4 ms·

Blockchains just are not a particularly good datastructure for most applications. In cryptospace you will easily find projects where the creators took an idea

Blockchains just are not a particularly good datastructure for most applications. In cryptospace you will easily find projects where the creators took an idea and inserted a blockchain instead of a traditional database. Web3 is the same idea. But this idea is obviously really bad, blockchains are an extremely niche datastructure. They are write only and established by consensus. Most databases shouldn't have these properties and forcing them in will lead into a giant mess.

cmrdporcupine 4y ago
And now... we've also hit the ricochet point where now people are taking things backed by a regular centralized database and using blockchain marketing terms for it. Because that's apparently where the $$ is. I'm crying too :-)
chrisco255 4y ago
They are not write only. Ethereum is a VM. Smart contracts are not databases, they are objects, classes, methods, and applications. They can be thought of as in memory applications or long lived processes.
- constantcrying 4y ago
  >They are not write only. I guess you can delete data from a blockchain by having a rollback. The etherium blockchain also is much more than a "block chain". But it is still very much a solution in search of a usecase. Digital money is the only area where a blockchain is a sensible choice from a technical point of view, that I know of. Although I am entirely unconvinced that bitcoin or etherium actually are even close to a viable interpretation of digital money. Say what you want about paypal, it isn't distributing my payment details publically.
  - chrisco255 4y ago
    Multisigs, permissionless lending, decentralized data storage, automated market makers, social protocols, exchangeable token based authentication, digital content ownership, digital rights management, etc. All of these things already exist on the platform. Feel free to do some research if you have an open mind: Gnosis Safe: https://gnosis-safe.io/ https://gnosis-safe.io/ Uniswap: https://app.uniswap.org/ https://app.uniswap.org/ Lens: https://www.lens.xyz/ https://www.lens.xyz/ Lit: https://litprotocol.com/ https://litprotocol.com/ GMX: https://gmx.io/#/ https://gmx.io/#/ Filecoin: https://filecoin.io/ https://filecoin.io/ Arweave: https://www.arweave.org/ https://www.arweave.org/ Aave: https://aave.com/ https://aave.com/ USDC: https://www.circle.com/en/usdc https://www.circle.com/en/usdc There's over $100B in USD backed currency alone on the Ethereum blockchain. It's programmable, liquid, exchangeble 24/7, interacts with thousands of smart contracts, and increasingly will be used for day to day commerce in addition to financial investment applications. You are pseudo-anonymous on Ethereum. It's free to spin up new accounts and you can always fund them from a centralized exchange if you want Paypal-like privacy. Even then, there are solutions in the works for privacy: https://aztec.network/ https://aztec.network/ Cheers.
    - gerbler 4y ago
      >Aztec is on the cutting edge of zero-knowledge SNARK development. Built on PLONK — the paradigm-defining universal zk-SNARK used by leading zero-knowledge projects like Mina, Dusk, Zcash, and zkSync. I have an open mind about this area and a technical background. I really struggle to understand what any of this means and what I can use it for. My experience with cryptocurrency (paying for things) has been challenging. As the whole ecosystem is very fragmented it's hard to know what to explore.
      - chrisco255 4y ago
        Zk-Snark means Zero-Knowledge Succinct Non-Interactive Argument of Knowledge. It's fairly cutting edge cryptography. It is very fragmented as the industry is rapidly evolving. If I may recommend, a good article on zk Snarks from Vitalik: https://vitalik.eth.limo/general/2021/01/26/snarks.html https://vitalik.eth.limo/general/2021/01/26/snarks.html ZK proofs allow you to prove you know the answer to something without revealing the answer itself. You may not need to know how these things work at a low level. If you're interested, feel free to dig into zk proofs and the cryptography behind them. The subject is quite deep and rapidly evolving. Ultimately, the important thing to know at a high level is that you can post a zk proof with just a few bytes of data, and that zk proofs, while computationally expensive to generate, are computationally cheap to verify. So they are being used as a blockchain scaling technology, since data storage is expensive on blockchains. If you're just interested in crypto as a user, I would generally recommend getting a Metamask wallet and interacting with some applications like Uniswap, Aave, or OpenSea.
    - constantcrying 4y ago
      >Multisigs, permissionless lending, decentralized data storage, automated market makers, social protocols, exchangeable token based authentication, digital content ownership, digital rights management, etc. All of these things already exist on the platform. And all of these either shouldn't exist or can be done better without a blockchain. It is a solution without a problem.
      - ajconway 4y ago
- nl 4y ago
  I think by "write only" they meant immutable. Which is true.
  - can16358p 4y ago
    Perhaps "append-only" might be the appropriate term there.
    - nl 4y ago
      Yes this is also correct.
      - chrisco255 4y ago
        Smart contracts themselves are mutable, though. They have their own memory allocation and data storage in the VM. While the transaction log itself may be immutable, there is still mutable state at the application level.
dcolkitt 4y ago
> Blockchains just are not a particularly good datastructure for most applications. Very few applications are bottlenecked by the efficiency of their data structure. Computing is filled with examples where we make significant performance tradeoffs to achieve other goals. E.g. software applications are orders of magnitude less efficient than dedicated hardware, yet we accept that tradeoff because software is far more flexible and faster to develop. Similarly decentralized blockchains may be orders of magnitude less efficient than centralized databases, but for many applications that simply doesn't matter. The most scalable chains cost $0.0001 per transaction and can process tens of thousands transactions per second. Yes, a very optimized database may only cost $0.00000001 per transaction and can process millions of TPS. But the fact is none of this matters to your typical CRUD app. What you do get for that tradeoff is much better guarantees around durability and reliability, observability, transparency, censorship resistance, decentralization, and a much more credible commitment to the users that the app will not stop operating, delete their data, or change its behavior in unexpected ways. That doesn't mean the tradeoff is always worth it. The most performance critical systems will still probably run in centralized databases, just as the most performance critical applications run on custom ASICs instead of software. (Even then, we'd probably still expect databases to commit frequently commit their Merkle root to decentralized blockchains for trust, transparency, trust, auditability) But we know that blockchain technology continues to improve at exponential Moore-like rates (in the past ten years, blockchain capacity has doubled about every 24 months). We'd therefore expect the performance tradeoffs vis-a-vis centralized databases to become increasingly de minims, whereas the positive characteristics of decentralized chains should only improve as the general tooling and technology gets better.
- constantcrying 4y ago
  That is not my really my point. It isn't just about a speed, but that the blockchain doesn't solve a problem. In most situations you have a central source of authority and you want that central source of authority to be able to manipulate the database. It is a feature. The applications where a blockchain is more useful than a traditional database have yet to be found. Money might be one such thing, but the Etherium/Bitcoin implementation is pretty bad.

cmrdporcupine 4y ago

And now... we've also hit the ricochet point where now people are taking things backed by a regular centralized database and using blockchain marketing terms for it. Because that's apparently where the $$ is. I'm crying too :-)

chrisco255 4y ago

They are not write only. Ethereum is a VM. Smart contracts are not databases, they are objects, classes, methods, and applications. They can be thought of as in memory applications or long lived processes.

constantcrying 4y ago
>They are not write only. I guess you can delete data from a blockchain by having a rollback. The etherium blockchain also is much more than a "block chain". But it is still very much a solution in search of a usecase. Digital money is the only area where a blockchain is a sensible choice from a technical point of view, that I know of. Although I am entirely unconvinced that bitcoin or etherium actually are even close to a viable interpretation of digital money. Say what you want about paypal, it isn't distributing my payment details publically.
- chrisco255 4y ago
  Multisigs, permissionless lending, decentralized data storage, automated market makers, social protocols, exchangeable token based authentication, digital content ownership, digital rights management, etc. All of these things already exist on the platform. Feel free to do some research if you have an open mind: Gnosis Safe: https://gnosis-safe.io/ https://gnosis-safe.io/ Uniswap: https://app.uniswap.org/ https://app.uniswap.org/ Lens: https://www.lens.xyz/ https://www.lens.xyz/ Lit: https://litprotocol.com/ https://litprotocol.com/ GMX: https://gmx.io/#/ https://gmx.io/#/ Filecoin: https://filecoin.io/ https://filecoin.io/ Arweave: https://www.arweave.org/ https://www.arweave.org/ Aave: https://aave.com/ https://aave.com/ USDC: https://www.circle.com/en/usdc https://www.circle.com/en/usdc There's over $100B in USD backed currency alone on the Ethereum blockchain. It's programmable, liquid, exchangeble 24/7, interacts with thousands of smart contracts, and increasingly will be used for day to day commerce in addition to financial investment applications. You are pseudo-anonymous on Ethereum. It's free to spin up new accounts and you can always fund them from a centralized exchange if you want Paypal-like privacy. Even then, there are solutions in the works for privacy: https://aztec.network/ https://aztec.network/ Cheers.
  - gerbler 4y ago
    >Aztec is on the cutting edge of zero-knowledge SNARK development. Built on PLONK — the paradigm-defining universal zk-SNARK used by leading zero-knowledge projects like Mina, Dusk, Zcash, and zkSync. I have an open mind about this area and a technical background. I really struggle to understand what any of this means and what I can use it for. My experience with cryptocurrency (paying for things) has been challenging. As the whole ecosystem is very fragmented it's hard to know what to explore.
    - chrisco255 4y ago
      Zk-Snark means Zero-Knowledge Succinct Non-Interactive Argument of Knowledge. It's fairly cutting edge cryptography. It is very fragmented as the industry is rapidly evolving. If I may recommend, a good article on zk Snarks from Vitalik: https://vitalik.eth.limo/general/2021/01/26/snarks.html https://vitalik.eth.limo/general/2021/01/26/snarks.html ZK proofs allow you to prove you know the answer to something without revealing the answer itself. You may not need to know how these things work at a low level. If you're interested, feel free to dig into zk proofs and the cryptography behind them. The subject is quite deep and rapidly evolving. Ultimately, the important thing to know at a high level is that you can post a zk proof with just a few bytes of data, and that zk proofs, while computationally expensive to generate, are computationally cheap to verify. So they are being used as a blockchain scaling technology, since data storage is expensive on blockchains. If you're just interested in crypto as a user, I would generally recommend getting a Metamask wallet and interacting with some applications like Uniswap, Aave, or OpenSea.
  - constantcrying 4y ago
    >Multisigs, permissionless lending, decentralized data storage, automated market makers, social protocols, exchangeable token based authentication, digital content ownership, digital rights management, etc. All of these things already exist on the platform. And all of these either shouldn't exist or can be done better without a blockchain. It is a solution without a problem.
    - ajconway 4y ago
nl 4y ago
I think by "write only" they meant immutable. Which is true.
- can16358p 4y ago
  Perhaps "append-only" might be the appropriate term there.
  - nl 4y ago
    Yes this is also correct.
    - chrisco255 4y ago
      Smart contracts themselves are mutable, though. They have their own memory allocation and data storage in the VM. While the transaction log itself may be immutable, there is still mutable state at the application level.

dcolkitt 4y ago

> Blockchains just are not a particularly good datastructure for most applications. Very few applications are bottlenecked by the efficiency of their data structure. Computing is filled with examples where we make significant performance tradeoffs to achieve other goals. E.g. software applications are orders of magnitude less efficient than dedicated hardware, yet we accept that tradeoff because software is far more flexible and faster to develop. Similarly decentralized blockchains may be orders of magnitude less efficient than centralized databases, but for many applications that simply doesn't matter. The most scalable chains cost $0.0001 per transaction and can process tens of thousands transactions per second. Yes, a very optimized database may only cost $0.00000001 per transaction and can process millions of TPS. But the fact is none of this matters to your typical CRUD app. What you do get for that tradeoff is much better guarantees around durability and reliability, observability, transparency, censorship resistance, decentralization, and a much more credible commitment to the users that the app will not stop operating, delete their data, or change its behavior in unexpected ways. That doesn't mean the tradeoff is always worth it. The most performance critical systems will still probably run in centralized databases, just as the most performance critical applications run on custom ASICs instead of software. (Even then, we'd probably still expect databases to commit frequently commit their Merkle root to decentralized blockchains for trust, transparency, trust, auditability) But we know that blockchain technology continues to improve at exponential Moore-like rates (in the past ten years, blockchain capacity has doubled about every 24 months). We'd therefore expect the performance tradeoffs vis-a-vis centralized databases to become increasingly de minims, whereas the positive characteristics of decentralized chains should only improve as the general tooling and technology gets better.

constantcrying 4y ago
That is not my really my point. It isn't just about a speed, but that the blockchain doesn't solve a problem. In most situations you have a central source of authority and you want that central source of authority to be able to manipulate the database. It is a feature. The applications where a blockchain is more useful than a traditional database have yet to be found. Money might be one such thing, but the Etherium/Bitcoin implementation is pretty bad.