search

autonomys / auto-mobile

Auto Typescript SDK
MIT License
0 stars 2 forks source link

Naming of AutoID #24

Open its-colby opened 8 months ago

its-colby commented 8 months ago

https://github.com/subspace/auto-mobile/blob/115a23df284065e00343877fc7b3c149e473df68/packages/sdk/src/library/wallet.ts#L76

AutoID should be standard asymmetric key-pair (ex: via elliptic curve), not Semaphore key-pair. Semaphore key-pair should be separate field in AutoWallet interface, such as “AutoGroupKey”.

@chrissotraidis

abhi3700 commented 8 months ago

I personally think that would lead to a different architecture.

The Semaphore based projects have taken the route similar to what has been currently implemented, I firmly believe.

its-colby commented 8 months ago

I don't see how this would lead to a different architecture. All it implies is...everytime you make a call for the AutoID in the code now, you would be making it for the AutoGroupKey.

abhi3700 commented 8 months ago

I think previously Auto ID was commitment, but then due to some reason (don't remember now), I switched to making secret (trapdoor + nullifier) as the Auto ID.

Now, I think commitment can be made as the Auto ID.

AutoID should be standard asymmetric key-pair (ex: via elliptic curve), not Semaphore key-pair. In my understanding, Auto ID was supposed to be a unique ID (like long number) attached/not attached to some wallet seed phrase. Now, if I make the Auto ID as asymmetric keypair, then I have few concerns:

  • Are we going to store the keypair private key or seed somewhere? If yes, what's the plan? We are already having the seed phrase that is stored locally via Shamir's secret sharing scheme as of now.
  • With Auto ID as asymmetric keypair, are you thinking of signing something? If yes, then it's already taken care by having private keys (determined from the recovered seed phrase) for respective chains (consensus + domains).