Pros

It auto records and calculate the creds for each contribution by merging GitHub PR, posting in forum, and chatting in Discord.

Cons

It is mainly automatic which oversimplifies the nuances and sophistications of values human created.

SourceCred ethnographic

1. The people (creators)

Team

Reference: Team at SourceCred and Medium and Twitter

2. Core product

1. 3 steps
   1. People contribute, contribution mint Cred
   2. Calculate/split Cred to contributors
   3. Contributors split Grain according to Cred%, epoch by epoch

1. Furthermore:

• Contributions may have a higher-level initiative. (oversee/break-down the contributions)
• Contribution may appreciate over time. (at beginning, it is underestimated)

1. Totally open source, run/host it yourself
2. References
   1. https://cred.sourcecred.io/#/explorer
   2. https://research.protocol.ai/blog/2020/sourcecred-an-introduction-to-calculating-cred-and-grain/
   3. https://medium.com/sourcecred
   4. Interview podcast: https://sourcecred.podbean.com/
   5. Real-world use: https://makerdao.sourcecred.io/#/explorer

◦ Graph的概念不好理解（即使用了spring-river-ponds的metaphor），edge和node都有weight，flow也容易引起misunderstanding（以为lose了）
◦ 还用了Markov chain的概念（预测只跟现状有关）
◦ 配置weight比较复杂
◦ 文档看了多个和好久才明白：设计者有点过度geek（tech和math）
◦ 连他们自己也觉得复杂后期想改造（CredEquate） |  |

| Graph：展示了contribution和人2类node，好处是你能全局看到贡献和人的关系 | 整个graph也是一个DB，设计比较复杂 | 只展示node是否更简单；然后贡献data可以解耦出来单独查询溯源？ | | 一个Time slot周期内，cred还会继续flow，早期有些贡献被低估后续会升值 | Cred好理解，代表network中贡献值/分值，but cred的自由flow逻辑不好理解（不同文档都有自相矛盾的地方）：

1. Flow=Cred从seed node mint，然后从某人flow out to 对应什么具体的contribution（可以积累看到具体的contribution）；
2. Flow不=lost，Cred在某人身上还是那个值X，每个周期根据这个分grain； | | | Grain split policies:

• Weekly (Fast incentive weekly)
• Balanced (Lifetime contribution → proportionate grain split)
• Recent (this week’s contribution gets higher weight than last week’s) | | Make the split flexible to meet diff use cases |

3. Concept/value proposition

1. Fuck capitalism, alt to it: capital on human over value creation