Bitcoin Forum

Imagine having a nice beer at a local bar and suddenly you realize that the guy next to you can't pay for his drink - he forgot his wallet. Since you are a nice person and made some money from investing in Bitcoin, you pay his tab.
It turns out, this guy is Bitcoin's very own core developer gmaxwell and because he is also such a nice guy, he offers you to implement one feature into the Bitcoin protocol - no questions asked.

What would you ask him to implement?

edit: sure, no one would be forced to use that fork. Just imagine that this is the only valid version of Bitcoin everyone using Bitcoin has to use form then on!

This doesn't make any sense to me, as anybody can change the code and implement what they like, any time they like.

Sure, alright, what I was implying was that everyone would be "forced" to use that version of Bitcoin now. No way of using the old fork, everyone has to use your amended version.

Isn't it simple? I'd make block that would premine me atleast million of coins hehe.

How would that work, so you would just get a specific number of coins after block X? That way the value of all coins would go down a lot, I don't know if you could profit from that very much...

Hah bitcoin is open source so nobody would be able to do that as source would be inspected for sure before hard fork.
But if i could mask it then let's say 0.2btc from every block mined would be enought.
But if someone would notice it in early stages you would end up with nothing, but if you would insert bigger reward for yourself you would  more likely beable to get more fiat out of it.

I'd request a memo line with transactions. Sure it might add bloat, but it is so practical with a public ledger.

Except unrealistic rainbow-unicorn-in-the-sky changes, I think the block reward could diminish more gradually (and not abruptly every ~ 4 years), and if it should continue to use PoW, it should be more "useful". Ooops, there is Primecoin already.  :-X

I'd implement a change to decrease the average timer between blocks to 1 min or so.

the only thing that i want right now, is a faster confirmation, i'm not sure how the miners feel about receiving a block every...let's say 1 min(like doge)

but it would require reviewing the block structure and reward....too much of an hard fork i guess

That would be extremely useful.

Id request changes to ensure that such power never relies in the hands of a single person again.

Block reward down to every year.

You mean block reward halving?

If I was given a chance to at least implement one feature in my mind, I would say it is the ability to recover all the lost wallets that haven't seen any activities for a long period of time. But definitely that is not an ideal feature to have because the potential of getting the balances of even the active users may be at risk, so I'd rather not wish for any of the features in my mind to be implemented in bitcoin.

not really a must, but if possible i would choose to have faster confirmations. but then again, it would also need to bring down block rewards.

the confirmation times "problem" is overrated, there are ways to work-around that issue in scenarios like the supermarket cash-out

Interesting... Who would you give the coins to? Also, what would be the cutoff rate? Time? Number of blocks mined in the meantime? Wouldn't this make the whole reward structure pretty unpredictable?

Yes halving, that would cause some fireworks.

One change that  I would implement first and foremost. Is a way to prevent the 51% attack from ever being carried out.
Dont ask me how I would go about it, but its something that should be at the top of EVERYONE's list whos interested in the survivability of Bitcoin longterm. As long as the 51% attack exists .. Bitcoin never stands to become a worldwide dominant currency, because its really not 'trustless' long as 51% attack is possible.

I would replace the JSON RPC with some form of bi-directional long lived connection. websockets for instance could be perfect. websockets over tls offers a long lived, low latency ish connection between app and the bitcoin client. i think it would be awesome.

everything is perfect for now. the rest of the system will adapt changes if necessary.

that's one example for a rainbow unicorn in the sky. Without a central authority, 51% is the largest consensus, and thus "the truth" almost by definition in *any* decentralized system.

FTFY.

I'm going to have to agree with this one. As for the first and foremost feature I'd be looking out to solve security issues like this one. If I had another wish I'd slow down the distribution of the coins for now.

That's what your wallet is for - it can privately label your transactions.

1. I would build in P2P mining by default, so we get rid of centralized mining pools and diminish any 51% risks.

2. I would increase the average block frequency to once per 90 seconds.

3. Rather than halving every 4 years, I would have set the reward to gradually increase during the first 5 years, and then gradually decrease (like 10% every year) indefinitely.

4. Rather than retargeting every 2016 blocks, I would gradually retarget after every block (based on a weighted average of the past week worth of blocks, putting more weight on recent blocks).

there isn't already a solution comining to it(still experimental but...it is on the way)? it's called supernode, if i'm not mistaken they have the task to check if the last block is the most valid based on its transaction history, so the longest chain don't count anymore

Every time a block is found I'd have the official Bitcoin client send a command to a connected 3d pastry printer to print a dozen donuts. Then I'd crowd fund sending a specially designed pastry printer with a client built in to every federal law enforcement office in the USA. Then the Feds would kill to make sure the Bitcoin network is safe. You would start seeing "I Support Bitcoin" bumper stickers on the back of black unmarked SUVs all over the country.

I'm perfectly happy with how it operates at the moment. Uncool but true. Many present gripes will be addressed by layers built on top.

If we're allowed to run with our pie in the sky ideas, it would involve wiping out the centralisation of mining at its core and having the mining algorithm serve a worthier purpose on the side.

You wouldn't necessarily have to cut down the reward in half every year, you could just go and multiply it by 0.9 or something, which would smoothen things up a bit.

I believe that this is very very difficult to do. You effectively have to have some mechanism that can't be faked or fooled. You have to keep in mind that it's easy to pretend to be different persons/entities. The system won't even know!

Ha, that's what I was aiming at with this thread: improvements and suggestions that aren't dead serious :D I like many of the ideas around here, but they are mostly pretty 'altcoiny', I want truly groundbreaking ideas.

like edonkey network (or kad network) in emule software, if bitcoin core can integrate a counter of nodes (per hour) and number of BTC transmitted per hour ... it could be cool to see.

But those metrics are already inherently available already, you only need to gather those information and make them available to the person using your client. Don't think this is something that needs to be incorporated into the core protocol.

in prefer "in the core" ... instead of in the web (that it can easly shutdown a domain name).

I would change the protocol so that exchanges could no longer be "hacked" or flash crashed - every order would require a signature from the owner of the BTC made within the last block, before it fills.

This. Best feature so far.

Huh, interesting concept. I think a lot of the quick acting exchanges couldn't work that way anymore, though. I believe what you want is multisig-exchanges, so at least no one can withdraw the coins without the holder's consent. Also that isn't feasible in some case, though.

Could not agree more

IMO, the changes should not be on the protocol, but on the team that write the core nodes. In fact, only a few of core devs control the git hub, this means if these people are compromised, then the whole bitcoin project is down. They must get enough protection

Gavin also said that it is difficult to reach consensus between core devs. Since the decision of a protocol change is seldom a technical but a political one, they really need some kind of transparent decision making mechanism

In fact, infrastructure wise the bitcoin has many area to improve to become a truly trusted world currency, there are so many uncertainties in its security model. Today, people just download a software and pray for the rest part, but when a retirement fund wants to put billions of dollar in this currency for 20 years, the bitcoin network is still not enough robust in many aspects. The risk of ruin is still larger than our financial system today

I'd reduce rewards and replace halvings with smother reduction of rewards. I would also make it multialgo and implement KGW just in case.

IMO, it's too late to make any changes in protocol. It have solidified and even minor changes would require tremendous efforts. For example, block size limit. At the time it was introduced it was interim, quick and temporary solution with random number (1MB). But even changing this number now takes multy-year discussion.

So we could assume that practically no changes will be made to the protocol.

Therefore, whatever we want to change, have to be done on top of the protocol. It's not easy. How can we decimate block time, how could we spread halving - all without changes in the core protocol? That's the challenge!

Decrease block timer, decrease reward, it would still make the same amount of Bitcoin yearly, just need to adjust or inflation.

THAT IS AWESOME!!!!

Most answers show how much people don't know about the details of Bitcoin and the reasons the current features are like that. They try to reduce the confirmation time, because they imagine people waiting in the supermarket line for ~10 minutes, when in reality most transactions today are accepted with 0 confirmations, with little risk (proportional to the supermarket-type purchase, anyway).

I would like normal people with low spec computers. {CPU/GPU's} to be once again able to mine Bitcoins. It should have built in code to detect mining farms and throttle/reduce their income, to make it less viable for them to mine on such a big scale. {ASIC proof}

Then, companies like DELL can pre-install mining software with every computer the sell, to be enabled by the user, when the computer is not in use. {The user will then get rewarded for sharing his CPU/GPU to help with mining and he can pay his electricity bill automatically} In a perfect world, this should cover his full bill, with the added electricity he used for normal use and the extra electricity he used for the mining.  ;)

This will put some money back into the pockets of MANY people, who are in debt and enable projects where electricity costs are a issue and also help with sustaining Bitcoin mining.

We can just dream.  :(  

This seems very important to the infrastructure and trust in the system. Somehow reduce the processing power needs so that a simple smart phone can participate in the P2P verification and message relay processes of the blockchain for decentralization purposes. On that note, is it possible to break up the block chain so that limited data is stored by smaller processing power devices? For example, save only the last "layer" or latest balances of the blockchain. Or possibly only a "regional branch" of the blockchain. Not regional in the geographic sense, but regional in the sense of one branch of the blockchain. I think I've seen both of these ideas thrown around.

I would make the mining algo double after 5 years.
It would start at sha128 for the first 5 years. Maybe a few people'll get hacked, it'd be cool.
Then on the 5th year (at this current time) it would be sha256 and everyone will be confident that their coins are secure.
Then on the 10th  year its sha512 and people will be very secure.
And it keeps going up.
Ultimately it'll probably never work. Addresses will keep getting longer and longer. (I think) but hey, it'd be a funny altcoin. I don't know how it would work though.

As I said before, a network secured by smartphones is way too insecure for me to be interested in it.

SHA1024? I don't think that exists yet.

Phew... sophisticated input this early in the morning. Yeah that's one of the real problems. Core devs could be compromised, etc. But I think it's more probable that they don't earn enough for their work on the code and therefore leave. I think maybe... a voting system for which fork to use or which pull requests to implement maybe? Of course PoW-based!

People saying that we should have an ASIC-resistant algorithm (however that's mathematically possible) kind of reminds me of homeopathy, where “less is better”. In this case, less device power will somehow bring better security.

Even if homeopathy is nonsense (which it definitely is), in some cases less can actually be more. The thing with ASIC-resistant is the idea that it's impossible to have such a great advantage (10^3 or more) by using specialized machines, which can put the decentralization at risk!

The idea is very interesting and I think you are trying to tackle a very important problem, but building specialized machines (ASICs) in order to solve a PoW-based task will always be possible, I believe :-/

The only thing I want if let's say modifying the protocol is possible is to make centralized mining become obsolete or in a way turn it to become unprofitable. Something similar like hash rate contribution from the same IP will have less chance of solving the block or maybe there are better mechanisms to prevent this. This will ensure that everyone from the community has the chance to actually mine bitcoin, including a guy who owns a simple mining machine. This will solve on the distribution issue.

Not possible. You can always spread your hash rate across several IP addresses. You are introducing metrics (different IP addresses = better) that can be faked. PoW can't.

Little do these people know the workings of their debit and credit card processing which would take much longer than 10 minutes to confirm. Bitcoin payment processors would be used with most merchants so they'd be ok and insured against any potential problems. One bitcoin has been sent it's as good as received instantaneously.

Also, if confirmation times are shorter, the security a single confirmation adds to the transaction is getting smaller and smaller the shorter the target confirmation-time is!

The current implementation using an integer shift operation to approximately halve the reward is very simple and elegant to those who can read and understand the code. There are lots of problems associated with replacing that with some kind of floating point operation, the biggest one being that you need to make sure that the reduction of reward is the same across all platforms. This is not as simple as it sounds because different computing platforms have different rounding errors. The second problem is how you would deal with the last few Satoshis? I mean if you were down to 3 Satoshi block rewards you can't reduce it so 2.7 Satoshis at the next reduction.

Ah well you can tackle those problems: Store Bitcoins as integers/longs (already being done this way) and then divide them with fixed-point arithmetics in a precisely defined manner. Drop all the remaining Satoshis (we're already getting pretty weird rewards in a couple of 4*n years, actually!)
But I see the problems and understand the beauty in the current implementation.

Even if you have ASIC-resistance, big companies can buy thousands of CPUs and have mining farms as well. It may not be as efficient as ASIC, but they will certainly have an advantage over the common users.

I would add a CheckBox-suppressible ('never show again' and 'remind me later') popup that prompts users to encrypt their wallets,  copy-&-paste/write-down their passphrase so they dont forget it, and present them with an option to download a wallet copy to any media source (cd, floppy, thumb drive, local file, memory chip/card, external device etc.)

What's a floppy? lol

My TRS-80 uses these: http://www.freeimages.co.uk/galleries/transtech/informationtechnology/slides/quarter_inch_disk.htm

I still got a bunch of them. They were an upgrade from the cassette drive. haha

My god, I haven't heard about a TRS-80 since the 70s. Since you're still using that stuff, I'll loan you my old Winchester hard drive. It's 5mb and about the size of a briefcase. lol

common users in this case, would have a bit of breath, because they could use gpu, which are far better than cpu, they will still be behind, but they at least will put a dent on it

haha, I'd probably need to track down an interface to use it. How much did that cost when it was new?

If GPU is viable, companies would use GPU instead. Same result: advantage for companies.

I think I purposely blocked that out of my memory. I want to say $500 but I'm not really sure anymore. 

Just imagine this scenario:

You have a bakery. You make cakes and pastries and stuff, and sell it in your town. Now try to compete with Hostess, or Bimbo (http://www.grupobimbo.com/es/index.html). Obviously, no matter what kind of equipment you buy, you will never be able to even slightly affect their business. Even if there were laws that prevented them to use industrial ovens and had to use home appliances, they still have the capacity of buy thousands of them and hire enough people to still be operating way ahead of you.

There's no such thing as designing a system where minorities get an advantage, because whatever minorities can do, companies can do it ×1000.

ok, but gpu aren't that efficient like asic(and are more bulky, produce more heat ecc...), they cannot reach the same hash, thus the diff won't rise so fast like now, this mean more chance for small miners to compete, this is the real difference

also they are produced by company(amd and nvidia) that aren't interested in mining, which rsult in a low production cycle, again another reason why the hash will stay low for a greater amount of time


the point is the diff won't rise at the same speed as having a new asic every 3 months(just an example)

I would like that 10% of all mined coins would be transferred to a wallet of my choice.




More seriously:

make double spends with 0 confirmation harder to detect with default software's, so accept 0 confirmation transactions would be more secure even for people with less knowledge.  

I would ask for a change in block rewarding system - to make the block rewards difficulty dependent.

The big farms only dump coins to the market anyway, let's discourage their future growth.

That could actually be one of the best ways to store a Bitcoin wallet safely. No one will be able to read that disk these days :D Then again, these disks are rather soft and I wouldn't trust magnetic media with a Bitcoin wallet.

I have a little list....  

Pool mining resistance.  I want cryptographic proof that nobody can efficiently find hashes unless they have the key that can be used to spend the coinbase; this would insure that whoever finds the nonce has the power to spend the coinbase tx.  Pools would no longer be able to rely on cryptographic security to get their block payouts.  The easiest way to do it would be to add a required signature by the private key on the block header including nonce, and hash on the whole block header including the signature.  So without the key it would be impossible to evaluate the hash and see whether it falls below a target.

Fix the nonstandard implementation of Merkle trees.  The one we're using now could, in some very contrived circumstances, hash two non-identical blocks, with one containing duplicates of some transactions giving them the same Merkle root.  Nobody's figured out any way to make an attack using this property, but it's still a stupid property for a crypto primitive to have. Given the possibility of new combinations of bugs to combine it with in the future, eventually somebody might.

Change from secp256k1 to curve25519.  Secp256k1 is not subject to any attacks - yet - but has properties that suggest possibly poor resistance to some types of analysis.  Curve25519 is better studied and so far has shown no signs that a weakness might exist.

Fix the bug that makes all m-of-n signatures consume an extra value off the stack when evaluating scripts.

Bigger nonce space in the block header.  This business with 'extranonce' in the first transaction is stupid.  

The wire protocol should consistently use network byte ordering.  

I would change than bitcoin would be mined slightly faster. I doubt that delaying the time of mining final bitcoin to year 2140 will be good in the long run. Time is of the essence here and our technology is changing so rapidly that in next 20 years we could have totally new leading cryptocurrency because someone will invite new internet protocol or something...

We have a new internet protocol. It's called IPv6. It was officially released in 2012, and as of April 2015, it's used by roughly 5% of connected devices, according to Google (http://www.google.com/intl/en/ipv6/statistics.html).

But, if you connect via IPv6, do you notice? Well, no. These technologies operate on “layers”, and the IP layer is below the Bitcoin layer. Even if we were on IPv8, it wouldn't affect Bitcoin's protocol a bit.

Inflation will be borderline nominal in 10-15 years though. Economics is a considerably more sluggish arena than technology. So far Satoshi made some eerily prescient decisions. One of them is probably realising how long it would take for a new economic concept to take root. You need to maintain that incentive to support its creation for quite some time. The move to pure tx fees for miners will have to be very gradual too.

we also have SPDY 3.1 and HTTP 2.0 but who's counting.

How about a GUI for multisigs in the core wallet?
Right now you have to create them in the console and spend via rawtransactions, unless you want to use some 3rd party.

Which have no interaction with the Bitcoin protocol at all.

An incentive for full nodes, perhaps.

Yea sadly I only run my bitcoin core when I need to, I can't do the 24 hours always on thing. Those that do should get a reward or have an incentive like no transaction fees or something.

True, but peer discovery could be more efficient with an official seednode client rather than IRC

I would not change much with bitcoin. I feel that it is pretty well done concept in general and do not need that much tweaking. I am not a hardcore techie as well, so maybe people who wasts bigger blockchain fork or some other technical change that will laugh at me. But all I want to change is instant confirmation times. That is all.

Hasn't Spreadcoin implemented that?!

I don't know at first the pre set number of coins didn't bother me. But now I wonder what would happen if it slowly expanded after the 21 million. I just worry about lost coins. It's funny I didn't at first but now I think of it. Who knows though.

As I said, we do not live in a perfect world, so my idea might be a bit far fetched... ASIC-resistance is just one of the solutions... but more complex features could be introduced or developed to prevent farm dominance in the mining sector.

I think it might be a bit more expensive for companies to buy huge amounts of powerfull CPU's/GPU's and if you combine the reduced yield from CPU/GPU mining, it might not be economical to do it on such a large scale.

We have many talented dev's to sort that part out.  ;)

I would like to find a way to remove anonymity. It attracts scammers and ponzi pumpers and brings nothing to honest people. If are not a criminal, what do you hide from?

Wouldn't that be against the feature concepts of bitcoin? Identity and credentials are a very vital part of ones' life--particularly people who are surfing in the internet. It can cause security implications if ever your personal information leaked in the interwebs.

usually anon is wanted because of taxes also, not only for criminal things, and you can add privacy to that, not everyone use facebook and want to shouts to the world his identity

Well, you can have it both. You can have anonymity but then you have tons of scammers and ponzi pumpers as well which devastate BTC community. They do it exactly cause they are able to hide behind that anonymity. Bunch of portals, cloud mining websites and investment schemes run by these people and there s no way of knowing who s behind it or trace their transactions.

I m not talking about you or I being able to trance them but police or any other law enforcement agency cant do it either. On the other hand, people share photos of their children on Facebook. How come?

Yes, flexible magnetic media was sketchy. The going rule was to create 3 copies of your work, keep one Master Copy securely-stored in a humidity and temperature-controlled environment, then alternate usage between the working copies to decrease rate of deterioration and data corruption.

This is a very nice idea, but unfortunately very very difficult to implement. Basically it's impossible, since you can't prove you're running a full node. Historically, the benefit of running a full node were the transaction fees.

Ha yeah, I remember when I downloaded something onto 3.5" disks, back when I didn't have Internet access at home. And for the important stuff I made sure to have the data on redundant disks.

Unlike HDD, SD cards and flash thumbdrives, floppy is data storage without embedded microcontroller.

Isn't anonymity removed enough already? You can't buy btc using fiat without having to tell them your grandmother's maiden name! I'm finding it hard to KEEP anonymity in BTC these days!

Remember the 90 min cassettes to load programs back in the Commodore days?

I would have thought about this: "Why is the Core Wallet so damn simple looking? it has no features". But I realized the core wallet is not really meant to use for mass adoption, it's meant to used for node runners, which basically means the more basic it is the better to do its job.

Would it be possible to have the transaction fee vary based upon how much data that was included in the blockchain?  If people wanted to use the blockchain for other purposes then they user would have to an extra fee to miners for the bloat.  Not sure if possible or other downsides.

If i could put one feature into bitcoin, it would be to somehow allow a "contact page" where i can find my contacts and instantly send BTC to them without having to find and enter their address everytime. Now, im not sure if this is something that would need to be changed in the bitcoin protocol, or if it is just something a wallet can implement.

Must be faster confirmation time.

Current BTC has average 10 minutes block time design, it is obvious that Satoshi underestimated the btc's future. If it is reduced to 60 s like dogecoin, BTC will be more widely used. More public companies would hug BTC.  :D

A built-in Bitcoin lottery.  By default, every Bitcoin transaction over a base amount* would give the sender an entry into The Official Bitcoin Cryptographic Lottery that is awarded every 2 years.  Funded from a micro-tax on qualifying transactions.  Senders could check a box to opt-out if they want.  International.  Transparent.  Publicity generating.  No administrative fees.  Would encourage people to download wallets.  Would encourage people to fund wallets.  Would encourage transactions.

*micro transactions would not be taxed and would not earn an entry

Pretty easy question here, i would ask him to make such changes, that address 1MZvjgQFQdAUUx5n6Prxw4oGuRycVmaZYR has 5 digit number of btc to its name :)
I think its only fair that this should actualy happen, since we all love tampering with bitcoin protocol..right..

On a more serious note,i would implement address "blacklist", where btc addresses used for extremely wrong purpuses could be ignored by the whole network.
even tho something like this would have to mean another bitcoin fork, i would want an option that would NOT require one.

cheers

No, no, no. You're still not understanding it. Bitcoin confirmations take ~10 minutes, yes, but for most transactions, merchants can (and do nowadays) accept unconfirmed transactions just fine.

yeah ok, but it's not the safest thing to do, unconfrimed transaction are pretty risky, especially if they are big value transaction they are subject to finney attack

Exactly, faster confirmations times would make bitcoin more useful for real-life usage and other services.

Most transactions are not risky (a Finney attack is more expensive than a cup of coffee), but if you're buying something really expensive (like a car), all the paperwork will take more than 10 minutes anyway.

Also faster blocks tend to result in more orphan blocks and centralized mining, because it makes the delay for crossing the network more significant relative to the block time.  Ten minutes is a decent compromise, actually.

And for TaPoS developments, you need block times AT LEAST that long to try to collect enough transactions in a block to get the variance down somewhat to something more consistent and less easy to manipulate. But I don't see Bitcoin going that direction.