2.6 KiB
BSIP: 0005
Title: Expiring Votes for Witnesses
Authors: Daniel Larmier <Dan@cryptonomex.com>
Fabian Schuh <Fabian@BitShares.eu>
Status: Draft
Type: Protocol
Created: 2015-12-16
Discussion: <https://github.com/cryptonomex/graphene/issues/265>
<https://bitsharestalk.org/index.php/topic,18109.0.html>
Worker: tbd
Abstract
Votes for witnesses are a measure of reputation in that they cast a persistent opinion - so not like votes in the traditional sense of the word. The witness voting system is currently designed to ensure stability by directing the apathetic to vote for incumbents. Whilst this may have advantages, it seems to present the danger that incumbents are likely to attain a near-unassailable position (excepting of course those who exhibit negligent/bad behaviour).
Motivation
If the aim is the create an autocracy - albeit seeded by the active witnesses or so original gainers of significant 'reputation' - then this is fine. However, it would seem to discourage serious efforts to try and break into the active circle by newcomers with much to contribute. At the end of the BitShares 1 network, circa. 220M votes were required and a quick browse of the delegates showed that with a few notable exceptions, the majority have been in place since a 4 or 5 digit block number - have all been of these current, active contributors?
Specifications
A blockchain parameter can be added that enforces a minimal freshness to votes that are counted and handle cases of lost keys or apathetic votes hanging around.
Discussion
Linear, Exponential, or other decay
The question is left open whether the freshness of votes decreases linearily, exponentially or by any other function. It is further not defined if the decrease should scale with the number of blocks or the absolute time.
Voter Apathy
The argument has been raised that a apathy of voters would result in only few active votes which could leave the blockchain open to attacks with only little stake.
Highly Volatile Voting Power
This proposal results in a high volatilty in voting power which has mostly negative effects when votes are removed from the actively voting stake, i.e. a big stake (e.g. from a proxy) expires. This could result in significant changes that take effect instantly. An alternative approach to fix this would be to have a window (e.g. several days) in which changes in votes take effect gradually.
Copyright
This document is placed in the public domain.