For a long time, I thought of the blockchain as almost synonymous with cryptocurrencies, so as I saw stuff like “Odyssey” and “lbry” appearing and being “based on the blockchain”, my first thought was that it was another crypto scam. Then, I just got reminded of it and started looking more into it, and it just seemed like regular torrenting. For example, what’s the big innovation separating Odyssey from Peertube, which is also decentralized and also uses P2P? And what part of it does the blockchain really play, that couldn’t be done with regular P2P? More generally, and looking at the futur, does the blockchain offer new possibilities that the fediverse or pre-existing protocols don’t have?
The “blockchain” I use on a daily basis is git, where the sha of the previous commit affects the next.
Given that git was invented before the word “blockchain” started being used, shouldn’t we call blockchain applications “git-like” rather than retroactively calling Git a blockchain?
I thought it sounded interesting when it was new but the more I’ve learned, the more convinced I am that it’s completely useless. I’ve never seen anything done on a blockchain that couldn’t be done faster, cheaper, and more securely in a SQL database. Even the not-a-scam applications are ridiculous and fall apart upon examination. Blockchain as a definitive record of ownership? Absolutely not. There’s no way to force a person to update a record. Lose your house in a bankruptcy? The sheriff on his way to evict you isn’t going to care that you’ve got some NFT saying you still own the house. Anything involving contracts at all? If a court can’t unilaterally update the blockchain record, then the record is unreliable. But if the government can unilaterally update a record, then you’re not relying on community consensus and immutability in the first place.
Blockchain isn’t useful for anything important, and it’s not a logical choice for anything trivial aside from literally just playing with blockchain stuff for the sake of playing with blockchains. I think it’s a dead-end technology.
Blockchain as a definitive record of ownership? Absolutely not.
Oh, its worse than you think.
https://www.cs.princeton.edu/~arvindn/publications/mining_CCS.pdf
Once BTC hits enough halvening-cycles, the entire protocol doesn’t work anymore. Its more beneficial to fork the blockchain (and collect ~50 transaction fees), rather than work on the head (and only collect ~5 transaction fees).
So if the last block confirmed 100-transactions (aka: collected 100 transaction fees), its more beneficial to undo that block and “steal” ~50 transactions, knowing that you’re leaving ~50 transactions for another miner to follow onto your block. (Ex: there are now two blocks: one with ~5 transactions available, the truth… and ~55 transactions available. The lie / false block you created. The lie is more economically beneficial to the next miner, so they’ll switch to your block).
It turns out that BTC forgot how to handle ties after the end of the “Free reward”, and there’s a good chance that “definitive record” is not so definitive.
Blockchain (simplified) is a giant excel spreadsheet that you can never edit, only add to. I struggle to think of any applications that is a benefit for, and even then append only databases would already do it better.
One of the benefits is supposed to be decentralization, but people tout that as a benefit for things like house deeds, or identification, or whatever. Imagine how massive an append only excel file of every house with every owner change etc etc included in it would be. Then we once again only have the people who can afford to store that much data storing it, and we are back to where we are now.
It doesn’t really solve any problems, it just is a worse version of what already exists.
Something about this comment didn’t seem right to me, so I did some quick math:
There are approx 144,000,000 homes (incl apartments, etc) in the US. https://www.census.gov/quickfacts/fact/table/US/VET605221
Assuming every home is sold 5 times on average, that’s 720,000,000 sale records/deeds.
Existing blockchain implementations use IDs that are around 32 bits, or 4 bytes.
A “home sales record” or deed on the blockchain needs to include the buyer and the time/date of sale (8 bytes), along with a cryptographic signature (4-16 bytes). The seller’s identity doesn’t need to be included because it’s always doing to be the previous owner.
So each record is 16-28 bytes, and there are 720,000,000 records. If we go with 28bytes, it would take about 20GB to store all of the deeds for the US. A 500GB hard drive costs $20.
You forget that the blockchain is all about not trusting some middle-man/site, so you need to stock that blockchain yourself, everyone needs to stock that blockchain.
So multiply not only the cost, but also the ecological impact just buying all those drives.
And that’s only for *US" housing (I didn’t get the timeframe you used to calculate it, is it for like year 2050? Old data stays forever.).
BTW found the guy buying 0.5TB Hard drives ;-)
Yes, everyone would need a copy of the 10s-of-GB blockchain. That’s a fraction of the amount of space a single computer game would use, does that seem unreasonable/impractical to you?
And I buy used enterprise 2-3TB drives on eBay :) . I was going to use a 32GB flash drive for my example, but a 500GB HDD is the same price
Fair enough about the size.
Checked out eBay, there are some cheap 2-3Tb drives there! How does it pan out quality wise? I guess they sell them off like after 5 years of usage right?
Yeah, that’s my understanding. Tbh I don’t have a lot of experience with them yet, but I’m building an 8 disk RAID6 array and I decided to go with those used drives. 10 matching disks will be around $120, and I’ll have 2 extra drives so I can rebuild the array asap if a drive fails.
Edit: I also backup all of my important files, so it wouldn’t be the end of the world if the entire array fails. And a little downtime isn’t that big of a deal for my home server, unlike a commercial data center.
…and 20G that needs to be replicated to tons of nodes if it should be really decentralized.
16-28 bytes seems extremely understated, I think it could easily be off by orders of magnitude.
What do you think I’m missing in my estimate? Do you have any experience in CS?
You’re just talking about ownership of a title right?
A deed contains a lot more information than the owner. Mine is 4 pages long. Contains a map of the street, various easements, et cetera.
Yeah, but that kind of data would be better in an auxiliary database. There’s no reason to include it in the blockchain.
well but doesn’t that beat the purpose of using the blockchain in the first place? why not just store everything in the auxiliary database?
Depends on why you want to use the blockchain, I guess. If you want a system to allows anyone to verify ownership of property without 3rd parties (government, etc) being involved then the auxiliary database should work fine.
What purpose were you thinking of, that would be defeated by an aux DB?
As we see here on the Fediverse, decentralization works fine without monetization using an actively anti-scaling append only database that emits the pollution of a medium sized country.
The only other good thing that came out of it is it increased the prevalence of digital payment system in the world, but I struggle to think of anything that would actually directly benefit from blockchain.
Merkel Trees are fine, and are how things like “Git” keep track of different files (and how distributed hash tables and file-sharing often work).
Merkel Trees are trees-of-hashes, which the cryptocoin world wants us to believe go by the new name of ‘Blockchain’, but people familiar with comp. sci history know that they’re just flailing about making shit up.
Blockchain is an application of Merkel Trees. Merkel Trees have lots of good uses, but Blockchain doesn’t seem to have much use after 10+ years of experimentation.
Yo!
You seems to know what you’re talking about, have you heard of Juliana trees? Like trees based only on the keys, so searching for a key takes len(key) time.
Bet there is an other name for it but I so remember like that and no web search says anything about it so I’m trying my luck here :-)
Same for robin hood hash trees :-D
Cheers
This sounds similar to a trie - maybe that’s what you were thinking of?
I mean look at big corpo/government servers. They’re running an OS from 1980 and software that hasn’t been updated since 1960’s.
We’ll get there. Eventually. Maybe.
My Android phone’s marketplace, Google Play, is literally newer than Bitcoin and I have government services + banking applications on it.
I think you’re blind to how slow and inconsequential cryptocoin’s entire world is at innovation. They’ve wasted 15 years. Literally wasted, the world has changed and they haven’t noticed.
ETH has DNS. I would think the fediverse would like to see adoption of DNS that governments and big companies can’t mess with or take over with lots of cash.
ETH staking is looking like its literally illegal in the USA, you know that right?
Coinbase Earn is quite possibly trading in unsecured, unregulated securities and is being sued over it.
As fast as money talks, you’ll be losing.
IMO. We should make global random networks and base our connections on top of them instead of clinging onto the hope of niceties because someone have the site google.com for example.
That kind of thing is actually possible with Ethereum DNS and hosting. It’s not mature enough to be viable yet, but the possibility is there.
A good start, but using crypto to just have a website seems overkill.
I have built a shared hosting protocol (and implementation) where you use link-files instead of website+DNS (nor crypto). Simple and lightweight, but with my communication skills it isn’t really taking off 😅
That’s cool. Got a link you’d be willing to share?
Here is the ‘official’ website : Tenfingers
There are versions for Linux and Windows and also the python source code.
Any feedback greatly appreciated!
I find it to be an interesting solution looking for a problem. There could be many applications but I’ve yet to see one that blockchain could solve better than anything else that we already have, outside of crypto currencies.
Web3 is an interesting thought experiment but I don’t see how it would work in real life. It would be extremely slow, data loss would be a daily occurrence and it would be a privacy/security nightmare.
NFTs could have a useful case of keeping online records of ownership. Being cars, homes and even cattle. Which coulld also make it easier and cheaper to sell or buy these things.
Ignoring privacy concerns of course.
But you would still need an authority that can unilaterally make changes to these ownership records. People die, things get lost, stuff happens. So it can’t all be based on signing with private keys of individual persons. At that point: Why not run a central database of it all. It’s cheaper, more efficient and you could still publish a public record for traceability.
I really don’t see any problem that Blockchain could solve better than other solutions. Except Cryptocurrency.
So Google, Amazon, Apple, and many other large companies in the IoT space are using a blockchain as a federated data store: https://github.com/zigbee-alliance/distributed-compliance-ledger
It stores the data needed for Matter [ https://en.m.wikipedia.org/wiki/Matter_(standard) ] device attestation.
I think its an interesting use case on how entities that don’t particularly trust each other can operate a federated system. Accounts are linked to an identity out-of-band in order to have write permissions to the chain. When an account writes, all the readers of the chain have reasonable assurances of the author of that write. No company can inject false state as another company without that company’s guarded private key. All transactions are also auditable as an additional assurance the data isn’t undergoing a malicious act.
tl;dr; interesting use cases for tamper proof federated ledgers.
