Hacker News Reader

Comment by canoebuilder | Hacker News Reader

>but the industry being built around it is very real even if niche. There's a lot going on (including research)

What are some concrete examples?

I think most of us who are critical are still interested, because we do see things that are going on, and we ask ourselves, are we missing something?

But we seem to run into the same road blocks again and again.

“Solutions” seem to come from people subtly or transparently “talking their book,” that is, they have an undisclosed vested interest in you not thinking too much about those roadblocks.

What is some fruitful research being done in the areas of well known roadblocks? What are some application areas with a genuine pathway to viability and no better non-blockchain alternative?

View on HN

I think smart contracts are actually extremely novel in the form they take in Ethereum EVM, Tezos, Cardano, DAML and Pact (I'm author of the latter). TBH I'm a little frustrated with the evolution of smart contract langs post-2017, as I think Ethereum's dominance has made most engs/projects uninterested in innovating.

If you look around, it's actually very hard to find a genuinely small language you can embed in your runtime, and offer to users. Lua is a common suggestion, but Lua doesn't let you lock down IO, or code loading (that is, without writing your own interpreter and changing semantics -- what's the point?). A truly pure smart contract language gives the containing app full control over the entire runtime.

Smart contracts to me are a modern stored-procedure language that you embed directly into your app, and give you a 100% deterministic way to store long-term state, including code. "Application DBs" like sqlite don't scratch this itch (for instance sqlite doesn't have SPs). Plus, smart contracts are metered (usually), which allows for a cost-model and prevents against DOS. This plus the turing-incomplete angle makes them reasonable to open up to the world.

Erlang and lambdas come to mind, and I can argue why they're different too. But this is already too long :)

>But this is already too long :)

Not at all, I appreciate the info.

Cardano I believe uses PoS of some sort.

What is the thinking about what keeps these distributed application servers running? Is it still mainly the mining reward model?

What sort of applications most excite you? Are there places the distributed model works better than the traditional? Are there applications the distributed model enables that the traditional does not?

And vice versa, what are the limitations of the distributed model? What things just aren't feasible? Could you build Twitter or YouTube like systems?

> Cardano I believe uses PoS of some sort.

Yes, but smart contract languages (Plutus in Cardano's case) generally are unrelated to the specific consensus algorithm of the blockchain platform.

> What is the thinking about what keeps these distributed application servers running? Is it still mainly the mining reward model?

In general, public blockchain platforms run by incentivizing the operators within the system itself. PoS and PoW are different versions of this (staking v mining).

> Are there applications the distributed model enables that the traditional does not?

Crypto, and DeFi (decentralized finance), is conducting a huge "live beta" in decentralized corporate governance and finance. There is a crapton of innovation happening with financial engineering and exchanges, completely new models like Uniswap (constant product DEXs) or new versions of existing things, like on-chain options.

Is it world-changing? No, or not yet certainly. But by construction, DeFi can only happen on a decentralized smart-contract blockchain, so it's genuinely new.

> What sort of applications most excite you?

Outside of crypto, there are many areas within industries which require the sharing and use of data across multiple organizations, markets, and even intra-industry. Think of Healthcare and the multitude of data sources, processes, and interaction between parties. In areas like this, where multiple parties have many different roles and responsibilities in a shared business process, I think the opportunity for private blockchains is promising. I admit that these are niche use cases and the use of these type of applications has not taken off yet.

Do note that these types of applications are very different from the decentralized implementations seen in the crypto sphere. These are the private, permissioned models that people speak of.

In that case they don't need a blockchain, they just need a database.

Daml is a nice functional smart contract platform, with the ambition of smart contracts being really contracts. This is a few years old video about the original idea: https://vimeo.com/219723741, and this is their web site: https://daml.com. Today Daml is integrated with the Canton synch platform, and can connect several kinds of blockchains and databases: https://www.canton.io.

>What are some concrete examples?

I like how you can use a PoS blockchain to earn interest on your coins. Cardano can do this without moving your coins to another wallet, xlm can do it but you need to trust the other party. Currently there are enough people in the world without banking access that this could make a huge difference to them.

Reply on news.ycombinator.com