The flood of new members, little help please?

2»

Comments

  • Maybe another way is to offer campaign services, coordinate with providers to get waves of paying users. This can be done via community member leads. This would help strengthen the community and help plan out sales for providers

  • @ascicode said:
    Code sent through PM with/without a gate of microsoft forms asking for usernames, this getting merged with signup dates of les. All before signed up are qualified. All signups during an offer getting rejected. Possible to choose by every provider.

    A sign-up date gate can be automated:

    1. The offer post uses a tag to gate both sign up date and karma count.
    2. It renders as a discount token that is unique per qualified member, which cryptographically proves the member qualifies for the offer, without revealing the identity of the member.
    3. The provider billing software can either implement the corresponding validation procedure or call an API from forum side to verify the discount token is valid.

    If a member leaks their discount token, it can only be used once, so it’s like sharing a grocery coupon with a classmate.

    It is unnecessary to implement the heavy Zero Knowledge Proof structures.
    Instead, the discount token should include:

    1. The discussion ID of the offer, in plain text.
    2. The username, encrypted by an AES key only known by the forum. The initialization vector is constructed from the discussion ID and numeric user ID, passed through HKDF with a per-offer salt.
    3. A signature covering the above, signed by the forum’s private key.

    The signature is the forum proving to the provider that the member qualifies for the offer.
    The encrypted username allows the forum to know whose code it is, without revealing to the provider.
    The initialization vector construction method avoids per-offer per-member storage, but still ensures a member gets only one same discount token per offer, and the encrypted username portion is different across different offers to avoid correlation.

    … until @forest picks this apart and deanonymizes the algorithm.

    The last flan is taken. 🍮 affbrr

Sign In or Register to comment.