just a tricky problem that will require a lot of experimentation. One of the disadvantages of open source code is that your adversaries get to see your algorithms (but you don’t get to see your adversaries algorithms to defend against them).
The ultimate solution will likely involve some neural network but even that has opposition in the form of adversarial neural networks. Again, it’s tricky. We look forward to running a bunch of experiments this summer.
Hmmmm, I wonder if there would be a way to link the scaling fees (precentage or other method of calcualtion) to the senders wallet vs. transaction amounts.
Thinking of how some debit accounts are set up. For example in a tiered/ precentage mix it would the user wallet could send x amount of transactions in y time frame under a certain threshold for a very low network fee. This would cover the coffee scenerio. Higher transactions would have higher fees.
Bad actors could get around this by having multiple wallets, but there could be ways to manage wallet spamming
Per 30 days 20 transacations
$10 = $.01 fee
$10+ = 1%
Over 20 transactions add 0.5% to anything under $10 and another 1% to anything over.