Towards a revised Oxford proposal

On October 2nd, the promotion period vote for the Oxford protocol proposal ended with a clear result – Oxford did not gather the supermajority of votes required to trigger its activation on Tezos Mainnet.

We thank the Tezos community for their engagement in the governance process and the discussions of Oxford’s features. It is a cornerstone of our governance system that protocol amendments are only accepted if there is a clear consensus that these changes will be beneficial to the whole Tezos ecosystem. This time around a consensus could not be found.

At Nomadic Labs, together with the other contributors to the Oxford proposal, we value the feedback provided during (and before) the governance cycle.

We are also grateful for the many efforts to test and integrate Oxford, notably Adaptive Issuance and Staking, into ecosystem tools. These were fundamental in finding a few non-critical bugs in the implementation of these features in Oxford.

A revised Oxford 2 Proposal

We are working towards publishing a revised protocol proposal, Oxford 2, which will not enable bakers to vote to activate Adaptive Issuance and focus on staking for bakers.

Our plan for Oxford 2 includes:

  • Including additional information in block receipts related to reward distribution and staking.
  • Reducing the double-baking slashing penalty.
  • Fixing the reported issues in Adaptive Issuance and Staking’s implementation – even when this feature would be disabled for the whole lifetime of Oxford 2.
  • Enabling automatic staking for bakers.
  • Disabling the features originally guarded behind the feature activation vote: Adaptive Issuance, the ability for delegators to become stakers, and the changes in weight for staked and delegated funds towards baking and voting rights.
  • Including all the features already present in Oxford (re-enabling Timelocks, completing the renaming of endorsements into attestations, and new improvements to Smart Rollups).
  • Potentially including a new feature from the alpha development protocol – private Smart Rollups – if it is mature enough come stabilization time.

We will provide more insight into the implementation of this plan in due time.

Moving forward

In addition to the technical work described above, we will reinforce outreach actions and improve feedback loops in the community. We want to ensure that all features and design choices in the Oxford 2 proposal (and its eventual successors) are well explained and that the community has access to necessary resources.

We look forward to continuing working together to keep Tezos ever-evolving.

4 Likes

@NomadicLabs Taking into account that Oxford proposal achieved majority (>50%) during the former promotion period… Could you clarify whether Adaptative Issuance will be included in the “Oxford 2” proposal or not, and the reasons for so? Thanks in advance

How about working on things people actually want instead of stuff that doesn’t really help adoption? The bakers clearly spoke with this rejection, so why continue down this path of resistance? Help adoption first then work on more incremental stuff.

1 Like

Oxford proposal achieved majority (>50%) during the former promotion period, how can you deduce that people rejected that proposal? And how can you deduce that Adaptative Issuance wouldn’t help for Tezos adoption?

Furthermore, I am also convinced that >80% of Tezos community are convinced that Adaptative Issuance is a good thing for Tezos, so it is just a matter of time that redelegations occur and supermajority is achieved.

2 Likes

Could you clarify whether Adaptative Issuance will be included in the “Oxford 2” proposal or not

They have confirmed this already:

Disabling the features originally guarded behind the feature activation vote: Adaptive Issuance

The code for AI will be inside the new protocol, but it will be impossible for it to be activated. A later proposal will re-enable the ability to activate it, after more discussions/plans/debate etc on its design take place

Oxford proposal achieved majority (>50%) during the former promotion period, how can you deduce that people rejected that proposal?

The target for accepting the proposal is >80%. A lot of bakers voted No, and did so for the first time. A lot more didn’t vote during this phase at all. Saying that it had a clear majority is arguably not true. This is a contentious issue at the minute, while I’m fully aware I don’t represent the entire community, I for one am not convinced this is a good thing at all. I’m deeply worried this is going to have a lot of negative blow back

1 Like

Quorum was met.

And a lot yay, >50%, for being specific: that means that the proposal is good in general terms but it needs to be rethinked a bit to achieve supermajority.

I really do not understand the purpouse of including Adaptative Issuance on “Oxford 2”, considering it will be impossible to activate it. Why not try to incorporate as much requirements as possible from the bakers feddback that voted no, and trying again with the new revised proposal? For example, including a proposal with a % staking more attractive for bakers like Oxlo and with all the technical bugs solved.

Quorum was met

and supermajority was not

And a lot yay, >50%

Not nearly enough to say with confidence “this is what everyone wants”. These are massive changes to an entire platform/economy, >50% is not enough to endorse such drastic changes

I really do not understand the purpouse of including Adaptative Issuance on “Oxford 2”, considering it will be impossible to activate it.

The technical work of taking it out is difficult. Why waste time performing work that doesn’t need to be done? If it takes 1 hour of dev time to leave it in place + make it impossible to pass, or 3 weeks to take it out and test to make sure this hasn’t introduced new issues … why choose extra time and risk?

They also want it to be in eventually. Again, why spend extra time taking it out, only to spend extra-extra time putting it back in later. Leaving it in and disabling it is simply the easiest + quickest path forward. Anyone who wants AI to eventually be enabled, should be happy that they’ve decided to not strip it out, because thats the easiest way to get it part of the next protocol

Why not try to incorporate as much requirements as possible from the bakers feddback that voted no,

What lead to the issue we are in now, is poor communication with the community (not trying to blame anyone). The worst thing that could happen right now, is making a snap decision to change it hoping that no other group of bakers have issues with the new changes

The core teams have stated repeatedly, their intention was not for AI to activate on O immediately, but for O to start a debate about whether or not AI should be activated. Their expectation is that AI would have taken quite some time after that to activate. This in part, is one of the issues, bakers didn’t like this. Many wanted the debate to happen off-chain first, and then those findings/learnings to be put on-chain. The core teams have other things that they want + need to get in to the protocol. Rather than risking another debate causing more delays, they’ve decided to have the AI debate off-chain and allow the other work to not be blocked by it

It is already done… just check the Oxlo proposal.

A) That doesn’t remove AI, that changes the metrics so it functions slightly differently. E.g. the target 50% staking remains the same as mentioned in those links

B) There is no indication that this addresses anyones concerns, or that this would pass instead. More debate/discussion is needed to figure that out

C) Everstake posted other concerns (e.g. overstaking), and thats not been addressed by this either

Pushing this instead brings a huge amount of risk of the proposal being rejected again, blocking other work from getting in to the protocol. Safest thing to do is to disable it until these conversations can be had, so other work can continue

1 Like

That does not remove AI but it addresses low-issuance concerns from some bakers that voted Nay to Oxford 1, like Baking Benjamins.

Other technical concerns can be amended without compromosing consensus (and most of them will be already addressed on “Oxford 2”).

without compromosing consensus

Consensus wasn’t reached, so therefore no action is compromising it

and most of them will be already addressed on “Oxford 2”

Technical in terms of bugs, yes. Technical in terms of AI’s implementation? absolutely not. Everstake listed issues which have not been addressed, and currently has no proposed implementation. If i’m reading it correctly, they also want to see changes in how the underlying calculations are made, and how AI rolls out to users. These sound like big changes to me, that would require a lot of research and not something that could be produced and tested in a few short weeks

Ecosystem teams such as wallets and indexers also want to see other changes related to unstaking, which are being taken on board now. So this is not a simple case of 1 or 2 bakers said No and everyone got screwed. There are lots of issues and genuine concern around this area.

Again back to the overarching point. The community didn’t agree, large changes are being requested, more debate is needed. Disabling AI for now is objectively the best option for everyone. The team that made the decision to do this, is the same team that wants AI enabled. The protocol can’t stagnate here for ever, the best thing for everyone is to debate AI while other work continues. Disabling it short term is the only way that can happen

You are saying that, not me.

As you shared your opinion, I will share mine for closing the discussion with you: that bakers are mostly concerned by the low-issuance factor… cause other technical things that also concerned to wallet devs etc… in the past, have been improved from protocol amendment to protocol amendment without having that multi-chain big profit bakers blocking that past amendments.

that bakers are mostly concerned by the low-issuance factor

Everstake has listed many issues, that was one of them, that wasn’t all of them. Fixing low issuance does not make AI pass. Once again, the team behind AI chose not to repurpose Oxford/Oslo/Oxlo with higher issuance … because that would not have addressed the issues raised. The team behind AI, that wants AI to pass would have done so if that was viable

in the past, have been improved from protocol amendment to protocol

These have been small bug fixes or entirely new features, not a complete re-write of core logic changing how operations work, which would have resulted in large client compatibility issues

AI was not ready in its current state, developers across the ecosystem didn’t know how to make use of it, it contained numerous bugs and issues, bakers were unhappy, it comes with a lot of risk that is poorly understood. It would have been very damaging for the ecosystem had that proposal have gone out. Teams are now playing it more cautious, which is the right thing to do. The fact that you wanted AI doesn’t change any of these facts

Yes, yes, I already said you it was my opinion… of course I can’t prove that big bakers aren’t honestly concerned for wallet devs and so on… as much as for their pockets.

Anyway, according to your argues I guess that when all that technical stuff is implemented, that big bakers will vote Yay for sure, no? We will see it, we have a long record of the big baker you mentioned saying A and voting B afterwards :slightly_smiling_face:

We have shared two draft documents describing the design changes for slashing and staking UX for bakers in the upcoming Oxford 2 proposal in this thread:

As the title suggests, we aim to gather feedback from the community, in particular from bakers and infrastructure developers.

:pray: Please use the quoted thread for feedback and questions from these documents.