Bitcoin Forum
May 05, 2024, 10:30:37 AM *
News: Latest Bitcoin Core release: 27.0 [Torrent]
 
   Home   Help Search Login Register More  
Pages: « 1 [2]  All
  Print  
Author Topic: fuck this shit, I want my own blockchain!  (Read 3319 times)
killerstorm (OP)
Legendary
*
Offline Offline

Activity: 1022
Merit: 1015



View Profile
July 28, 2012, 08:36:30 AM
 #21

@killerstorm
Have you read this ?
https://bitcointalk.org/index.php?topic=79837.0;all
--------------------
Also about 3.6.1 and 3.6.2 :
Given that bounty will be a constant, you can completely eliminate tx fees

With dynamic block frequency, however, there is an alternative approach: block weight is computed via a rather complex function which takes into account difficulty, quality of transactions and block size.

If we have two blocks with same difficulty, one with 'good transactions' would win, so miners have an incentive to include them.

This function can also automatically filter out spam by penalizing large blocks.

The only downside is that it is really hard to implement this without sacrificing security. I.e. a double spend attack might exploit some weakness in function to make attack cheaper.

However it is good against 51% attack which simply produces empty blocks: those blocsk would have low weight making attack much more expensive. To block normal economic activity attacker needs a large supply of old coins to fake 'good transactions'.

Chromia: a better dapp platform
1714905037
Hero Member
*
Offline Offline

Posts: 1714905037

View Profile Personal Message (Offline)

Ignore
1714905037
Reply with quote  #2

1714905037
Report to moderator
1714905037
Hero Member
*
Offline Offline

Posts: 1714905037

View Profile Personal Message (Offline)

Ignore
1714905037
Reply with quote  #2

1714905037
Report to moderator
1714905037
Hero Member
*
Offline Offline

Posts: 1714905037

View Profile Personal Message (Offline)

Ignore
1714905037
Reply with quote  #2

1714905037
Report to moderator
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
1714905037
Hero Member
*
Offline Offline

Posts: 1714905037

View Profile Personal Message (Offline)

Ignore
1714905037
Reply with quote  #2

1714905037
Report to moderator
1714905037
Hero Member
*
Offline Offline

Posts: 1714905037

View Profile Personal Message (Offline)

Ignore
1714905037
Reply with quote  #2

1714905037
Report to moderator
killerstorm (OP)
Legendary
*
Offline Offline

Activity: 1022
Merit: 1015



View Profile
July 28, 2012, 09:19:25 AM
 #22

Do you plan to implement some sort of pruning/reducing blockchain's size from the start ?

No, I don't see what value it brings from the start. And I don't see any solution which actually solves scalability problem. Pruning schemes only buy you some time.

Chromia: a better dapp platform
Etlase2
Hero Member
*****
Offline Offline

Activity: 798
Merit: 1000


View Profile
July 28, 2012, 10:36:24 AM
 #23

No, I don't see what value it brings from the start. And I don't see any solution which actually solves scalability problem. Pruning schemes only buy you some time.

Pruning, as it has to be implemented for bitcoin, currently does not even buy any time. Unless you trust someone, you will always have to download the entire history of bitcoin before becoming a full client. Now, down the road, the developers might at some point say hey, this is ridiculous, and attempt to push a fork-inducing change that will once again be decided by a few people in the community apart from the rest. See the BIP16/17 "scandal" if you are unfamiliar with what I am referring to. The solution is an account based ledger rather than a transaction ledger.

NASDAQEnema
Full Member
***
Offline Offline

Activity: 182
Merit: 100


View Profile
July 28, 2012, 11:39:05 PM
 #24

Account based leads to a whole bunch of vulnerabilities. Transaction based is inherently resistant to common problems in the physical world.

What should be done is to store a compressed version of the last 1024 blocks in a new block and don't download intervening blocks.

If you feel Universe has trolled you exclusively, please donate to Emergency Butthurt Support Fund:
1Jv4wa1w4Le4Ku9MZRxcobnDFzAUF9aotH
Proceeds go to Emergency Butthurt Escape Pod none of you will be allowed to use. If you have read this far, you must pay Emergency Butthurt Internet Tax.
Etlase2
Hero Member
*****
Offline Offline

Activity: 798
Merit: 1000


View Profile
July 29, 2012, 07:28:10 AM
 #25

Account based leads to a whole bunch of vulnerabilities.

Care to elaborate?

Quote
Transaction based is inherently resistant to common problems in the physical world.

Uhh cryptocurrencies are digital, fyi. Tongue And I don't see anything inherent.

Quote
What should be done is to store a compressed version of the last 1024 blocks in a new block and don't download intervening blocks.

This is similar to what an account based ledger would do, except you'd have to store all of the unspent transactions instead of account values.

killerstorm (OP)
Legendary
*
Offline Offline

Activity: 1022
Merit: 1015



View Profile
July 29, 2012, 07:49:15 AM
 #26

This is similar to what an account based ledger would do, except you'd have to store all of the unspent transactions instead of account values.

Why do you think that there are significantly fewer account values then there are unspent transactions?

What prevents person from starting a new account each time he receives money?

Why wouldn't he do that if that improves privacy?

You can introduce some account-creation fee, of course, thus encouraging people to reuse them, but it won't be like Bitcoin at all. It's a very radical, fundamental change.

If you're willing to do that, you can instead force txn defrag by making only outputs in last N blocks spendable. I.e. people will be forced to send money to themselves to save it. (It has numerous advantages: it fixes "grandfather's wallet problem", prevents deflation and subsidizes fees, thus encouraging miners.)

This requires only a tiny change to Bitcoin code. The only problem is that it will be hugely unpopular, but so will be switch to accounts, I think.


Chromia: a better dapp platform
killerstorm (OP)
Legendary
*
Offline Offline

Activity: 1022
Merit: 1015



View Profile
July 29, 2012, 07:58:23 AM
 #27

I believe that the right solution is 'divide and conquer', i.e. instead of having a huge-ass list of all transactions or list of all accounts we should split it into manageable parts and store on different servers. Like in a DHT, e.g. Kademlia. (Used in p2p file sharing, for example.)

Unfortunately DHTs do not provide features which are required for safe transactions, at least ones we know. But it doesn't mean that such technology won't be developed in future.

Chromia: a better dapp platform
Etlase2
Hero Member
*****
Offline Offline

Activity: 798
Merit: 1000


View Profile
July 29, 2012, 08:05:40 AM
Last edit: July 29, 2012, 10:35:51 AM by Etlase2
 #28

Why do you think that there are significantly fewer account values then there are unspent transactions?

Because, at least with bitcoin, you are encouraged to use a different address for every transaction. And even if you don't, each transaction still creates a transaction in the ledger with unspent outputs. If one person doesn't care as much about anonymity, they can receive 50 transactions and thus have 50 pieces of data in the log whereas an account ledger would still have 1. And when he goes to spend all that money, 50 transactions worth of outputs will need to be combined. Then those outputs could be pruned, but the intervening bandwidth usage is still heavy.

Quote
What prevents person from starting a new account each time he receives money?

Why wouldn't he do that if that improves privacy?

You can introduce some account-creation fee, of course, thus encouraging people to reuse them, but it won't be like Bitcoin at all. It's a very radical, fundamental change.

People have varying needs for privacy but bitcoin forces a square peg into a round hole, basically. An account creation fee is exactly the answer and it does not have to be expensive at all, just enough to say hey, if you want better privacy, you will have to pay a small premium. You are demanding to use more network resources, so it only makes sense. In the future when bitcoin mining rewards are low, combining transactions is going to be expensive, and you are not going to have a choice in the matter.

Quote
If you're willing to do that, you can instead force txn defrag by making only outputs in last N blocks spendable. I.e. people will be forced to send money to themselves to save it. (It has numerous advantages: it fixes "grandfather's wallet problem", prevents deflation and subsidizes fees, thus encouraging miners.)

Yeah good luck convincing bitcoin users who believe that the smallest transaction must be stored on thousands or tens of thousands of computers for all time because money is sacred and everyone should share the burden. Huh

I believe that the right solution is 'divide and conquer', i.e. instead of having a huge-ass list of all transactions or list of all accounts we should split it into manageable parts and store on different servers. Like in a DHT, e.g. Kademlia. (Used in p2p file sharing, for example.)

With the account ledger system, assuming elliptic curve DSAs, you're looking at 100-150 bytes per account. A terabyte could hold the world's accounting.

PS - I forgot to mention that individual coins cannot be traced using an account ledger, so complex coin mixers are not required to "clean" coins.

killerstorm (OP)
Legendary
*
Offline Offline

Activity: 1022
Merit: 1015



View Profile
July 29, 2012, 11:32:42 AM
 #29

What about name of the new coin ?
BBQcoin sounds funny, but this was a joke. right ?!

More-or-less, I actually considered taking over a dead block chain instead of starting a new one, but now taking over bbqcoin is pretty much pointless.

But with this discussion list of potential features got really long and it would take quite a lot of time to make it into a coherent vision and implement them.

However, I now got an offer to take over maintenance of a dead blockchain, one a bit more established than bbqcoin Smiley

So my current plan is to start developing relatively tame features in an established blockchain, and make a new one with radical features later, when/if I'll get a coherent vision.

But I need to evaluate feasibility of reviving of an old blockchain...

As for new block chain name, I haven't settled of something particular yet, but I kinda like "notacoin". I.e. "Not A Coin". Quite in GNU tradition, I think. It's supposed to mean "It's not a coin, it's better" or "This is not a currency you're looking for". Also can be interpreted as "Nota-Coin", as in "notary".

Chromia: a better dapp platform
Etlase2
Hero Member
*****
Offline Offline

Activity: 798
Merit: 1000


View Profile
July 29, 2012, 01:25:18 PM
 #30

So my current plan is to start developing relatively tame features in an established blockchain, and make a new one with radical features later, when/if I'll get a coherent vision.

I hate to sound like a broken record, but I'd like to hear your input on decrits/encoin. I have clear pseudo-code for a lot of it written in my head, but actually taking it to task is a whole nother story.

Quote
But I need to evaluate feasibility of reviving of an old blockchain...

Not in any way worth it, imo. Everyone will have to download new software if you break the chain, so why not just release a new coin.

Pages: « 1 [2]  All
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!