What i find concerning is that offchain transactions are not really going to help with a lot of the strange smart contract or internal tokens. It sounds more like a fantasy that you will be able to make the same tech that is far, far ahead on the Bitcoin and Litcoin projects fit into your world but you have completely different designs. This is something suddenly appearing — almost as an admission that “sharding is too hard” and “we can do payment channels too, like Bitcoin”. So the innovations are way ahead in Bitcoin and not Ethereum where smart contracts are still really simplistic and mostly just for the crappy ERC20 token (cough ponzi). This took a really long time with a fairly simple design to find traction — so really this is a kind of Ethereum maximalist wet dream indicating that you’re way behind on scaling but you’ve been and continue to spout future transaction throughput predictions as if you can do them now.
Saying sharding would be exciting if i had any expectation to really see it and I’ve heard talk that it actually has quite some limits to growth in itself. Whereas we know that offchain, multi-blockchain structures can handle millions of transactions.
Many of the “scaling” ideas listed in that sheet aren’t going to really improve the throughput very much at all or they are just “this is terrible” we need to replace it. Parallel process transactions is non-trivial because you need to make sure state changes don’t form contention — this is something you design in a CPU with all kinds of ultra cool pipeline look ahead, shared memory … at the protocol level its orders of magnitude more likely to yield years of unexpect bugs and then you have to go back to the drawing board when other changes to the protocol happen. So that one i wouldn’t expect to be seeing any time soon. Parallelism tends not to increase power to process — because ultimately computing is only one thing at a time and processing is on hold while the computer jumps to another task and then returns. Only certain things work in parallel — so unless its sharding and different nodes doing things in parallel you don’t get any real processing power upgrade.
The conclusion “tokenize everything” — we could have been doing that well before cryptocurrency came along. We have had huge structure for frequent flyer miles and rewards programs. A lot of these ICOs are barely even at that level of usefulness — we’ll make some rewards tokens … Sure some have much more complex and interesting ideas like, sell computer time (already existed), sell resources like electricity … but to me those will work better if people can interchange tokens and cryptocurrency from multiple blockchains because many will be weary of centralizing into “the one true blockchain”. A good developer would make a blockchain/crypto solution that was tuned up to be specialized for an industry, not use a kind of broad prototype platform like ethereum which is trying to be all things to all people. That always ends up badly, that is typical captain Ahab going down with the ship, my project can do everything IT epic failures that more experienced developers distance themselves from.
Tokenize everything has little value — it doesn’t match real problem/solution analysis. What we have is a problem, i want to get rich making a token and call it the solution, inappropriate design system. Mostly lead by suits and not developers. Yes I have a Computer Science background and i didn’t go and work for a scummy bank and don’t write convenient “believe it” analysis to help pump the price when it is flagging. What about some talk about the test net working and its going to be this long before we can go live … that is what would be beyond pie in the sky vapourware.
Don’t forget to through DAO and governance in there … those are genuinely interesting but I think its clear governance is about people with money, investment and pulling strings and not a for the benefit of the public when it comes to Ethereum.