6 ms·

The next Ethereum? I'm still waiting for the current ethereum to show any kind of reason for its existence beyond an admittedly cool tech demo. I'm not seeing

by sho 9y ago

The next Ethereum? I'm still waiting for the current ethereum to show any kind of reason for its existence beyond an admittedly cool tech demo. I'm not seeing anything in this offering which couldn't be done more easily and cheaply by a ton (sorry) of other payment options. Notably they say they want to be WeChat - and they've done it all without any need for blockchains.

pavlov 9y ago
I would have to agree about Ethereum being not much more than a tech demo at present. A few months back I tried to learn building an Ethereum distributed app, or "dapp" as they call it. I installed my own Ethereum node to see how that works. The Ethereum toolchain is a strange combination of actual JavaScript (command line tools + REPL to a live node) and a pseudo-JavaScript called Solidity: a custom language inspired by JavaScript but with abundant new syntax for developing smart contracts. That doesn't seem like a good idea from a robustness point of view -- did anyone ever say "I know, I'd like my money to be handled by immutable JavaScript functions"?! Still, I can see the mindshare appeal of reaching the widest possible developer demographic by making everything look "web-like". I would be OK with that choice if the platform worked as advertised, but it doesn't, and it just keeps getting worse. I spent about $100 on deploying my toy dapp's smart contracts, and then I gave up. Debugging on Ethereum is a nightmare. It seems very easy for a non-obvious bug that's not caught on the testnet to make a deployed program go into a loop and burn through all the money you allocated for it. With today's USD/ETH rate, my failed $100 experiment would have cost $300. Even if I had got the app deployed, who's going to use it when fees are this high? Nobody wants to pay $10 to store a piece of data in my dinky little dapp. The whole concept of dapps is ridiculous with the current Ethereum implementation. Ethereum is a victim of its success: the network is congested, fees are running sky-high, and my Ethereum node has stopped working because of the exploding demands of storing the blockchain. Realistically there doesn't seem to be anything the Ethereum founders can do to fix this on the short term, which means that there will be a painful implosion as the network goes useless until it's replaced by a new sharded Ethereum (which will naturally come with its own growing pains). My conclusion is that Ethereum today is like MongoDB in its first year: breathless promises of "web scale" are in no way matched by the reality of what the platform can do... But like Mongo, it's in JavaScript and has a formidable hype machine, so people will flock in. Eventually things will get better, I'm sure.
- shp0ngle 9y ago
  I'm not sure why people latch to the javascript thing. Javascript is to Solidity what C is to Java. It is similar on the very surface level, but other than some syntax similarity, it's very different. There is plenty wrong with Javascript and there is way more wrong with Solidity. But the languages are not very similar
  - davidgrenier 9y ago
    I'm not sure the analogy holds. C is a very specific programming language you can characterise as a byte-array manipulation language. That's a very specific building block for making software if that building block is what you need, hence its merit. Java "on the very surface level" has absolutely nothing that relates to C, modulo some syntax if you think that has a lot of weight.
    - bigato 9y ago
      This is exactly the point he was making
- acover 9y ago
  Are most DApps silos? Is there a market for a direct clone of ethereum that is just less congested? Lots of apps just don't make sense at $3 per transaction. It's quite discouraging.
  - pavlov 9y ago
    I don't know if a clone would solve anything. My humble opinion is that the Ethereum design is just fundamentally flawed. It doesn't do any specific thing very well at all, but it looks enough like Bitcoin to encourage speculators. As a product, it's like: "The gold coin that also works as toothpaste!"
- cocktailpeanuts 9y ago
  I agree that the ecosystems is very shitty and amateur. Just look at web3.js--they pushed the alpha version to NPM so anyone getting started will automatically get their alpha. I wouldn't say anything if this alpha was working, but it's one of the buggiest piece of software I've ever seen in a while. They shouldn't ship something if it's so obviously not ready, especially if it has to do with people's money. I actually tried so hard to make it work with this alpha version but finally gave up after a while and manually "downgraded" to the earlier version, because my time is precious and I don't have time to help them develop that buggy software by sending pull requests for every single bug I've found. And I don't even want to get started with Solidity. Everyone in the industry knows Solidity is a bad language which encourages unsafe and careless programming. That said, I think it's a wrong way to think about Ethereum from the perspective of building a normal web app. A lot of people confuse this point and think they're building "Twitter, but for crypto", "AirBnb, but for crypto", etc. but it's a whole new class of software, and the faster you drop the old frame of "what web programming is", the faster you'll adapt and more likely that you'll come up with neat ideas to build. The gist is, you have to think like an economist than a programmer. Lastly, you could just use testnet to test your stuff. I've been mostly running my demo apps on testnet and only deploy when I think there's a chance for any traction. Which means you don't need to spend any money to try building these apps.

by sho 9y ago

pavlov 9y ago

I would have to agree about Ethereum being not much more than a tech demo at present. A few months back I tried to learn building an Ethereum distributed app, or "dapp" as they call it. I installed my own Ethereum node to see how that works. The Ethereum toolchain is a strange combination of actual JavaScript (command line tools + REPL to a live node) and a pseudo-JavaScript called Solidity: a custom language inspired by JavaScript but with abundant new syntax for developing smart contracts. That doesn't seem like a good idea from a robustness point of view -- did anyone ever say "I know, I'd like my money to be handled by immutable JavaScript functions"?! Still, I can see the mindshare appeal of reaching the widest possible developer demographic by making everything look "web-like". I would be OK with that choice if the platform worked as advertised, but it doesn't, and it just keeps getting worse. I spent about $100 on deploying my toy dapp's smart contracts, and then I gave up. Debugging on Ethereum is a nightmare. It seems very easy for a non-obvious bug that's not caught on the testnet to make a deployed program go into a loop and burn through all the money you allocated for it. With today's USD/ETH rate, my failed $100 experiment would have cost $300. Even if I had got the app deployed, who's going to use it when fees are this high? Nobody wants to pay $10 to store a piece of data in my dinky little dapp. The whole concept of dapps is ridiculous with the current Ethereum implementation. Ethereum is a victim of its success: the network is congested, fees are running sky-high, and my Ethereum node has stopped working because of the exploding demands of storing the blockchain. Realistically there doesn't seem to be anything the Ethereum founders can do to fix this on the short term, which means that there will be a painful implosion as the network goes useless until it's replaced by a new sharded Ethereum (which will naturally come with its own growing pains). My conclusion is that Ethereum today is like MongoDB in its first year: breathless promises of "web scale" are in no way matched by the reality of what the platform can do... But like Mongo, it's in JavaScript and has a formidable hype machine, so people will flock in. Eventually things will get better, I'm sure.

shp0ngle 9y ago
I'm not sure why people latch to the javascript thing. Javascript is to Solidity what C is to Java. It is similar on the very surface level, but other than some syntax similarity, it's very different. There is plenty wrong with Javascript and there is way more wrong with Solidity. But the languages are not very similar
- davidgrenier 9y ago
  I'm not sure the analogy holds. C is a very specific programming language you can characterise as a byte-array manipulation language. That's a very specific building block for making software if that building block is what you need, hence its merit. Java "on the very surface level" has absolutely nothing that relates to C, modulo some syntax if you think that has a lot of weight.
  - bigato 9y ago
    This is exactly the point he was making
acover 9y ago
Are most DApps silos? Is there a market for a direct clone of ethereum that is just less congested? Lots of apps just don't make sense at $3 per transaction. It's quite discouraging.
- pavlov 9y ago
  I don't know if a clone would solve anything. My humble opinion is that the Ethereum design is just fundamentally flawed. It doesn't do any specific thing very well at all, but it looks enough like Bitcoin to encourage speculators. As a product, it's like: "The gold coin that also works as toothpaste!"
cocktailpeanuts 9y ago
I agree that the ecosystems is very shitty and amateur. Just look at web3.js--they pushed the alpha version to NPM so anyone getting started will automatically get their alpha. I wouldn't say anything if this alpha was working, but it's one of the buggiest piece of software I've ever seen in a while. They shouldn't ship something if it's so obviously not ready, especially if it has to do with people's money. I actually tried so hard to make it work with this alpha version but finally gave up after a while and manually "downgraded" to the earlier version, because my time is precious and I don't have time to help them develop that buggy software by sending pull requests for every single bug I've found. And I don't even want to get started with Solidity. Everyone in the industry knows Solidity is a bad language which encourages unsafe and careless programming. That said, I think it's a wrong way to think about Ethereum from the perspective of building a normal web app. A lot of people confuse this point and think they're building "Twitter, but for crypto", "AirBnb, but for crypto", etc. but it's a whole new class of software, and the faster you drop the old frame of "what web programming is", the faster you'll adapt and more likely that you'll come up with neat ideas to build. The gist is, you have to think like an economist than a programmer. Lastly, you could just use testnet to test your stuff. I've been mostly running my demo apps on testnet and only deploy when I think there's a chance for any traction. Which means you don't need to spend any money to try building these apps.