When discussing web of trust I think it’s worthwhile discussing whether we agree (or not) with the following two proposals, in the abstract:
1. The goal is not to produce one curated result for one community. It’s to produce one curated result per user. If Alice and Bob don’t have identical WoT, they should see different results.
2. It is generally preferable to use explicit attestations rather than scraped data. (Don’t assume Alice “trusts” Bob just bc she follows him and it’s easy to scrape a follows list. She should explicitly attest “I endorse Bob in context X.”)
I have written these up as DIP-01 and DIP-02, respectively, which are linked to on this page: