Skip to content

Milestones

List view

  • https://en.wikipedia.org/wiki/Concurrency_control Core should allow users to either take locks or use CaS or something, to enable transactional edits of their key-value store, fully or partially while scoping which ranges of keys are relevant to the transaction.

    No due date
  • Core should allow users to gate private data in two ways: 1. normal access control over namespaces of the keys, like Google Drive or S3 2. e2ee keys-values so even hosting provider can't read the data, the main idea here is to encode a binary tree in the keys, but that requires concurrency control first.

    No due date
  • To enable gossip and trustless CDNs/indexers, it is valuable to have all data self-authenticating, all the way from the user's Pkarr key, to the homeserver signer key, through a possible homeserver Pkarr key that is kept in the Authenticator. We should also allow new homeservers, to incorporate (bless) all snapshots signed by older homeservers... most likely using Bamboo, append-only Bao, or some other efficient append-only log of snapshots. This is most likely not viable for clients, but good for servers to receive data from untrusted sources, like in a gossip network.

    Overdue by 1 year(s)
    Due by October 1, 2024
  • Store user data in an append-only fashion (using seq number) to help with initial sync, and add a subscription API for streaming live changes. This milestone should help decouple pubky-app indexer from the homeserver, since it can subscribe to updates from the homeserver knowing nothing about its implementation.

    Overdue by 1 year(s)
    Due by July 4, 2024
  • A homeserver that has the basic features for account and session management, as well as an Authenticator app for Pkarr management, signing up to homeservers, and basic safe holding of keys.

    Overdue by 1 year(s)
    Due by July 1, 2024
    3/3 issues closed