Topic: If you could design Bitcoin from scratch today... (Read 249 times)

Yes, that's true. I forget this topic is about theory, not practice (since in practice people usually fork one of full node software/implementation for new coin).


Monero use such PoW algorithm. There's also algorithm called ProgPoW which claim it greatly reduce ASIC efficiency gain.


RSK has better decentralized compared with other Bitcoin sidechain since it perform merge mining with Bitcoin. But i don't know decentralized is it in practice. And those aren't sidechain, but Taro and RGB which build on top of LN can be used for smart contract.

built under, not ontop

(dont fall for a subliminal trick of saying certain networks are on top(better)

data/science in levelling is not what some promote

take website domains
the top level domain is the main director network. which then has domains and subdomains below it

top level code is the entry point. the main file .. which then contains the reference modules (imported resources) which other down level functions then need for their niche purposes

Interesting question.  If I could design Bitcoin from scratch today, there are a few changes I would make to improve the system.  Here are some of the changes I could make:

Improved transaction speed
Privacy improvement
Energy sustainability
Scalability management

But maybe that's okay already

theories about more coins/more devisions
these idea's fall flat because. even if coin supplies change from 21m to 42m or 2.1b or changed from 4decimals 8 decimals 11 decimals or 14 decimals.. eventually people will still "want more"

side note: more divisions of sats or more "coins" equal the same end result. more shareable units in circulation
at data/binary level of value stored in a blockchain
210m 'coin' at 8decimals.. AND 21m 'coin' at 9 decimals are
1001010100110110110001110000100010010001000000000000000
yep its the same measure at data/blockchain level

the difference only occurs only at GUI expression (human display translation)

anyways..
bitcoin already has ways where you dont have to buy a whole bitcoin
gold has ways where you dont have to buy whole tonnes

infact at data level of blockchains there is no btc. everything is in satoshi measures

people can easily choose to display their balance(GUI level) in bits mbits and other such measures

i personally dont count bitcoin value in whole coins. i measure value in 0.001 allotments
EG $22.28 right now

when i bought bitcoin and mined and traded from 2012
1btc then was $6 so buying say 1000 'coins' was $6k
now using $6k i can buy 270 of millibits
or 270000bits
which psychology is meeting the buying "many whole units" mindset

..
with all that said, people psychologically are stuck under the mindset of needing to buy whole shares/whole products/whole baskets
so no matter the division idea. people will still argue/demand more divisions or more coin(more supply either way of shareable units) because people cant escape the mindset of whole unit required purchasing

education should be done instead that people can buy divisions
informing people they can buy 1000bits is the same as 0.001btc. thus not needing to endless add supply

..
as for worrying about when 1 sat is 1cent
inflation means that compared to 2010 where chewing gum was 50cent and 2023 its $1
by the time economics moves $ value more. the cost of real world comparive prices of goods also moves

fees should remain less than chewing gum prices
so when things move forward where 1 sat=1cent no one can buy anything for a cent. and cheap items convert to higher prices
where in the future people treat $1 the same in the future as they treat a dime today

where cents wont even be part of a conversation about money

psychology of economics of inflation
todays Dime name and value. will be the replaced psychology of cent
todays $1 name and value. will be the replaced psychology of dime

EG years ago chewing gum at 10cent was talked about in cents
now at $1 chewing gum is talked about as 10 dimes or $1
in the future due to inflation when chewing gum is $10
people will talk about it as 10 bucks/100dimes. and no mention of cents are made

no one will measure things in cents. they will measure in dimes as the smallest useful measure

thus a sat is still 10x lower then anything useful(dimes) in the US real world

you can see this because in america prices of goods that are nearly $10 are not $9.99 but instead $9.95
because they dont want to deal in cents. they at worse want to deal in nickels

fast forward a couple decades and nickels will lose psychological meaning and usefulness

21 billion bitcoin:
I like the idea someone put up of 21 billion Bitcoin thus making the price in the double digits rather than double digit thousands. This is just purely an optics things for newcomers. I mean I do think a high price gives the idea that Bitcoin is extremely valuable, but also it is a fact that many people new to Bitcoin just assume it can't increase in value much anymore because it is already so high. I talked to a friend of mine who only had doge and I asked him why the heck he bought something as silly as doge (cuz of Musk of course...this was back in early 2021) and told him he should buy Bitcoin and he said it's already at $50,000 so it can't possibly go much higher. People new to Bitcoin/Crypto don't have an idea of how big global money can be, and how small Bitcoin's market cap is compared to its future potential. They just see a price high in the thousands and think it can't increase much. There is definitely a price-unit bias against Bitcoin for people uneducated on Bitcoin which makes newbies more likely to buy some cheap crap altcoin rather than Bitcoin.

I don't think a satoshi should necessarily be the unit called a bitcoin (ie 2,100,000,000,000,000 bitcoin) because then Bitcoin would seem like some worthless penny stock to newcomers. So I like the idea of 21 billion bitcoin instead of 21 million. That way even when Bitcoin gets up into the millions it would still seem affordable to noobs cuz it would only be in the low thousands, but it would still seem valuable with a fairly high price.



Add another decimal place:
I would also give Bitcoin more digits of division. By which I mean increase the number of satoshis by 10 by giving one more decimal place to the currency. Once the price of Bitcoin goes over a million dollars one satoshi is going to be more than one US cent. So probably sometime next decade the smallest unit of bitcoin won't even be smaller than the smallest division of USD and I don't think that is good because it'll make converting bitcoin to USD imprecise. So one or even two extra decimal places would be ideal.



Stop Ordinals NFTs:
I guess the most obvious recent thing is fix the unintended mistake in taproot that allowed ordinal NFTs to be created and stop stupid pictures from clogging up the network.



Mining (wouldn't change):
Someone said make mining less energy intensive but I would never want that and nobody who understands PoW and Bitcoin should want that. Also Bitcoin isn't energy intensive, its the economics of bitcoin that create the desire to consume all the energy. The economics of bitcoin mining are incredibly strong and make bitcoin the most secure decentralized network on the planet, so to say it shouldn't do that would mean using something different than PoW which would be very bad for Bitcoin and for the world. Bitcoin only uses as much energy as miners find economically viable. Bitcoin itself is not energy intensive, it simply creates a very valuable market and it is that market that is energy intensive. It'd be like saying you don't like how much volume is traded on bitcoin exchanges and you wish it was less, it isn't bitcoin that creates the volume, its the market.

Also, the fact that Bitcoin uses PoW is a huge benefit to society for several energy related reasons - bitcoin mining can prevent the massive waste of unused energy in society, it can strengthen the energy infrastructure of the world, and help drive the world to renewables by making them more profitable and increasing their manufacture by using them to collect stranded energy for mining.

Someone also said have all the statistics of mining - energy source (ie renewable vs fossil), and who, what, where, when of mining recorded but that simply isn't possible because all those things have nothing to do with mining and they are external data. There is no way to put/force that external data into the blockchain. This is why Bitcoin relies on energy for security and not something much easier to do because there is no way to fake energy input, while inputting any sort of external data can always simply be made up and there is no way to enforce accuracy. Even if the system made miners fill in some data fields with all this information there would be no way whatsoever to confirm its accuracy. I mean if you could make a blockchain's security run on any external data and still keep it decentralized you could just make it so each human can only count as one miner, that would be ultimate decentralization, but inputting enforceable external data isn't possible in a decentralized system.



Slightly larger blocks:
If designing Bitcoin from scratch now I'd probably up the blocksize by maybe 4x. There isn't much point in going really large because no amount of on-chain scaling would be enough to handle mass transactions, whereas trying to do so would just turn Bitcoin into a centralized altcoin and then it would lose its entire value proposition as the best money humanity has ever had. L2 will always be needed and is the only way to scale. But I'm pretty sure with internet speeds and disk space improving we could increase the blocksize a few times over without affecting people's ability to run nodes and propagate transactions through the network. This would allow on-chain txs to remain cheap for years longer while L2s mature and become ready for mass use.

Alternatively something like a small increase built into every halving so the blocks grow a little bit once every four years as internet speeds and disk space increase, and once halvings stop the increases would stop so it wouldn't just be a run on thing that gets out of control. This would stop the stupid idea that bitcoin should be a big block centralized coin (like the failed centralized tokens BCH and BSV...or any other crypto) while still giving tx fees some relief over time as adoption increases, and done in a small gradual way that doesn't hurt decentralization because it would allow technology to keep up with the increases. And L2 would still be needed for the 99%+ small every-day payments and for in-person situations where payment needs to be made within seconds.



Design with Layer 2's in mind:
I'd want to figure out the best L2 networks that could be designed and figure out what on-chain features are necessary to allow those L2s. L2 is absolutely necessary, so a blockchain built with L2 in mind and with L2s created at the same time as the blockchain to make sure everything works as designed would be key.



ASIC resistant mining:
I guess some PoW chains have ASIC resistance don't they? I mean if there is a way to allow CPUs or GPUs to be the drivers of mining that'd be great because then anyone could mine because it would greatly lower the cost of getting into mining.



Access to lost addresses (wouldn't want because not possible in a decentralized way):
Someone mentioned this and said people don't want this because people want bitcoin to get lost because it makes bitcoin more scarce haha. That is not remotely the reason people don't want this. Nobody wants bitcoin to get lost. But the question is how would this even be possible in a secure decentralized network? I don't think it is possible.

Ethereum just implemented the ability to get back lost funds but I think the only reason it is possible on Ethereum is because they have the EVM running on top of the blockchain and so I think accounts run on the EVM and so I guess they can give access back to accounts without needing the keys I guess? Also from what I can tell there's basically a council or something on the network that will decide requests to re-enable access to accounts. It all sounds very centralized and also I imagine this will open up a whole new attack vector on people's Ethereum accounts as hackers try to convince whoever is in charge of granting access to accounts to give them access to accounts that aren't actually there's. Granted it all sounds amazing from a user experience making Ethereum more usable and simple for newcomers, but it's just an example of further centralization and new attack vectors, and it sounds like Ethereum is essentially turning into a bank, while Bitcoin's intent is to be secure sovereign decentralized money so people can bank themselves and not need something like the Ethereum foundation or a bank to have ultimate control over their finances and their money, so I don't think there is any way to have this sort of thing on Bitcoin without trashing Bitcoin's whole purpose.



Quantum Computing resistant:
From what I understand when quantum computing gets mature they will be able to hack private keys from public keys, but not from public addresses. And apparently there are still 2 million or so bitcion held in the original transaction type in which money was sent to public keys rather than public addresses. Meaning at some point there's gonna be around 2 million bitcoin (I guess half of that is Satoshi's) that will be able to be stolen and put back on the market by the first company/govt with a fast enough quantum computer. So if this is all correct, I would obviously design Bitcoin today to only user the latest tx types which would not only make it more efficient than the older ones but also remove this quantum threat.



Side Chain protocol for smart contracts:
I suppose, in addition to as stated earlier designing L2s from the beginning alongside Bitcoin, I would at this point want a L2 or sidechain or something like that with smart contract abilities to be designed from the start with Bitcoin. Keeping it as a L2 or sidechain or whatever would keep it from bloating the blockchain, while also essentially making all of Crypto worthless because there would be no need for anything but Bitcoin. I know there are sidechains and stuff on bitcoin that allow smart contracts but they seem to be centralized like altcoins in that they are products of companies and run by those companies or by a small group, rather than set up in a decentralized protocol sort of way.



Namespace for addresses:
I would want a namespace stored on the blockchain to relate public addresses with comprehensible names. And you would be able to assign an address a name simply by sending say a 1 satoshi transaction from that address with the name you want....or something like that. This would make sending transactions to people wayyyyy easier because you wouldn't have to carefully type in some crazy set of unintelligible characters. And if obviously wallet software would make sure the name is in existence or even let you know if the name on an address you've sent money to before has changed.



On-chain upgrading
Some altcoins do this, where upgrades are voted on-chain and so I assume the whole process is somehow done through the protocol. This would be great for improving Bitcoin's 'open source-iness' and I think it would allow small desirable changes to the protocol to happen quickly rather than big upgrades only once every few years.


That's everything I can think of.

you currently can ask a friend you trust for their existing copy of the blockchain folders

no protocol changes are needed. just a portable storage device that is more then 500gb to pass the data to you (its a simple folder copy and paste) outside of any software requirement

as for "offline payments"
well there were offline paperwallets called "bitnotes" and also physicalcoin wallets called "Casascius coin" which people can trade hand-to-hand without need of the internet

this is where the private key of funded amounts is passed between people in paper or metal format. hidden with a physical seal

 but you would need to trust the person handing it to you wont move the funds before you were able to get online to move them to a new key only you owned.. years ago there was a guy that done the casascius coin where he blindly created the private key thus he didnt know the private key of the funded coins he handed out. thus allowing people to not have to rush to move funds off the coin as the private key was sealed under a sticker no one has seen

I could have added one important program related to offline synching of bitcoin and thus adding the transaction information later whenever the device is connected to the network. This functionality feels so important because there are regions or there are instances when we dont have the network. Let us say I am roaming around the globe and there is just a country where my SIM is not supported and I wanted to pay in the bitcoin payment then I could just grab  my device and do a quick QR scan and send type of payment. Since I have done this offline, network would not know about the spent bitcoin's until I go back online. But once I do, it can get embedded into next available block and time stamp it as confirmed transaction once it does.

The circulating supply of the bitcoin would not get hampered with this because it is getting calculated when miners are minting it and releasing for usage over the blockchain.

I might not be technically correct here but this is best feature and I still wish it can be added to it.

this topic is about theorising a "from genesis" "from scratch" of idea's

this topic is not advocating a new bitcoin or a new blockchain or a new proposal for the current one
this is just DISCUSSING (forums purpose) random idea's people have of an "what if"/"if you could" statement and someone had an idea so i discussed how it could be possible

treat this topic as a topic to free your mind and think of possibilities. like a game. where your free to express yourself

its a theory topic not a proposal topic

....
yes producing a  hash ID of a live/continually updating utxoset takes time. which is where some previous idea's fall flat. .. which is why others like the ball and chain method uses periodic snapshots. thus plenty of time between blockchain trims

If I have to restart it, only the total supply to 21 Billion. If the price will be just below $25, its better so that everyone can be onboard. It doesnt have a value of $50k to be recognized.

But I will leave the global crisis still so that people will be aware of BTC that government officials are using it and some official orgs are using BTC.

Designing bitcoin from scratch required a very high level of skill that not many coders could do. And I also don't know if I can design bitcoin alone from scratch because I'm sure it is very complicated and would not have been possible without high expertise. And given the development of bitcoin the way it is today, I'm not sure what design I need to add for bitcoin because I see bitcoin is pretty well developed right now and just need to see it adopted by more and more people. In addition, I also did not follow the history of bitcoin creation from the beginning.

If we were to design Bitcoin from scratch today, it would be a bomb-ass project! There's so many dope directions we could take it in, but the main focus gotta be on makin it more secure and user-friendly, ya dig? We gotta figure out ways to integrate it with other fly technologies to create a smooth experience for all the users out there. But, at the same time, we gotta be mindful of what could go down if we create a decentralized currency, ya feel me? We gotta think about how it could impact the global economy and all the power dynamics at play. It's gonna be a tricky balance, but if we plan it right and think it throuh, we could create a revolutionary system that'll be off the hook!

solution to this is simple.
multisig 2 of 3
have 3 keys somewhere at youhave access to but not same location


that way if you lose one. you still have 2 more. (separating locations reduces risks of losing all three at same time)
and when you lose one its then time to move funds to a fresh 2 of 3 to have 3 keys again to avoid the risk of losing a second key if left to long on an already lost key set

chances of losing all three at the exact same time is high if same location stored. so spread them out. give them to friends or relatives knowing they cant do anything with only 1 key they hold.


as for any protocol that can move value without key signatures(your permission/consent). a system that allows that is a system i wont want to use. the chances of abuse becomes rife
(EG fiat account forfeiture)

EG. find an old stash. pretend its yours but you forgot your proof. and boom you win someone elses stash.. um no thanks

Redundant transactions do not need to be recorded or reverified.

There would be a set of conditions for redundant transactions, which would allow the data to be culled after a period of time had elapsed.

There are no chargebacks or reversals in the transaction system. Which are some of the main reasons for long and detailed transaction chains being important. Once something has been verified for a period of time. It can be considered etched in stone. Which makes the record of it less important, as it will never be amended or revised. The old data can be archived on a separate chain, which would be considerably larger. While the smaller chain would be the one in active use for a leaner and more streamlined architecture.

Crypto miners might also be tasked with the culling process which would function similar to current BTC rewards.

Say there are 1,000 transactions in the chain preceding the current transaction. Do we really need 100% of those 1,000 transactions? Why don't we cull 900 out of 1,000 and keep only the last 100. With the raw data being archived in a big file chain. If this is done on every transaction branch, it might shrink the size of the chain by a factor of 90%. If the current bitcoin blockchain is 250 gigabytes, this might shrink it down to a far more manageable 25 gigabytes.

I don't think fees are pegged to dollars. You pay with sats, and you can manually select the rate of the fee you're choosing to send your transaction. I've never seen a clogged blocks that requires you to pay more than $50 to make $ a transaction. What wallet are you using right now? You might want to switch if it always sends overpaid transactions.

I don't really see how it will work though, as mentioned above. And who can guarantee that somebody else would steal your keys in one way or the other?

i prefer full nodes to be full nodes.. where its more of a majority. rather than 'just 1000'

but i was just displaying how what today is termed 'blockchain culling' is able to work. as one quick example(just as a thought experiment not a proposal)

some people have already theorised, desired and implemented methods have fool* nodes but you asked how a culled blockchain could work. so i just responded with just one idea that has been whispered around in the last few years

i personally would prefer full nodes to actually be full nodes. where they have a strong consensus ruleset not a soft ruleset. where all transactions are verified and all tx formats have actual byte limits..

and if people then want to be fool* nodes. then so be it. let them know they are a step down from full node. and its their choice for their personal use to accept the step down.

*fool node. sounds the same as a full node but doesnt do everything. EG it doesnt validate all data or it prunes/culls blocks.. basically its not a full node even though it wants to sound like it is

I don't think it's the fact that people don't want it. It's the fact that there's no way of doing this without having a centralized entity requiring AML/KYC. The closest thing we have to this is social backups, like what Argent(ethereum wallet) has.

Sounds like a good idea. But if you have a wallet, you'll lose information about all the transactions you made more than 4 (for example) years ago.

Instead of replacing all full nodes with these "part nodes", I'd suggest that the majority of nodes become part nodes with the culling feature, but there are still about 1000 or so nodes with full blockchain history connected to the network, so that these transactions can still be fetched, albeit more slowly, as they are only expected to be referenced rarely at most.

It is something that can be done now (with another program, without modifying Bitcoin Core) and not an "if only" idea.

Bitcoin is already good enough.  I thought there is nothing to re-design again. But, look at comments, people have more ideas of how it can be improved. Of course I do not have enough knowledge to design or talk about the functionality. But, if I could be a part of the team, I may ask them to keep it minable with GPU's as well.

As we know, at some point, Bitcoin was minable with Graphics cards. Unfortunately, I wasn't aware of Bitcoin. I wish I was there and I could mine Bitcoins with my Graphics card. I accumulated some sats with my Gaming PC from Nicehash when ETH POS update.

the codebase may seem very vast. but it also has alot of cludgy code in it to seem vast
(usually developers purposefully make it cludgy and hard to read/follow to stop outsiders from  competing/changing things)

there are simpler ways to add checks/rules for certain data types. far simpler then the cludge sipa implemented over the years

even things like checking block rewards.
todays code had the coin= units / 100000000 and then more code to then show a 50btc
and then more code to show a depreciation factor of 50% per 210k block and then counting them out

however simply by saying block reward starts at
binary: 100101010000001011111001000000000    (5000000000sat aka 50btc)
and each halving is
binary: 100101010000001011111001000000000   (2500000000sat aka 25btc)
binary: 100101010000001011111001000000000  (1250000000sat aka 12.5btc)
binary: 100101010000001011111001000000000  (625000000sat aka 6.25btc)
and so on

can be done in code form far far far easier


a method can be what has been theorised before as a 'ball and chain'
(ball is a unspent snapshot of a milestoned point dataset. then the normal chain updating from it)

imagine at block 210,000 a snapshot of the utxo is saved and a hash ID is created.
this hash id is put into an op return of coinbase of a block so that its hash is locked into the blockchain as a reference and then after another far later milestone when it shows acceptance and agreement of that snapshot by the masses, by not rejecting blocks containing it
meaning masses agree that their snapshot complies with the same hash as the one in the opreturn reference.. that then becomes the list of unspents from the standing point of block 209999

then instead of storing full blocks of 0-209999 it stores just the unspents of that milestone
then repeat at block 420k and 630k and 840k

---

as for what id change
all tx formats have rules.. with byte limits to keep tx as lean as possible

where a certain opcode has the anyonecanspend(no limit/format/check) but the opcode is disabled be default UNTIL good clean code is reviewed independantly and scrutinised for validation bugs. and then activating the opcode, which requires nodes to upgrade. then activates the new tx format with all checks and validation rules implace including byte length thats fits it sole purpose

Most people would not be able to make a redesigned Proof-of-concept Bitcoin clone, even if they happen to be a programmer, because the codebase is really vast and so you'd have to use one of the early versions that are small enough for a single person to successfully tweak.


What? How is that even possible, without destroying full blockchain verification?