A Post Mortem Report: The Constantinople Ethereum Hard Fork Postponement

hardfork
constantinople
ethereum-cat-herders

#1

Bringing in this article so that we can have a discussion around it that isn’t trapped on Twitter or Medium.


#2

One of the things I noticed was that 1052 and 1014 never went past Draft (so didn’t go through Last Call), and a bunch of us (me, @axic, @5chdn) were pointing this out and at least trying to get them set as Final / Accepted in the weeks leading up to the scheduled hard fork date.

This isn’t related to any of the issues around postponing the fork or the 1283 EIP (which did follow process), but it is an example of how we should track these sorts of things and actually follow process.

It seemed hard to get the attention of EIP editors (no reviews, no comments) on what was a basic administrative clean up (setting status on EIPs that were scheduled to be included).

There are two “tracking” issues in the EIP repo around this with more discussion, filed by @fulldecent.

Do we need more EIP editors to help with these admin tasks? How can we help?


#3

Some of the big takeaways are about overall process improvement, and failure to adhere to the listing process. I drafted a process improvement based on some FEM threads in this PR:


Some more links about EIP process improvement suggestions:


#4

Thank you very much for recognizing this. :smile:

I have worked hard on getting the two-week process adopted (this was part of getting ERC-721 approved) and I believe it has benefitted the community for the ERCs process. It certainly makes it easier for me to find which EIPs I want to review! At the time this was really a big change and risked much criticism so again I appreciate your kind words.

However this process also applies to consensus-changing EIPs and it was not followed this time.

I believe the issue is that the attitude of EF is /much/ too centralized. They prefer to push their own software changes on the world by announcing on their own channels, rather than using a mildly community-based process like the EIP two-week review.

(EDIT: Maybe “prefer” was too strong there, maybe not everyone knows about EIP-1 changes yet…)


#5

I don’t really see the EF as being in charge. There are humans who work for the EF involved, but at this point we are in the process of having many independent clients.

I see lack of consistency by editors / maintainers — if there is open-need to GitHub labels, more maintainers, and so on — I think the process can improve.

P.S. pretty please thumbs up my RSS changes or give me feedbac.