I just did a pre release (v0.1.1-alpha) for pretty good apps. Desktop (currently mac and linux, but win forthcoming).

Anyone can make a list of hardware wallets or veggies or whatevs. Anyone can add items to any list. Anyone can endorse / reject any item on any list, and anyone can endorse / block any user as a trusted curator of that list.

For testnet, I’m using kind 9901 for lists and items; endorsements are kind 39901, parameterized replaceable events. Your app downloads data, crunches some numbers and says “here is your crowdsourced list of Awesome Hardware Wallets: A, B, C …”

Curator trust is contextual, transitive and calculated in a PageRank-like fashion, so if Alice trusts Bob to curate wallets, Bob trusts you, and you trust Marty, then Marty will have influence over Alice’s list even if she doesn’t know who he is. Influence of unvetted users is set to a default which is usually low or zero to screen out sybil attacks, but can be adjusted if you’re not currently experiencing sybil attacks and want a more permissive system so you can get more input from more sources. There are several other parameters you can play with; as you adjust them, you can watch items pop on and off the list and hopefully get an intuitive feel for how the parameters work.

I welcome any feedback that anyone may have! 🙂