Opening this article, I was hoping to get a technical discussion as to why an application might or might not need blockchain. Unfortunately, I was met with a ranting article with a few straw man arguments about why blockchain is flawed.
The first point the author makes is the disadvantage of immutability. Claims that few systems benefit from immutable transaction records. First of all, I don't think that's true. Many systems do, not just financial systems with double entry bookkeeping. Event sourcing is a well defined architecture pattern in back end systems, and it turns out that immutable event logs are actually incredibly useful. It's also pretty well known that you don't erase history in financial systems. That is one of the pillars of double entry bookkeeping.
A more balanced article would point out the pros and cons of immutable systems, but to flippantly argue that immutability is useless? That's just lazy.
Tracking shipment of goods on a blockchain has many advantages. Namely, you can create a shared protocol for the data of shipment tracking, and you can open that blockchain up to certain parties to sign as the good moves along the supply chain. It doesn't have to be "trustless". A particular app's blockchain might only be open to a few trusted parties to write to. Not all transactions have to be written to a single, global, immutable ledger. There are still great benefits to sharing a common blockchain among specific, trusted, cooperative entities.
On POW and the electricity argument, I think we waste electricity on lots of frivolous things. I'm willing to bet Christmas lights also consume more than the state of Delaware. However, it's not the processing of transactions that the electricity is being spent on, it's the creation of new coins.
POS is a good alternative. I think many coins are starting to leverage POS. Ethereum will be switching to that soon.
Still, POS vs POW has nothing to do with "why you probably don't need a blockchain". These are implementation details that vary per blockchain/coin.
Scale also varies dramatically among coins. Ripple already processes 1000s of transactions / second. Lightning network on BTC is growing rapidly. Ethereum has scaling solutions in the works. Other chains have completely different properties than the top two and have different answers for the scaling issues. These issues will be solved with time.
As for rants on ICOs and scams, etc. I mean, do you know how many people have been scammed from phishing e-mails? Does that mean we should not use e-mail? No. As always, caveat emptor, do your own research, and don't take unnecessary risks.
> On POW and the electricity argument, I think we waste electricity on lots of frivolous things. I'm willing to bet Christmas lights also consume more than the state of Delaware.
Delaware used 11.26 TWh of electricity in 2016 ("total retail sales", i'm ignoring the direct use bit, sue me):
A light user of LED christmas lights ("1 to 3 wreaths, a garland, and a total of approximately 10 strings to wrap their outdoor trees"!) will use 96 watts while operating for 5 hours a day, 30 days a year, for 14.4 kWh in total:
That is based on a dubious estimate that miners will spend roughly 60% of the price of one BTC in electricity to ensure profitable mining. The price of BTC is highly volatile, the hashing power of the network is not nearly as volatile as the price. No one knows exactly how much energy is being spent.
I don't think you are right that electricity is consumed only for the creation of new coins. The proof of work is done for every block of transactions, and would still be part of the system if no new coins were created.
Correct. Producing new coins is a side effect of mining in POW systems in order to help pay the miners.
For bitcoin specifically, once the reward falls below the Satoshi, the entire cost of mining will fall on the shoulders of users in the form of transaction fees. This will not work very well.
I would not bother if I were you, I've never seen a coverage so negative of blockchains as the one we have on HN. Every week you have a new article looking like this, full of straw man arguments, never any pro/cons. "You Probably Don't Need a Blockchain", yeah you probably did not need a website in 1999 as well, and web technologies were crap. Cryptocurrencies are completely new field which needs time to mature, basing the potential future of it on current technologies isn't very wise.
I remember the Dropbox post on HN full of people asking why they would need it when they can use USB keys and ftp, I feel cryptocurrencies are clearly underestimated the same way here on HN.
I think HN is responding to the almost hyperbolic proliferation of "blockchain" as some digital panacea and social zeitgeist. It's cool to talk about "crypto" at dinner now. Every time I hear some suave new BTC owner at the table next to mine undressing his two lady friends with technically inaccurate explanations of how crypto currencies work I die a little.
But I think you're right. Time will tell. I would love to see more distributed software that empowers participants to grow stake in the digital systems they depend upon. And having well-curated frameworks and working knowledge for operating blockchains will help us grow our digital gardens. The renaissance has only just begun. Patience is a virtue.
The first point the author makes is the disadvantage of immutability. Claims that few systems benefit from immutable transaction records. First of all, I don't think that's true. Many systems do, not just financial systems with double entry bookkeeping. Event sourcing is a well defined architecture pattern in back end systems, and it turns out that immutable event logs are actually incredibly useful. It's also pretty well known that you don't erase history in financial systems. That is one of the pillars of double entry bookkeeping.
A more balanced article would point out the pros and cons of immutable systems, but to flippantly argue that immutability is useless? That's just lazy.
Tracking shipment of goods on a blockchain has many advantages. Namely, you can create a shared protocol for the data of shipment tracking, and you can open that blockchain up to certain parties to sign as the good moves along the supply chain. It doesn't have to be "trustless". A particular app's blockchain might only be open to a few trusted parties to write to. Not all transactions have to be written to a single, global, immutable ledger. There are still great benefits to sharing a common blockchain among specific, trusted, cooperative entities.
On POW and the electricity argument, I think we waste electricity on lots of frivolous things. I'm willing to bet Christmas lights also consume more than the state of Delaware. However, it's not the processing of transactions that the electricity is being spent on, it's the creation of new coins.
POS is a good alternative. I think many coins are starting to leverage POS. Ethereum will be switching to that soon.
Still, POS vs POW has nothing to do with "why you probably don't need a blockchain". These are implementation details that vary per blockchain/coin.
Scale also varies dramatically among coins. Ripple already processes 1000s of transactions / second. Lightning network on BTC is growing rapidly. Ethereum has scaling solutions in the works. Other chains have completely different properties than the top two and have different answers for the scaling issues. These issues will be solved with time.
As for rants on ICOs and scams, etc. I mean, do you know how many people have been scammed from phishing e-mails? Does that mean we should not use e-mail? No. As always, caveat emptor, do your own research, and don't take unnecessary risks.