Skip to content


Switch branches/tags

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Reputation Data Types | 'Likes'

For more Reputation Modules see this url.

Design is in keeping with principles of modularity that are discussed here.

Reputation Modules are designed as separate DHT's that can bridge into any existing app. This allows a community to pick which reputation module is appropriate for them and choose between specific features.

For example, a community can swap out a 'likes' module and instead decide to incorporate 'claps' or 'upvotes/downvotes' based on how they wish to articulate culture within their collective.

They do so by picking a DNA template, and adding their community's unique identifier in the 'properties' field to spin-off a new DHT.

Assuming they pick the 'likes' DNA, they will also have the choice of whether:

  • Likes received and given is private information i.e seen only by those who give, and those who receive.
  • Likes received and given is public info
  • Likes received and given cannot be seen by receiver.

Implementation as a Separate DNA

  • Any front-end module should be able to call this 'Like' DNA for its users to validate other users.
  • Likes are always issued by a user towards a 'base entry' that is created by another user. The 'base entry' is undefined by us. It could be a post, a photo, a blog, audio file etc.
  • The maximum likes a user can attach to a base entry is 1. (discuss)

Issuing Certifications, Badges, Stars etc.

The community can use the Likes DHT to aggregate the number of likes, and issue Reputation Scores (Badges/Labels/Certifications etc.) calculated on a specific DHT.

  • Mathematical operations possible on the number of likes: + - / * < >


Entry "like" {
    struct : Like {
        base: string,
        timestamp: u64,
        auth_address: Address,

Links {
  users -> Likes
  Base Entry -> Likes
Entry "base" {
    base : Unique universal identifier for the entry that is being commented on. 


Reputation Likes through mixin-zomes or callable DNAs







No releases published


No packages published