That's already in the requirements:
(This sounds complicated, but there will probably be an easy-to-use UI for simple tasks such as subscribing to a single user's ignore list.)
Quote from: Requirements
Allow users to export/import lists of good posters, hidden posts, and ignored users through three separate webs of trust. Most people will want to use a single trust list for this, though, so there should also be an easy way of editing all trust lists at once.
Quote from: Requirements
About webs of trust
Each web of trust is composed of data, a separate trust list per user, and a trust limit setting per user from 0 to 100. A trust list is a list of users with associated trust scores from 0 to 100. Users see (by default) only the data of users in their trust networks. Trust networks are calculated from their trust lists.
The trust network algorithm should work something like this: To find a user's trust network, put everyone in his trust list with score above his trust limit into his trust network. Then get the trust network of everyone in his trust network. Combine all of these trust networks into a single list. Ensure that each user is listed only once -- average scores if a user appears more than once. Then for each user in this list, add them to the original user's trust network only if they do not already exist there. There should be some recursion limit.
Each web of trust is composed of data, a separate trust list per user, and a trust limit setting per user from 0 to 100. A trust list is a list of users with associated trust scores from 0 to 100. Users see (by default) only the data of users in their trust networks. Trust networks are calculated from their trust lists.
The trust network algorithm should work something like this: To find a user's trust network, put everyone in his trust list with score above his trust limit into his trust network. Then get the trust network of everyone in his trust network. Combine all of these trust networks into a single list. Ensure that each user is listed only once -- average scores if a user appears more than once. Then for each user in this list, add them to the original user's trust network only if they do not already exist there. There should be some recursion limit.
(This sounds complicated, but there will probably be an easy-to-use UI for simple tasks such as subscribing to a single user's ignore list.)
It does sound quite complicated, yes. Thanks for notifying me that this already exists, though.