I just posted an update on github (
https://github.com/bitshares/bsips/#2) concerning BSIP #5.
In it I said that particular BSIP does not include a minimum level of content for a BSIP, and further suggested less formal discussion should take place elsewhere to distill the issues involved and see if a more formal BSIP to address the perceived problem is warranted.
There is much I could say here, but I will try to keep this brief. Most of my thoughts can be found in the BSIP issues on github. For some recent background see
this forum poll. For older discussions about governance and voter apathy use the
forum advanced search, use the keywords VOTER APATHY and change the "Search order" drop down to "Oldest topics first".
BSIP #5 attempts to address the issue of voter apathy by allowing the votes for witnesses to expire. IMO this is not a very well thought out approach. Moreover, if the approach is sound why not apply it to other votes, such as for committee members? The BSIP approach could cause instability in the witness ranks, or allow unqualified witnesses to gain the position. The role of witnesses in our DPoS ecosystem is extremely important. Reputation is a key factor, as well as technical abilities. However this BSIP does not address those issues. In fact, it weakens stability and the role of witness reputation.
We saw what can happen in BitShares 1 with witnesses that didn't take their role seriously, were lax in maintenance yet had huge votes and were difficult to get rid of, or witnesses that abandon their posts and thus reduce shareholder accountability. It has been suggested on several occasions that we codify some automatic performance metrics for monitoring witnesses so that if they fail to meet certain standards they loose their position. Missing blocks is not a sufficient disincentive IMO to keep witnesses operating properly, but it may serve well as one of several metrics (version / info publishing, frequency of reports or info updates, feeds, their number, accuracy and frequency just to name a few) in an automated system of accountability / performance evaluation.
I consider this post to be a starting point for discussions about how to strengthen the role of the witnesses, including BSIP #5. I believe we as a community should think carefully about BSIP #5 as well as the following and focus our discussion along these lines:
1) The issues (voter apathy, lack of fresh witnesses, stability, incumbent witness accountability)
2) Considerations (reputation, technical abilities, ecosystem knowledge, conflicts of interest)
3) Candidate process (mentoring / training, accomplishments, voting campaigns, motivation)
4) Policies (metrics for monitoring witness performance, decentralization, anonymity, feeds, communications, grounds for termination, process for termination, minimum vote threshold, minimum technical standards)
5) Standards and best practices (some of these overlap with items above)
IMO this community has been playing fast & loose regarding a very important aspect of our DPoS ecosystem. Whenever the voter apathy topic arises it doesn't seem to resolve much. I think we're at a place of maturity now that demands we put more thought into this, and not just apply a patch to experiment on how it may affect the ecosystem. I highly applaud xeroc and his efforts to establish quality documentation and formal procedures such as BSIPs which are sorely needed in BitShares. A more regimented, structured approach such as what jakub and Bill Butler bring to the table will help us all to stay focused and on point.
As I said in my comments on github, my concern is not so much about resolving or affecting voter apathy as much as it is on strengthening the witness role. I believe we can do that more effectively through policies, procedures and standards to improve awareness, accountability and core strengths of our witnesses.
I hope this thread will demonstrate the strength of our witness community by the number of witnesses that comment here. Let's see if we can get all 25 to show up and offer their input, no matter what it is.