Hard and Soft Protocols

Overview

The toy model from the previous section illustrates how references can revaluate older posts in order to reward and punish contributors in the long term. Our next purpose is to analyze how various choices of parameters may incentivize different behaviors. For example, some choices preference protocol development versus off-chain work, some choices preference recruitment versus retention, some choices preference conservative policing versus liberal productivity. At different stages of any organization’s life cycle different qualities will be more valuable, so a healthy DAO will require a dynamic governance system, which can select and transition between all of these qualities.

In this chapter we give an overview of the generalities of DAO governance: the types of rules experts will follow in order to win validation pools (called hard and soft protocols), and the mechanism for developing those protocols in a democratic fashion. In the following chapter we explore how a DAO can choose parameters which promote their desired governance structure.

There are two categories of rules by which a DAO’s evolution is governed: hard and soft protocols.

Hard protocols

Hard protocols​ are rules which are programmatically enacted, such as how and when tokens are created and distributed in the validation pool and how their values are calculated due to references. From a legal governance perspective, hard protocols can be compared to directives (EU Law) or constitutions (US Law).

Soft protocols​

Soft Protocols are socially agreed upon rules which dictate orthodox behavior for its members, clarifying the voting culture of the DAO. A history of posts in the forum specifying proper behaviors and values will inform members on how to successfully vote in future validation pools. Faithfully following these soft protocols means a member’s vote properly polices each contribution, reinforcing the consensus of all parties in the DAO. Violating these protocols would ideally be equivalent to voting against the majority, and would justly result in loss of sem tokens. Such validation pools strongly incentivize uniform consensus on routine policing of work-evidence posts, since valuable reputation is lost every time an expert votes against the majority. The purpose of such validation pools will be to police off-chain work, algorithmically checking that the protocols established by the DAO in the forum are properly followed. Almost every pool of this type is expected to have unanimous agreement. Therefore, typical validation pools are structured in a way which inhibits authentic deliberation, and so this structure is not appropriate for deciding contentious issues, such as policy or protocol changes.

Loosely vs. Tightly Coupled Voting

Protocols, both hard and soft, will be developed in the forum through a continual series of posts which posit the values and behaviors of members that should be rewarded. Such protocol development posts will likely be contentious since posters will not be sure it will be accepted until the members are polled, and therefore will involve posts with minimal fees since posters will be less likely to risk larger stakes until the sense of the body is determined. Therefore, Semada offers a mechanism which allows the poster to initiate a ​contentious validation pool with loosely-coupled voting or tightly-coupled voting. ​Tightly-coupled voting​ refers to votes where all sem tokens staked in the minority are lost and distributed to the majority voters. Tightly-coupled validation pools are appropriate for work-evidence posts where consensus is almost always unanimous.

Under ​loosely-coupled voting the experts do not risk their tokens in case they fail to vote with the majority. This encourages members to register minority opinions on contentious posts without risking loss of reputation. Consensus on contentious protocol-development posts is then loosely gauged by the strength of vote. Ideally, novel protocols will not be instituted until overwhelming participation and consensus is achieved.

In order to enforce a limit on the time of a debate, the DAO may choose to include a mechanism where validation pools transition gradually from loosely-coupled voting to tightly-coupled voting. This transition allows authentic deliberation to take place during the loosely-coupled voting phase, since divergent views are not punished. When the debate transitions to the tightly-coupled voting phase, the tragedy of the commons problems of meaningless votes (the “nothing-at-stake problem”) and non-participation (i.e., free-riding) are avoided, fairly and stably, because members who do not participate effectively will lose reputation or at least lose opportunities to gain reputation tokens, whose future value will greatly increase if the protocol improvement successfully attracts fees to the platform.

The above table categorizes the difference between contentious and non-contentious posts. Non-contentious posts, such as work-evidence posts, will be rigidly policed and binding in the allocation of tokens. Protocols will start as flexible and nonbinding when they are validated as contentious posts. “Flexible” means members will not be required to follow initial contentious post protocols subjected to loosely-coupled votes. “Non-binding” means no sem tokens are sacrificed for violating the content of such posts. As consensus is gradually achieved and protocol suggestions become validated in tightly coupled validation pools, the protocols become rigid and binding, so that violating them results in loss of sem tokens.

This process of moving protocol development posts from contentious to non-contentious is meant to address wicked problems which are generally intractable, yet consensus must still be reached in a timely fashion. Therefore a well-designed method for encouraging careful deliberation is crucial. Approaches from operations research (e.g., problem structuring methods) to social and political science and history (e.g., quaker governance) will help organize the process, depending on the particular application and DAO.

***

The problem of “non-representativeness” is naturally addressed by the structure of Semada. This problem is that a user’s power in most blockchains derives from their fungible currency holdings. In those cases, the vast majority of token holders will generally not have any experience or interest in protocol upgrades, nor will their interests and incentives necessarily align with the goal of long-term profitability and health of the platform. Further, wealth (not expertise) is power in such systems. So token holdings in most systems is not a fair representation of one’s ability to make sage governance decisions. Non-representativeness is a type of tragedy of the commons problem which leads to expensive and degenerate free-riding. In Semada, however, governance questions will be answered by experts according to their weighted sem reputation token holdings in that specific expertise.

***

The collection of currently validated hard and soft protocols comprise the legal constitution of a DAO. Similar to nation states, most members of a DAO will never exhaustively analyze every legal facet of the constitution. So the UI design is as important as the constitution to the success of a DAO. If hard protocols are compared with a nation state’s constitution, then the UI is similar to supporting administrative agencies.

To protect minority interests and ensure stability, DAOs often would choose to institute certain initial charter rules which are hard coded to require near unanimous consensus to change, similar to a nation state’s foundational constitution. By protecting minorities from the inevitable transient interests of the majority, this would encourage recruitment of a diverse population of members.