All rounding would presumably be rounded down until they reach 0.0000,0001.
If a hub's fees do not exceed 0.0000,0000,9999, it would always round down to zero.
The total accumulated micro fees will only be important at the settlement unto the bitcoin blockchain.
For example, total hub fees for the day:
(1) 0.0012,7187,4912 recorded in LN = 0.0012,7187 settled to bitcoin chain
(2) 0.0012,7187,9999 recorded in LN = 0.0012,7187 settled to bitcoin chain
(3) 0.0012,7188,0000 recorded in LN = 0.0012,7188 settled to bitcoin chain
You are implying fractional reserve banking that can be applied to the bitcoin blockchain?
That would violate the agreed 21 million coin cap limit, so all rounding must go down.
exactly. your getting it..
however i can see hubs not agreeing to lose out on earning by rounding down. which is (before carlton butted in) the point i was making.
that LN has many flaws and needs to address many things to fix many bugs, remove many opportunities of abuse, etc etc
i think using millisats is opening a big can of worms of possible issues, compared to just handling satoshis, to emulate bitcoin and allow signing and settling to happen more fluid
I think the LN system can be designed so that the hub operator would not be able to
determine their own rounding aspect. If the hub operator is concerned with losing out
on a whole satoshi, then at one of the last LN tx they receive prior to the "settlement"
they can jack their fee up a few millisats so that it evens out to a full satoshi. The
"LN fee rise" will still be in accordance with the previously approved hub fee boundaries,
that I outlined in a prior posting.
LN is still an experiment and in development, but I believe that we can get it to work as
envisioned if we work together and create "checks and balances" within "checks and balances".
LN allows us to protect Bitcoin from failure, by outsourcing that potential failure to a secondary layer.