A Fast, Scalable Protocol For Resolving Lightning Payments

I should note that this is a change. The protocol as written forbids a cooperative close and return of base funds if any HTLCs were disputed.

(Despite forbidding a cooperative close, there was no requirement to close the channel unilaterally after a disputed HTLC, rather than continuing to add new HTLCs. Sorry if this wasn’t clear.)

The intent was to have at least a minimal self-griefing cost (the base funds), but such a cost should be per-HTLC, if it is needed at all.

I’ll update the paper to reflect this and any other changes that come out of this discussion…