Author: Andrew Miller 2015-11-24 13:31:37
Published on: 2015-11-24T13:31:37+00:00
The author of the email is responding to a previous message about SNARKs and their application. They note that while SNARKs may not be necessary, it's impressive that the sender tried them out. The author then goes on to question some performance estimates regarding snarklib/snarkfront, stating that checking a single snark proof should take around 10 milliseconds instead of 500 milliseconds and should only require about a kilobyte of memory instead of 150MB. Additionally, the verification key should be very small, in the kilobyte range. The author intends to reproduce the experiment using libsnark later that week and mentions that there are students at UMD with a tool for composing circuits that complements libsnark and is hoping they release it soon. The email then shifts to a separate discussion about the segregated witness proposal and new opcodes in SW script. The proposal suggests that the first byte of SW script is a version byte and if you don't understand that version, the script succeeds. This would make adding ops that change the contents of the stack a soft fork instead of a hard fork. The author doesn't think this is necessary for their current discussion.
Updated on: 2023-05-23T21:29:02.057920+00:00