I have a real concrete need as a blockchain engineer. I need to actually develop my programmes (smart contracts) and I need them to work fast. Also I would like the entire process to be free. Is that too much to ask in 2018? According to Ethereum, apparently it is.

Using steemit, I started to wonder if I should still use Ethereum for developing smart contracts. There is a cognitive dissonance when you click an upvote button and wait a second or two, while in Ethereum, once you fix all the "out of gas" errors you wait minutes just to call a function on a contract because the network is clogged.

So, since steem does not allow for smart contracts (if not in the form of plugins/side-chains) I wondered around a bit and found some other potential alternatives.

Name	Scalable	Fees	Live	Language	Notes
Ethereum	Nope	High	Yes	Solitidy	De facto standard for smart contracts
Lisk	Nah (2.5tps/10s)	Not free	Yes	Javascript	Still side chains :/
Steem	Yes! (10k tps/3s)	Free	Yes	C++	Sidechains :/
Cardano	Not currently	Not free	No	Haskell-like	Not much is known of the final architecture
Rootstock	Nah (100tps/10s)	Not free	No	Solidity
EOS	Yes! (50k tps/3s)	Free	No	WASM	Will it deliver?
Tezos	Probably not	Not free	No	Schema-like	Disastrous management
Qtum	Nah (70tps/2m)	~0.05€	Yes	Solidity
Neo	Maybe? (10k tps)	Free for now	Yes?	Many	Not sure if true or CHAYNA!
Bitshares	Yes	Low to free	Yes	C++	Side chains...

EOS

So by looking at the table it seems that the obvious choice is: just wait for EOS. Fair, we should be almost there, and the promises should be worth the wait. But what if I want a proven solution.

Neo

Neo transactions are free for the moment, but users could vote for nodes that actually take a fee. Moreover Neo is still in the making and the scalability is still not proven.

Special mention for Nano

Nano does not allow for smart contracts, still I wanted to include it here because I think it contains a technology that is even more scalable than graphene blockchain promise.

Nano blockchain is not a sequence but a partially ordered graph. This means that it is already sharded by user. Furthermore there is no block time, which would make it as quick as the propagation times itself.
The thing that I don’t understand about Nano is, while it’s actually simpler than Steem since it has smaller and simpler transactions, it actually has a slower throughput. Capped by I/O writes to the database, Nano can only do ~10k tps. Which seems a lot, but given the improved architecture I wonder if Nano has still to learn from the highly optimised Graphene engine.