Closed nicola closed 5 years ago
@nicola Have we concluded that this is the strategy we will take? If we have decided that, then I think this resolves #157. If we have concluded this, but #157 still requires more investigation into the options enumerated above, can you comment and/or update the issue there to reflect where we are at on the question?
This does not resolve #157 this is one strategy to investigate and step number 1 for both aggregation or recursive composition. We are no further of where we left before, all updates are written on github
Got it. Thanks.
Hi all,
just talked with Matteo who I knew had to deal with Snarks composition [and then with the problems of sampling the different curves].
He told me he actually did not do it in the end, but talking with Rihad he was pointed to this paper: https://www.shiftleft.org/papers/decaf/decaf.pdf
technical name for this in geppetto is: bounded proof bootstrapping
Cocks-Pinch method is the one we should use to find the right curve. Luckily Scippr-lab has an implementation here: https://github.com/scipr-lab/ecfactory/tree/master/ecfactory/cocks_pinch
I am a bit unclear on what the inputs to the Cocks-Pinch algorithm should be:
Zcash is thinking about his too! https://github.com/zcash/zcash/issues/3425#issuecomment-428011202
r = 52435875175126190479447740508185965837690552500527637822603658699938581184513
closing this, @nicola please open in a research repo if this still needs work
From Ben: "aggregation just requires setting the field of the second snark to be the order of the curve in the first"
Next steps are: