Ethereum should have a dataflow programming smart contact language, Ziliqa style.
I added another paper: Snowflake to Avalanche: A Novel Metastable Consensus Protocol Family for
RapidChain: A Fast Blockchain Protocol via Full Sharding
is also an interesting idea posted on IACR’s eprint archive.
I haven’t read all of it.
To verify cross-shard transactions, RapidChain’s committees discover each other via an efficient routing mechanism inspired by Kademlia  that incurs only a logarithmic (in number of committees) latency and storage. In contrast, the committee discovery in existing solutions [45, 40] requires several “gossip-to-all” invocations
We’re already working on this with gossipsub
Third, to ensure the members of C out agree on the same block, they participate in a Byzantine consensus protocol that we construct based on the synchronous protocol of Ren et al. .
It’s a synchronous protocol for committees, and partially synchronous for the rest of the protocol.
The paper is also in a UTXO context, which makes the whole protocol substantially different.
A node who wishes to participate in the protocol for the next epoch will be admitted to the system only if it can solve a fresh puzzle before the next reconfiguration event.
It’s also PoW.
Yep, definitely a different design space, but it’s interesting to come across a paper focusing on this from a UTXO standpoint.