Bitcoin 20MB Fork

[bambou]

A mix between pro and DGAF as I don't care but it will need to happen eventually so why not now?

Not a tech-dude here, but i, on the contrary, feel that this is somehow of a crucial matter.
Personally I like bitcoin the way it is now.

I just started reading this thread, but let me show you my point of view regarding this.

In order for bitcoin to succeed as a protocol, not as a currency it needs services running on top of it. Either centralized or decentralized bitcoin is here for everyone(not only for those paying the higher fees)! A first service would be the Lighthouse project which is supposed to be a decentralized way of raising funds Kickstarter-style but which is already limited by the 1MB block size cap (see https://www.vinumeris.com/lighthouse/faq#max-pledges). Having a 1MB block size limit hinders innovation and tech development in my view and don't forget that I'm also a miner, but I agree and support Gavin in this matter. I view the blockchain as a decentralized and neutral informational highway. Let data flow freely if you want a higher value of the whole ecosystem. Let's find solutions to the potential problems like blockchain bloat by innovating, not by rejecting and imposing limits.

Consider bitcoin as the internet. Everyone is using it whether he is poor (free wi-fi) or wealthy (Google, Facebook etc). There are a lot of great things built on free services (twitter, instagram, facebook etc). I bet that you wouldn't like it if your ISP would have a 100MB/s download speed and would action it for the highest bidder. Forcing fees hinders evolution!

So please let's fork already so we can see some real services coming to life!

hum, you should keep reading,

i'm not sure a full consensus is reachable at that point.
i hear both sides but it is the centralization part that bugs me.
im a long term speculator and thus more incline to keep my coins 'out of it' until the market settles.

anyhow you better start packing on popcorn, bitcoin is such a great adventure!

[BitUsher]

A first service would be the Lighthouse project which is supposed to be a decentralized way of raising funds Kickstarter-style but which is already limited by the 1MB block size cap (see https://www.vinumeris.com/lighthouse/faq#max-pledges). Having a 1MB block size limit hinders innovation and tech development in my view

Excellent point I never thought of. This is a perfect example of how having a 1MB cap hurts decentralization by limiting projects like lighthouse, open bazaar and twister.

The fact that lighthouse only can allow for 684 pledges with the possibility of up to 1k means that it is severely limited. Popular crowdfunding platforms will sometimes get 20-50k pledges for a fundraiser or project .

So I suppose a 20MB block will allow for up to 20k pledges with optimizations which is much better.

[zimmah]

Upgrade the limit sooner rather than later.

Don't be the one to say things like these famous quotes:

  "We will never make a 32-bit operating system." -- Bill Gates, speaking at the launch of MSX in 1983

"Almost all of the many predictions now being made about 1996 hinge on the Internet's continuing exponential growth. But I predict the Internet will soon go spectacularly supernova and in 1996 catastrophically collapse." -- Robert Metcalfe, 3Com founder and inventor of Ethernet, 1995

 “There will never be a bigger plane built.” — A Boeing engineer, after the first flight of the 247, a twin engine plane that holds ten people

Don't assume 1mb is enough just because it is now. If another boom comes, we are likely going to shattered 1mb, and we don't want it to crash just because the network can't handle it do we?

Better to be prepared.

[hdbuck]

This thread makes the centralisation issue clearer:
https://bitcointalk.org/index.php?topic=144895.0

lol this is from feb 2013.. ^^

[zimmah]

blockchain security will be weakend with this a lot

(less fees for miner which is important after rewards decrease, larger chain, more central mining) = blockchain security to your anus

How do you work that one out? 20x the transactions means 20x the fees.

On the fence, low block size encourages payment infrastructures to build around the limitations but the 10min interval is enough incentive for that anyway and pruning can keep the bloat down but hard drive space is cheap even at 1TB a year anyway... if Gavin reckons go for it then +1, his anorak is much cooler than mine

Since more transactions will fit, the fee per transaction will be lower because most of the time there is room left in the block to include your transaction. But even though the fee per transaction is lower, the total sum of fees is larger because there are more transactions.

[inBitweTrust]

This thread makes the centralisation issue clearer:
https://bitcointalk.org/index.php?topic=144895.0

lol this is from feb 2013.. ^^

I remember reading that very interesting thread, excellent read back in the day .

This is why I laugh at all the hero members making claims that Gavin is trying to rush this through and it is too quick when we have been discussing it for years.

Perhaps there may be a compromise(not really because we also want decentralization too) made and we implement some change in the hardfork which addresses privacy or centralization concerns at the same time?

[altcoin hitler]

let's be honest: Gavins' proposal is bullshit and there needs to be other solutions found

[Peter R]

Since more transactions will fit, the fee per transaction will be lower because most of the time there is room left in the block to include your transaction. But even though the fee per transaction is lower, the total sum of fees is larger because there are more transactions.

Is this true?  We can try to answer this question empirically using the data we already have.  

The chart below shows the correlation that exists between the number of transactions per day and the total daily fees paid to the bitcoin miners.  This reveals that indeed the total fees have in fact increased with the increased transaction volume.  



We can take our analysis further.  Let F be the daily fees paid to bitcoin miners, and let N be the number of transactions per day.  Performing a least-squares regression between log F and log N reveals that the total daily fees have actually grown as N raised to the power of 2.7.  What this means is that the fees have historically grown much faster than the number of transactions per day; a doubling of the number of transactions has on average resulted in a 6.5x increase in the total daily fees paid to the miners.  The correlation between the two time series, log F and log N, is 0.92 (strong coupling).

If this correlation continues to hold moving forward in time, the total fees collected by miners will increase sharply with increased transaction volume.  (EDIT: I'm not suggesting that this correlation will necessarily continue--I'm just describing what's happened so far and projecting that into the future.)

[Peter R]

Another interesting correlation that exists is the one between Bitcoin's market cap and the number of transactions per day^1,2.  If the correlation continues to hold, it suggests that the 1 MB blocksize limit may also limit Bitcoin's future market cap.   



¹Actually between the square of the number of transactions and the market cap (Metcalfe's Law).

²I've plotted the number of transactions excluding popular addresses to remove the on-chain gambling bubble of 2012/2013.  This has only a minor effect on the 2014/2015 portion of the data.

[tvbcof]

let's be honest: Gavins' proposal is bullshit and there needs to be other solutions found

An amazing assortment of the most respected and active developers are working on 'sidechains'.  It looks to me like this fits the bill.  I couldn't be more delighted with things going in this direction since it seemed like a logical way of scaling to me since shortly after I read the whitepaper and saw the problem.

Actually I am even more interested in 'treechains' in principle than I am in sidechains, but even if that can be made to work at all it still seems entirely unlikely that it could every be put under the hood of Bitcoin.  Certainly it could form the functional core of some of the sidechain attempts though.  I'll look forward to playing with some of these.

[viboracecata]

let the block size increase, I've hard to be tolerant to the database resident in c harddisk, I thought the fork is necessary, otherwise, the nodes number will decrease in long time

[tvbcof]

Since more transactions will fit, the fee per transaction will be lower because most of the time there is room left in the block to include your transaction. But even though the fee per transaction is lower, the total sum of fees is larger because there are more transactions.

Is this true?  Rather than guessing, we can try to answer this question empirically using the data we already have.  

The chart below shows the correlation that exists between the number of transactions per day and the total daily fees paid to the bitcoin miners.  This reveals that indeed the total fees have in fact increased with the increased transaction volume.  



We can take our analysis further.  Let F be the daily fees paid to bitcoin miners, and let N be the number of transactions per day.  Performing a least-squares regression between log F and log N reveals that the total daily fees have actually grown as N raised to the power of 2.7.  What this means is that the fees have historically grown much faster than the number of transactions per day; a doubling of the number of transactions has on average resulted in a 6.5x increase in the total daily fees paid to the miners.  The correlation between the two time series, log F and log N, is 0.92 (strong coupling).

If this correlation continues to hold moving forward in time, the total fees collected by miners will increase sharply with increased transaction volume.    

Even with tx/day logarithmic it doesn't look like a linear fit is very appropriate.  Looks like the whole upper half is flat which would negate your point completely as I understand it.  In fact, I don't really see a good reason to plot this on the log scale in either axis.  Plot it without logs, with data for a reasonably modern time-frame (say, after the unspendable output stuff went in), and knock off the absurd outliers (the only reason to go logarithmic in the first place) and let's see what she looks like.

[Peter R]

Even with tx/day logarithmic it doesn't look like a linear fit is very appropriate.  Looks like the whole upper half is flat which would negate your point completely as I understand it.  

I don't expect the F ~ N^2.7 relation to hold forever; I think the coefficient in the exponent will decrease over time and perhaps that's what we're already seeing.  I do expect the correlation to remain positive, however.  But that's speculation about the future.  I wanted to look at the facts as of today.  And as of today, the fees have, on average, increased faster than the square of the number of transactions [the correlation is 0.92 (strong)].

In fact, I don't really see a good reason to plot this on the log scale in either axis.

The total fees range from sub $1 to $10,000+.  On a linear scale, only the recent data over the last few years is clearly visible (and indeed looks flatter as you point out).  But the purpose of this exercise was to inspect the relationship between F and N over as long a time frame as possible (why not use all the data?).  Once again, this doesn't preclude a change in the relationship between fees and transaction volume moving forward in time.  

[benjamindees]

Many of the arguments being presented are so poor that I feel compelled to argue against my position for the sake of elevating the discussion.

+1

[CoinCidental]

sorry for the stupid question but if a block today is 1mb, after the hard fork, will it still be 1mb? or will they all be 20?

I mean, if the size automatically adjusts to the need, then it will not create a larger blockchain at all, because the transactions are there and need to be processed sooner or later

Bigger size and lower fees lead to more traffic and thus higher blocksize. If you implement it, the space will be used. The chain also becomes vulnerable to attacks by spam. Malicious parties can bloat the chain to the max size and they will.

malicious parties could overbloat the chain no matter what size it is ,just get 2 wallets and send transfers back and fourth to yourself 24/7
or is that not possible ?

[zimmah]

sorry for the stupid question but if a block today is 1mb, after the hard fork, will it still be 1mb? or will they all be 20?

I mean, if the size automatically adjusts to the need, then it will not create a larger blockchain at all, because the transactions are there and need to be processed sooner or later

Bigger size and lower fees lead to more traffic and thus higher blocksize. If you implement it, the space will be used. The chain also becomes vulnerable to attacks by spam. Malicious parties can bloat the chain to the max size and they will.

malicious parties could overbloat the chain no matter what size it is ,just get 2 wallets and send transfers back and fourth to yourself 24/7
or is that not possible ?

sure you can, but this is unlikely to happen, and limiting the block size just because it could potentially be used against us is asking for disaster.

We're just one bubble away from needing the extra space, and what do we do when that bubble comes?

If the blocksize becomes a bottleneck at just the time where people start gaining trust in bitcoin, that could be the end of bitcoin.

[hdbuck]

This thread makes the centralisation issue clearer:
https://bitcointalk.org/index.php?topic=144895.0

lol this is from feb 2013.. ^^

I remember reading that very interesting thread, excellent read back in the day .

This is why I laugh at all the hero members making claims that Gavin is trying to rush this through and it is too quick when we have been discussing it for years.

Perhaps there may be a compromise(not really because we also want decentralization too) made and we implement some change in the hardfork which addresses privacy or centralization concerns at the same time?

3) Centralization fears are exaggerated and can be reversed in the future. Mining pool centralization is a much bigger fear regardless. I'm not to comfortable with what essentially amounts to a 3 of 5 multisig to secure us against a 51% attack.

please explain

[phillipsjk]

This thread makes the centralisation issue clearer:
https://bitcointalk.org/index.php?topic=144895.0

lol this is from feb 2013.. ^^

I remember reading that very interesting thread, excellent read back in the day .

This is why I laugh at all the hero members making claims that Gavin is trying to rush this through and it is too quick when we have been discussing it for years.

Perhaps there may be a compromise(not really because we also want decentralization too) made and we implement some change in the hardfork which addresses privacy or centralization concerns at the same time?

Two years is not a long time. As far as I can tell, the issues are still exactly the same.

As mentioned up-thread, by current ISP is now offering new users less bandwidth than they did two years ago. One major reason is that they are rolling out "free" wireless access points that are obviously not garnering as many new users as they had hoped. (The cheapest Internet plan is $50/month, while the average cell phone plan in Canada is about $40/month)

I am thinking that staying with 1MB blocks would not be a horrible outcome. As it is, Bitcoin is not secure for the average user in part because privacy is very fragile.

If someone can identify a user’s bitcoin addresses—in Ulbricht’s case, by seizing the laptop he was actively using at the moment of his arrest—then they can often be used to trace his or her transactions.

Coinjoin helps, but I have my doubts that it will ever be used for the majority of transactions. The only Bitcoin wallet that I am aware of that tries to use coinjoin by default is the highly experimental Darkwallet (not to be confused with Darkcoin)

One of the "scamcoins" (general term for alts mainly designed to enrich the developer) actually came out with some strong privacy features in order to hide an 80% pre-mine. Seriously, the white-paper on https://cryptonote.org/ is worth the read.

As I mentioned in the other thread, Monero appears to be the generally-accepted community fork of this. Cryptonote-based currencies take advantage of something called "ring signatures" in order to allow you to automatically do coinjoin-like operations with an arbitrary number of addresses. Not only that, but transactions are automatically broken down into common units, so tracing by quantity is mostly pointless as well. Edit: "stealth address" appear to be implemented as well. And, most relevant to this discussion: the block-size automatically adjusts.

   A good example of a hardcoded limit change leading to disastrous consequences is the block
size limit set to 250kb[1] . This limit was sufficient to hold about 10000 standard transactions. In
early 2013, this limit had almost been reached and an agreement was reached to increase the
limit. The change was implemented in wallet version 0.8 and ended with a 24-blocks chain split
and a successful double-spend attack [9]. While the bug was not in the Bitcoin protocol, but
rather in the database engine it could have been easily caught by a simple stress test if there was
no artificially introduced block size limit.

An extremely popular version of Bitcoin may well be a disaster for personal privacy: since the block-chain is public. Coinjoin operations can help, but are not really encouraged. It is easier to use a privacy-compromising "well known" address.

If the "masses" were moved to something like Monero instead, I have more faith that the privacy features would be used. MPcoin and Monero can co-exist as experiments in how block-size impacts both adoption and mining fees. Maybe the discussions from two years ago will turn out to be true. In that case, Monero will just be another payment system controlled by cartels. Maybe, the UN security council members will agree that "crypto-currencies" are of strategic importance akin to nuclear weapons. In that case, Monero and GavinCoin (20MB block Bitcoin) will be regulated out of existence. The more agile MPcoin (1MB block Bitcoin) would (hopefully) survive.

[R2D221]

The more Agile MPcoin would (hopefully) survive.

I'm sorry, I'm kind of lost in this discussion, but what is MPcoin?

[tvbcof]

...  Maybe, the UN security council members will agree that "crypto-currencies" are of strategic importance akin to nuclear weapons. In that case, Monero and GavinCoin will be regulated out of existence. The more Agile MPcoin would (hopefully) survive.

The way I look at it:

 - As long as the mainstream systems hold up and tolerate cash as legal tender and gold/silver is tolerated as not illegal (which could be indefinitely) crypto-currencies are cool and neat and all that, but mainly a toy.

 - If there is a failure in mainstream-land and cryto-currencies are able to step in and fill a non-trivial void then it is almost a sure thing that they will be attacked by interests vested in the mainstream and whatever 'acceptable' instruments step up to the plate (e.g., SDR's or whatever.)  These interests generally control today's policies and there is little reason to imagine that changing.

 - Ergo, an 'agile' (as you put it) crypto-currency is the only thing worth investing in.  Bitcoin still has the potential to be such a thing.  Barely.  And only because development and hardening could and would happen very quickly if there were an actual need.  Once it bloats all bets are off.