Traversing the modification course of


Tezos is a self-amending blockchain community which includes a formal, on-chain mechanism for proposing, deciding on, testing, and activating protocol upgrades with out the want to laborious fork.

On this publish, I describe how the Tezos modification course of works in follow and suggest a number of methods to enhance it within the near-term. Deeper questions, like splitting voting rights from consensus, or how amendments needs to be funded on-chain are outdoors the scope of this publish, however necessary matters for future consideration.

This piece assumes a primary data of Tezos baking. I like to recommend studying “It’s a baker’s life for me” and the Tezos proof-of-stake documentation to get began.

* I take advantage of the expression “modification course of” as an alternative of “election cycle” to keep away from confusion with “cycles” of baking.

Understanding the Tezos modification course of

The modification course of could be damaged into 4 discrete durations: the Proposal Interval, the Exploration or “Testing” Vote Interval, the Testing Interval, and the Promotion Vote Interval. Every of those 4 durations lasts eight baking cycles (i.e. 32,768 blocks or roughly 22 days, 18 hours), comprising nearly precisely three months from proposal to activation.

As summarized within the flowchart diagram beneath, any failure to proceed to the following interval reverts the community again to a Proposal Interval. In different phrases, failure to proceed restarts your entire modification course of.

That is exactly what has been occurring because the Tezos genesis block was adopted by the group in June. Within the absence of any proposals, a brand new proposal interval has restarted each eight cycles since Block 1.

So how does it work?

  1. Proposal Interval

The Tezos modification course of begins with the Proposal Interval, throughout which bakers can submit proposals on-chain through the proposal operation, which takes “supply”, “interval” and “proposals hash” as parameters. The “supply” is the baker who submits the proposal operation, the “interval” specifies the particular proposal interval by which the proposal was submitted, and the “proposals hash” is the hash of the tarball of concatenated .ml/.mli supply recordsdata.

Bakers could submit as much as 20 proposals in every Proposal Interval. When submitting a proposal, the baker can be submitting a vote for that proposal, equal to the variety of rolls in its staking stability at the beginning of the interval.

For these eager to observe alongside, (and hopefully different Tezos block explorers) lets you watch incoming proposals.

Different bakers can then vote on proposals by submitting proposal operations of their very own. As described within the white paper, the Proposal Interval vote is completed through approval voting, that means every baker could vote as soon as on as much as 20 proposals. Consider it as a type of “upvoting.”

On the finish of the Proposal Interval, the community counts the proposal votes and the most-upvoted proposal proceeds to the Exploration Vote Interval. If no proposals have been submitted or if there’s a tie between proposals, a brand new Proposal Interval begins.

2. Exploration Vote Interval (referred to as “Testing Vote” within the present protocol)

Within the Exploration Vote Interval, bakers could vote on the top-ranked proposal from the earlier Proposal Interval.

Voting is completed on-chain through the Poll Operation, which takes “supply,” “period_n,” “proposal,” and “poll.” The “supply” is the baker, “period_n” is the particular voting interval by which the operation was submitted, “proposal” specifies the proposal, and “poll” specifies the baker’s vote, both “Yay,” “Nay,” and “Abstain.”

As within the Proposal Interval, a baker’s vote relies on the variety of rolls in its staking stability on the starting of the Exploration Vote Interval. Every baker could ship a poll operation solely as soon as throughout the voting interval.

On the finish of eight cycles, the community counts the votes. If voting participation (the entire of “Yays,” “Nays,” and “Abstains”) meets the quorum (defined beneath) and an 80% supermajority of non-abstaining bakers approves, the proposal proceeds to the Testing Interval.

If the quorum or the 80% supermajority should not met, the modification course of reverts to the start of the Proposal Interval.

A notable distinction from the Proposal Interval is that after the Exploration Vote Interval, the quorum is up to date as follows the place Q_t is the earlier quorum and q_t is the entire participation:

This adjusts the quorum to match an exponential shifting common of the previous participation charge.

Within the genesis block, the quorum started at 80% and can stay at 80% till the tip of the primary Exploration Vote Interval.

3. Testing Interval

If a proposal is authorized by a supermajority within the Exploration Vote Interval, the Testing Interval (eight cycles) begins with a testnet fork that runs parallel to the primary Tezos chain for 48 hours earlier than terminating. This 48-hour fixed has been set conservatively with a purpose to cut back the chance of the community perceiving the testnet fork as the primary chain.

Simply as described within the 2014 white paper, proposals have entry to a small commonplace library, however are sandboxed and cannot make any system calls.

The aim of the Testing Interval is to guage whether or not a proposal is a worthy modification to the protocol.

The testnet fork ensures the improve doesn’t corrupt the context and that, ought to the improve be adopted, the community would proceed making legitimate state transitions. Nonetheless, 48 hours of testing by itself is decidedly too quick to find out whether or not a proposal could be a worthwhile and secure modification. Particularly, it could possibly keep liveness longer than 48 hours and doesn’t include safety vulnerabilities). Consequently, a testnet matching the modification proposal is prone to run off-chain for the remaining ~7.Three cycles of the Testing Interval, permitting stakeholders to guage and focus on the modification as achieve higher data of its properties.

4. Promotion Vote Interval

On the finish of the Testing Interval, the Promotion Vote Interval begins. On this interval, the community decides whether or not to undertake the modification based mostly on off-chain discussions and its habits throughout the Testing Interval.

As within the Exploration Vote Interval, bakers submit their votes utilizing the poll operation, with their votes weighted proportionally to the variety of rolls of their staking stability on the starting of the Promotion Vote Interval. As within the Exploration Vote Interval, every baker could ship just one poll operation throughout this era.

On the finish of the Promotion Vote Interval, the community counts the variety of votes. If the participation charge reaches the minimal quorum and an 80% supermajority of non-abstaining bakers votes “Yay,” then the proposal is activated as the brand new mainnet.

In any other case, the method as soon as extra reverts again to the Proposal Interval. The minimal quorum is up to date based mostly on the participation charge utilizing the identical method which was used on the finish of the Exploration Vote Interval:

As soon as the amended protocol is activated, a brand new Proposal Interval begins and the modification course of begins anew.

Enhancing the Tezos Modification Course of

Because it stands, the present Tezos modification course of is designed with a easy, however efficient mechanism for amending each the protocol and the modification course of itself.

The present design is well-suited to bootstrapping enhancements to Tezos in its early days, however beneath are 5 obtrusive alternatives for enchancment within the present on-chain implementation.

There are additionally alternatives to encourage wholesome coordination off-chain. Two examples that come to thoughts are incentivized testnets (a la Cosmos’ Sport of Stakes) and a dialogue discussion board based mostly on repute/stake), however these fall outdoors the scope of this piece.

  1. Lengthen the proposal interval and introduce proposal charges

Lengthening the proposal interval to 12 or 16 cycles would enable extra time to debate and coordinate round proposals. Moreover, a requirement that proposals be submitted throughout the first eight cycles would guarantee ample time for dialogue earlier than the modification proceeds to an Exploration Vote Interval. This requirement would additionally stop proposers from sniping the tip of a Proposal Interval by which no proposals or solely weakly upvoted proposals had been submitted.

Proposal high quality additionally represents a modest danger within the present Tezos community, notably if some bakers face a coordination price in distinguishing reliable proposals from malicious or mediocre ones. In the end, it needs to be cheap to vote on proposals, however submitting one ought to incur a value to discourage spam and frivolous proposals.

Introducing an upfront proposal price of say, 500 xtz looks like a easy but affordable precaution that reduces the chance of proposal spam within the early days of the community.

Alternatively, a time-locked deposit might be priced at a better degree (e.g. 2,500 xtz) and returned to the proposer if the proposal is upvoted by at the very least 10% of the voting weight within the Proposal Interval.

2. Consider voting weight on the finish of voting durations, quite than the start

As a result of bakers’ staking balances are counted at the beginning of every voting interval, delegators can’t change their delegation as soon as voting has begun. This forces delegators to belief delegates to commit honestly to the votes they signaled earlier than the beginning of the interval.

Evaluating a voter’s staking stability on the finish versus the start of the voting interval would enable delegators to maneuver their stake as soon as they’ve seen how their delegate voted. Not less than for now, this additionally encourages delegates to create one delegation tackle for these preferring “Yay” for a given proposal and one other for “Nay.” This might enable delegators to maintain their most popular delegate(s) however nonetheless maintain affect within the modification course of.

At first this will likely appear counterintuitive, as a result of delegators are used to ready 7 cycles earlier than their delegated tokens really present baking rights. In actuality, rolls are counted in actual time when delegated, despite the fact that they have an effect on baking rights solely after 7 cycles.

That stated, switching additionally incurs different significant transaction prices. There are search prices related to discovering a brand new delegate who votes as desired, however who doesn’t tradeoff payout construction or reliability. One other concern is {that a} delegate might attempt to retaliate in opposition to delegators who change throughout a voting interval. One attainable answer could be to allow delegators to secretly override their delegate’s vote.

3. Longer testnet forks and an extended testing interval total

Lengthening the general Testing Interval to 12 cycles from eight would offer stakeholders extra time and information by which to guage and focus on the properties of a given proposal.

That stated a 2-day testnet fork could not present an extended sufficient interval to guage the persistence of the amended protocol. Conducting the testnet fork for one week might incrementally enhance the realism of the on-chain take a look at fork.

Proposals can have been examined in off-chain environments lengthy earlier than the Testing Interval and the Testing Interval could evolve into extra of an “Analysis interval”.

One other chance could be to finally create a testnet that’s permissioned on-chain through a DAO or elected council. The choice-making latitude of this council might be narrowly specified: the council’s functionality might be restricted to delaying development to the Promotion Vote Interval by a hard and fast variety of cycles for the categorical goal of permitting extra time for additional analysis. In contrast, a extra aggressive strategy would entail the council voting to advance the proposal to a Promotion Vote Interval at predefined intervals (e.g. at eight cycles or 16 cycles).

4. Quorum flooring

The Tezos quorum adjustment mechanism is a straightforward, but highly effective software to keep away from stagnation because of low stakeholder participation and misplaced cash. Nonetheless, many successive Voting Intervals (e.g. 1 yr) with low participation might drive the quorum to low ranges, permitting proposals to cross with out a lot help. Though malfeasance could be blocked by mobilizing “nay” votes, setting a minimal quorum could also be worthwhile as an extra safety.

One chance could be for the supermajority approval threshold within the Voting Intervals to regulate based mostly on the participation charge. Underneath this mannequin, the community would require an 80% supermajority to vote “Yay” at a 60% participation charge and a 90% supermajority at a 50% participation charge.

Alternatively, it’s additionally attainable to require the quorum to be higher than the typical participation charge. For example:

Following this method, if the typical participation charge had been 0%, the community would keep a flooring of 20%. If the typical participation charge had been 50%, then the minimal quorum could be 60%, and if the typical participation charge had been 80%, then the minimal quorum could be 84%.

5. Delay activation of the amended protocol

The present Tezos protocol prompts amendments immediately on the finish of the Promotion Vote interval. Nonetheless, it could be prudent to delay activation by a predefined variety of baking cycles to make sure that the community is ready for the brand new protocol (notably if the protocol incorporates adjustments to the consensus itself). This might additionally, conceivably, delay the beginning of the subsequent Proposal Interval as a buffer interval to permit bakers to regulate to the workings of a brand new protocol.

Modification RPCs

This week’s Proto3 replace has paved the best way by including a bunch of RPCs to help the modification course of. This can enable the Tezos group to work together with proposals, test the quorum, see the variety of ballots solid, and discuss with different necessary amendment-related information talked about all through this piece.


As detailed, Tezos implements a multi-stage modification course of that permits stakeholders to formally suggest, choose, take a look at, and activate upgrades to the protocol on-chain.

However we must always bear in mind, in the beginning, that blockchain governance stays a nascent subject. There may be nonetheless little in the best way of immediately analogous expertise to attract upon.

This calls for that we wade in slowly and take nice warning. Incentivizing proposal high quality, encouraging higher delegate accountability, and offering an extended testing interval appear affordable first steps to enhance the modification course of. However these are merely locations to start out.

What’s most necessary is that we study from our early experiences with the modification course of. And even with early success, we needs to be cautious of the prevailing system’s limitations, notably its reliance on voting, over the long term. It’s vital to discover complementary mechanisms: particularly, futarchy, programmatic constitutionalism, and on-chain councils.

In spite of everything, blockchain governance shouldn’t be merely concerning the “subsequent” improve, however quite about creating internet-native establishments which can be constructed to final, ones that may be adopted sustainably by a whole bunch of tens of millions, if not billions of customers some day. Introducing the Tezos modification course of goals to be a step in direction of that aim.


Tezos White Paper

Tezos Place Paper

Proto3 replace documentation (consists of modification RPCs)

Read the original article here