☑️ the concept of decartography
The concept of DeCartography is a Relational Computation Oracle that utilizes Web3 crowdsourcing to provide a social graph with reliable neutrality. Dapps that utilize this oracle can leverage quantified data on cooperation and social distance as a social graph. One specific use case is the allocation and prioritization of funding for public goods, where DeCartography will be utilized in the social identity layer. This collaboration with Gitcoin, the first beta test partner, aims to reduce "collusion" in Quadratic Funding, where socially related agents cooperate to ensure that specific projects receive more grants. The unique feature of DeCartography is its social graph generation based on the Schelling Point mechanism, utilizing Web3 crowdsourcing. While existing social graph providers also analyze data and provide the results to other apps, DeCartography functions as a computational oracle based on the Proof of (Human) Work, assuming equilibrium based on the concept of the Schelling Point by assigning certain tasks to crowd workers who act as validators of the decentralized oracle. This mechanism differs from traditional decentralized oracles where validator nodes are computers (computational and mathematical approaches). DeCartography adopts a complex system that determines the results of the oracle through the intuition and common sense of crowd workers. Unlike the mathematical nature of analysis in the former approach, DeCartography avoids the need for specific analytical indicators for social graph generation by utilizing crowdsourcing and peer prediction methods, thus avoiding the risk of "perverse incentives" where indicators can be hacked.
As a user of the oracle, if you have data that you want to analyze, you can submit the data to the DeCartography platform along with the required fee. After a certain period, the analyzed data will be returned. Currently in beta, the platform primarily analyzes data from projects that have contacted us.
As a crowd worker, you can earn cryptocurrency without the need for KYC by completing tasks assigned by DeCartography. By signing in to DeCartography using your wallet, completing tasks, and waiting for the specified period, rewards will be automatically transferred.
In this article, we will introduce the idea of the DeCartography project.
---
DeCartography is a governance technology for mapping, as the name suggests. It is a Dapp that utilizes crowdsourcing to generate a social graph that quantifies cooperation among contributors. There is no universal model for measuring and applying a social graph in a "pluralistic" manner. Instead, a diverse social mechanism based on its own algorithm can be built to evaluate and distribute rights and responsibilities to individuals by consulting a journal imbued with souls. This process could serve as an important foundation for pluralistic mechanism design that transcends social organizations and mechanisms, including democracy, markets, data economies, commons, and identity. Research and development of identity and policy modules for plurality could become a completely new business model for Gitcoin. For example, we may collaborate with diverse ecosystems in grant rounds and provide infrastructure for plurality for self-organizing communities beyond secondary funding. For example, civil resistance and digital authentication. For more details on these potential applications, please check out this article on cross-social data: . As mentioned by Leon Erich in the above link, DeCartography is planned to be used as an infrastructure for various Dapps as an identity and policy module for "plurality". The word "cartography" comes from the Greek words "chartis" (map) and "graphein" (to describe). DeCartography is a project that aims to represent the state of objective reality by adding levels of abstraction (plotting to matrices, dimension compression, clustering) to the initial state of objective reality (on-chain activity of each wallet) in a reliable manner.
Therefore, it is necessary to ensure reliable neutrality from the mechanism layer in this system. What does "reliable neutrality" mean in this system? Vitalik Buterin describes a good mechanism as follows: "A mechanism is a tool that takes input from multiple people and uses these inputs to determine the values of participants and make decisions that people care about." In a well-functioning mechanism, the decisions made by the mechanism are efficient, meaning they are the best possible results considering the preferences of the participants, and they align with incentives, meaning people have incentives to participate "honestly". To establish reliable neutrality in DeCartography, it is necessary to avoid writing specific individuals or specific results into the mechanism, ensure open and verifiable execution, keep it simple, and avoid frequent changes.
To adhere to these rules, DeCartography is built in a public space where everyone can refer to the following resources:
- All source code is publicly available on Etherscan and GitHub.
- The Cartography-Protocol, which includes task generation and reward calculation, is written in just 100 lines of code.
- All change history is publicly available, and updates are always announced.
Of course, the social graph data generated as the output of DeCartography is derived from the input of crowd workers as participants in the mechanism. However, it is expected to be very challenging for the core team to build while maintaining momentum, especially in the early stages, so it is planned to be built openly and gather opinions and criticisms through an iteration cycle called "Round". Additionally, to ensure reliable neutrality for crowd workers, DeCartography adopts the peer prediction method mentioned in Vitalik's article on Credible Neutrality. The peer prediction method refers to mechanisms for extracting information from human agents when it is not possible to directly verify the information obtained from them (when there is no Grand Truth). These mechanisms are designed to have a game-theoretic equilibrium where everyone has an incentive to honestly reveal their information.
---
DeCartography adopts a Proof of (Human) Work based on the Schelling Point and is a decentralized computation oracle. The mechanism of DeCartography is entirely dependent on the concept of the Schelling Point. The mechanism works as follows: imagine you and another prisoner are placed in separate rooms, and a guard gives each of you the same piece of paper with several numbers written on it. If both of you choose the same number, you will be released; otherwise, you will be imprisoned for life, as human rights are not particularly relevant in the game theory country. The numbers are as follows: 2, 5, 9, 25, 69, 73, 82, 96, 100, 126, 150. Which number will you choose? Theoretically, these are all arbitrary numbers, and the probability of choosing the same number and being released is 1/8. However, in reality, most people choose 100, so the probability is even higher. Why 100? Because each prisoner believes that the number 100 is "special" in some way, and each prisoner believes that the other prisoner also believes that the number 100 is "special". Therefore, each prisoner believes that the other prisoner is likely to choose 100, so they also choose 100. Clearly, this is a recursive chain of logic that goes on indefinitely, with no "backing" other than itself. The tasks that DeCartography assigns to crowd workers are based on this mechanism. Specifically, after displaying nine wallet addresses and the NFTs they hold, crowd workers are asked to select three that are "similar". After repeating this task 10 times (one session), they submit their answers with a signed message via a smart contract. After the round ends, all crowd workers' answers are aggregated, and correlations are calculated. Voting power is calculated based on the Humanity Score obtained from Gitcoin Passport and the amount staked before work. Based on the correlations and voting power, the proportion of rewards for each agent is calculated and transferred from the round's budget.
Why is this mechanism effective, even though it may seem like a puzzle? Essentially, it is for the same reason the prisoner's example above works. We assume that "looking similar" is a common sense Schelling Point. Everyone wants to provide their intuition, expecting that everyone else will provide the same answer, and the protocol encourages that.
Law enforcement officers have been using Schelling Points for centuries. They put prisoners in separate rooms and ask everyone about what happened in a certain event. It is easy for many people to agree if they tell the truth, but it is almost impossible to coordinate on specific lies, relying on the fact that it is difficult to coordinate on specific lies.
In short, what is a Schelling Point? The most important element is that it is special. What is the specialness here? Specialness, like beauty, lies in the eyes of the beholder. David Friedman writes the following: "Two people are given a list of numbers 2, 5, 9, 25, 69, 73, 82, 96, 100, 126, 150, and they are told that if they choose the same number individually, they will be rewarded. If both are mathematicians, they are likely to choose 2, which is the only even prime number. Non-mathematicians might choose 100, but for mathematicians, it is not as unique as the other two perfect squares. Illiterate people might choose 69 because of its peculiar symmetry. And people who are more interested in numbers than in mathematics might choose 69 for different reasons." ---
Why a "slow" oracle?
Some may question the need for an oracle to generate a social graph by clustering a list of addresses. It may seem that it takes much more time and resources to generate a social graph using crowdsourcing than using computers. The reason is that we believe there is value in analyzing data based on the intuition of crowd workers. Since all code executed on Ethereum is essentially public, if we have to write code somewhere to generate a social graph, it can be seen externally and can be avoided in reverse. For example, if we write code to "cluster based on Crypto Punks ownership," "cluster based on owning more than 100 NFTs," or "cluster based on Uniswap pools," the code for generating the social graph can be seen and avoided. By deciding on certain indicators for analysis with good intentions, people who benefit from manipulating the social graph (manipulating the social graph) only need to focus on hacking those indicators.
Therefore, we hypothesize that an oracle driven by people's intuition would be beneficial.
---
Issues and limitations
The vulnerability of this system lies in the manipulation of oracle results through collusion. If one entity controls more than 50% of the total votes, they can unilaterally set the median as they please. To mitigate the risks of Sybil attacks and collusion, we have implemented the following measures:
- We use Gitcoin Passport to prevent individuals from creating multiple accounts.
- Crowd workers can stake before starting a task. This mechanism is based on peer prediction and rewards increase based on the correlation between the staked amount and the answers. Conversely, if the correlation is too low, the staked amount may be slashed.
- We plan to mix the addresses participating in DeCartography as crowd workers into the analysis to calculate the social distance between crowd workers.
Bridging-based mechanisms like this help incorporate higher-quality answers and reduce the risk of a single community manipulating the results, as shown in academic research.
The rewards DeCartography pays for crowd workers' work are determined based on this "common sense". Accounts with higher "common sense" scores have a higher chance of receiving higher rewards. For more details, please refer to the reward determination mechanism.
---
The success of completely new ideas like DeCartography is not yet well understood, and it will require repeated experimentation to determine which versions work well and what rules yield good results in various contexts. We appreciate your feedback.