[Bitcoin-development] Proposed alternatives to the 20MB step function

[Ricardo Filipe]

I've been following the discussion of the block size limit and IMO it
is clear that any constant block size limit is, as many have said
before, just kicking the can down the road.
My problem with the dynamic lower limit solution based on past blocks
is that it doesn't account for usage spikes. I would like to propose
another dynamic lower limit scheme:
Let the block size limit be a function of the number of current
transactions in the mempool. This way, bitcoin usage regulates the
block size limit.

I'm sorry i don't have the knowledge of the code base or time to make
simulations on this kind of approach, but nevertheless I would like to
leave it here for discussion or foster other ideas.

cheers