Compare Babe Spec and Babe Implementation with Babe paper

[drskalman]

We need to compare babe implementation with babe spec and submit any discrepency to research-security-issues repo.

[drskalman]

Currently the more important part is the relative time protocol part of it because the slot selection part is going to be replaced by the SASS.

Babe research write up can be found here:

https://github.com/w3f/research/blob/master/docs/polkadot/BABE/Babe.md
But check with @hndnklnc for the latest write up.

[hndnklnc]

It is the latest writeup. If you have any questions, you can ask me.

[FlorianFranzen]

I am updating our spec to the latest write up right now. Current progress it tracked in w3f/research-internal#110

[drskalman]

@FlorianFranzen is going to the current host spec according to the latest babe spec with recent constant hardcoded.

[drskalman]

The Babe spec is being followed up in issue #168
And the descrepancy between the rust implementation and the Spec is discussed in https://github.com/w3f/research-security-issues/issues/7

[burdges]

I found the randomness was not being recycled correctly, to which wei proposed a fix in paritytech/substrate#5788 although they'll compare with another fix approach.

We should figure out if the randomness recycles from one epoch waits long enough before being used in another epoch.

[burdges]

Fix merged paritytech/substrate#5876

I donno if they merged the whole migration machinery yet though.