Author Topic: Hard Problems in Cryptocurrency  (Read 4622 times)

0 Members and 1 Guest are viewing this topic.

Offline luckybit

  • Hero Member
  • *****
  • Posts: 2921
    • View Profile
  • BitShares: Luckybit

I don't think youve solved scalability.  Running Bts X on a massive storj based ledger would be extremely slow not lending itself to trading.
That is debatable. There is already a project running Bitcoin nodes in the cloud and it doesn't seem extremely slow. http://www.blockcypher.com/

You could end up being right but shouldn't we let the tests determine?

It's up for debate though.

Also, what would the shift to storj do for the economics?

I would hope it would be symbiosis, mutually beneficial. I would think it would be one of the primary reasons for mass adoption of Storj because I know I don't want to have to store a 100 GB Bitcoin blockchain. It's growing way too fast now and it's reaching the point where it has to be stored in the cloud to make economic sense in any case.

So if it's a centralized cloud or decentralized cloud I would expect it to be in the cloud. I would prefer it be the decentralized cloud.
https://metaexchange.info | Bitcoin<->Altcoin exchange | Instant | Safe | Low spreads

bitbro

  • Guest

I don't think youve solved scalability.  Running Bts X on a massive storj based ledger would be extremely slow not lending itself to trading.
That is debatable. There is already a project running Bitcoin nodes in the cloud and it doesn't seem extremely slow. http://www.blockcypher.com/

You could end up being right but shouldn't we let the tests determine?

It's up for debate though.

Also, what would the shift to storj do for the economics? 

Offline luckybit

  • Hero Member
  • *****
  • Posts: 2921
    • View Profile
  • BitShares: Luckybit
I don't think youve solved scalability.  Running Bts X on a massive storj based ledger would be extremely slow not lending itself to trading.
That is debatable. There is already a project running Bitcoin nodes in the cloud and it doesn't seem extremely slow. http://www.blockcypher.com/

You could end up being right but shouldn't we let the tests determine?

It's up for debate though. I think with Storj because they aren't like SAFE Network the Storj software will probably choose the cloud which is fastest and cheapest. SAFE Network on the other hand can't connect to another cloud at the backend if it's slow.
« Last Edit: July 28, 2014, 01:20:19 am by luckybit »
https://metaexchange.info | Bitcoin<->Altcoin exchange | Instant | Safe | Low spreads

bitbro

  • Guest
I don't think youve solved scalability.  Running Bts X on a massive storj based ledger would be extremely slow not lending itself to trading.

Offline luckybit

  • Hero Member
  • *****
  • Posts: 2921
    • View Profile
  • BitShares: Luckybit
"searching"?
What I wanted to ask was: Is the fact that it is deterministic the reason that forks cant happen unless there is collusion of at least 50+% of the delegates?

Exactly.  Forks only happen due to bugs or an attack.
If the fork is there once, how likely is it that both chains have the same length? What does it depend whether they have the same length?

I will finish this https://blog.ethereum.org/2014/07/05/stake/ and discuss it with the Ethereum community.
We need more inter community discussions that are constructive, rational and neutral, not the typical reflexive camp bashing.
I made a post on the NXT forum with the goal of coming to a consensus about ad-/disadvantages of both systems with the NXT community  https://nxtforum.org/general-discussion/nxt-pos-vs-bitshares-dpos/msg70304/#msg70304
I made an updated summary in the OP. If I did not miss anything all the contributors to the discussion on the NXT forum do not disagree with my assessment.
There are two questions left that are contradiction (to a degree) what you said about NXT. Maybe you find time to comment on them:
- just saw that you already commented on the one regarding the operational costs of a delegate.
- the other one is: https://bitsharestalk.org/index.php?topic=6172.msg82399#msg82399

What we need to do is track problems which have been solved and rate the different solutions.

Scalability solved

For example scalability may be solved by using decentralized clouds. If SAFE Network or Storj work then full nodes could be hosted there. I think that solution is an alternative to the Bitshares X solution of multiple blockchains. Both solve scalability and both are decentralized so there is room to debate each solution's pros and cons.

Volatility solved

Volatility is also solved. There are two solutions which could work. The contract for difference market pegging which Bitshares and Ethereum will be using and the escrow based solution that Mastercoin will be using. Either solution could work and both have their risks but it seems the market pegging solution is more popular and easier to implement. We should compare and contrast these solutions.

Consensus solved

Proof of Work is now becoming obsolete as a means of providing security. Proof of Stake has proven itself adequate for security. There is some debate as to whether or not Proof of Stake is optimal for distribution and marketing as it seems the Proof of Work currencies have mechanisms which generate free marketing which Proof of Stake currencies do not have.

Delegated Proof of Stake and Transparent Forging are in competition for title of "state of the art". Delegated Proof of Stake seems more likely to scale than Transparent Forging at this point but we will only really find out when they go head to head. It is likely that both paradigms will work.

Proof of Work still in my opinion has useful purposes if the work itself is beneficial. For instance if it uses BOINC then the work might be to fold proteins. In the case of Primecoin it was to find prime numbers. I think if there is some creativity with regard ot Proof of Work it will still survive and have a role but not in the form we have now with ASICs hashing SHA-256 creating useless numbers precious numbers. The ideal proof of work will do useful calculations to generate precious numbers.

Autonomous growth and sustainable development partially solved

One solution for autonomous growth and development is the Bitshares delegate approach where delegates use dilution aka inflation as a means of funding development. This solution has it's pros and cons but really it's not all that different from what has been discussed previously except for the dilution factor which isn't politically popular. The Mastercoin approach was to set aside a certain amount of coins which get released over time so that it could be factored in at the moment of purchase. This doesn't have as much political consequences because people already factored in the developer Mastercoins from the beginning.

Another approach is to use smart bounties which I advocated a while back but that approach would require smart contracts and a level of complexity which may only be practical to try with Ethereum and even in that case it required something similar to delegates (DAC operators). So far all of these are only semi-autonomous which means humans have to vote in some way, rely on delegates, etc.

These are just some solutions I could think of which are emerging. The main problem is the media and majority of the community cannot keep track of all this stuff because its not in one place. If someone has a solution to a problem it's hidden in a forum somewhere in some community and unless you're looking in all of them you might miss something.

Random number generation without mining

The only way I could think of to do something like this is with an external data feed. Delegates could act as oracles who could supply the initial entropy in the form of trusted entropy data feeds. Then do the algorithms on those feeds to produce seeds for random number generators. This is not a hard problem if you can trust the delegates to generate the initial entropy and can somehow chain the numbers together on the blockchain using transactions (like with the Lottery DACs).

But collusion is a possibility which makes it hard. It's also hard because any flaw in the process could make the whole chain of numbers deterministic and ruin it.

« Last Edit: July 28, 2014, 01:16:53 am by luckybit »
https://metaexchange.info | Bitcoin<->Altcoin exchange | Instant | Safe | Low spreads

Offline santaclause102

  • Hero Member
  • *****
  • Posts: 2486
    • View Profile
"searching"?
What I wanted to ask was: Is the fact that it is deterministic the reason that forks cant happen unless there is collusion of at least 50+% of the delegates?

Exactly.  Forks only happen due to bugs or an attack.
If the fork is there once, how likely is it that both chains have the same length? What does it depend on whether they have the same length?

I will finish this https://blog.ethereum.org/2014/07/05/stake/ and discuss it with the Ethereum community.
We need more inter community discussions that are constructive, rational and neutral, not the typical reflexive camp bashing.
I made a post on the NXT forum with the goal of coming to a consensus about ad-/disadvantages of both systems with the NXT community  https://nxtforum.org/general-discussion/nxt-pos-vs-bitshares-dpos/msg70304/#msg70304
I made an updated summary in the OP. If I did not miss anything all the contributors to the discussion on the NXT forum do not disagree with my assessment.
There are two questions left that are contradiction (to a degree) what you said about NXT. Maybe you find time to comment on them:
- just saw that you already commented on the one regarding the operational costs of a delegate.
- the other one is: https://bitsharestalk.org/index.php?topic=6172.msg82399#msg82399
« Last Edit: July 30, 2014, 08:03:12 am by delulo »

Offline bytemaster

"searching"?
What I wanted to ask was: Is the fact that it is deterministic the reason that forks cant happen unless there is collusion of at least 50+% of the delegates?

Exactly.  Forks only happen due to bugs or an attack. 
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline santaclause102

  • Hero Member
  • *****
  • Posts: 2486
    • View Profile
"searching"?
What I wanted to ask was: Is the fact that it is deterministic the reason that forks cant happen unless there is collusion of at least 50+% of the delegates?

Offline bytemaster

Producing blocks is the term we use... no searching is necessary.  It is deterministic.
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline santaclause102

  • Hero Member
  • *****
  • Posts: 2486
    • View Profile
At min 38:30 he says that if a fork happens there is no reason to decide for one of the forks. With POW miners have to choose which fork to support so only one of the chains will survive. That would be one advantage of seperating shares/votes into mining votes and holding-only-shares.
How does DPOS solve this (compared to existing POS Systems / NXT)?

With DPOS there can be no forks unless the delegates collude in which case they would be fired and 51% of delegates have signed there is no problem.
Quote
With DPOS there can be no forks unless the delegates collude
That means collude with at least 51% if I am right. Reasons: Finding blocks (what term is actually used with DPOS that is equivalent to mining / forging?) is deterministic -> there can not be two blocks found at a time.
Valid?
« Last Edit: July 27, 2014, 09:09:02 pm by delulo »

Offline bytemaster

Underneath DPOS is TAPOS meaning shareholder votes decide longer chain (more block-share-votes)
As according to BM there is no tapos anymore and DPOS made TAPOS irrelevant. How is the case described above solved with the DPOS only? Would the remaining delegates have to "manually" talk to each other?
Quote
he says that if a fork happens there is no reason to decide for one of the forks. With POW miners have to choose which fork to support so only one of the chains will survive. That would be one advantage of seperating shares/votes into mining votes and holding-only-shares.

If a fork happens the delegates will pick the longest one or they don't get paid.  They cannot pick both because they will be fired for signing two blocks at the same time stamp.    Only one fork can be the longest and delegates cannot 'make up for lost time' by back-dating block production.  So it is always clear.
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline santaclause102

  • Hero Member
  • *****
  • Posts: 2486
    • View Profile
Underneath DPOS is TAPOS meaning shareholder votes decide longer chain (more block-share-votes)
As according to BM there is no tapos anymore and DPOS made TAPOS irrelevant. How is the case described above solved with the DPOS only? Would the remaining delegates have to "manually" talk to each other?

Offline toast

  • Hero Member
  • *****
  • Posts: 4001
    • View Profile
  • BitShares: nikolai
Underneath DPOS is TAPOS meaning shareholder votes decide longer chain (more block-share-votes)
Do not use this post as information for making any important decisions. The only agreements I ever make are informal and non-binding. Take the same precautions as when dealing with a compromised account, scammer, sockpuppet, etc.

Offline santaclause102

  • Hero Member
  • *****
  • Posts: 2486
    • View Profile
At min 38:30 he says that if a fork happens there is no reason to decide for one of the forks. With POW miners have to choose which fork to support so only one of the chains will survive. That would be one advantage of seperating shares/votes into mining votes and holding-only-shares.
How does DPOS solve this (compared to existing POS Systems / NXT)?

With DPOS there can be no forks unless the delegates collude in which case they would be fired and 51% of delegates have signed there is no problem.
Quote
With DPOS there can be no forks unless the delegates collude
True, that makes sense. But I didnt get the other part: Take the case that it would still happen -> then the delegate(s) would get fired and then there are still two chains right?

Offline bytemaster

At min 38:30 he says that if a fork happens there is no reason to decide for one of the forks. With POW miners have to choose which fork to support so only one of the chains will survive. That would be one advantage of seperating shares/votes into mining votes and holding-only-shares.
How does DPOS solve this (compared to existing POS Systems / NXT)?

With DPOS there can be no forks unless the delegates collude in which case they would be fired and 51% of delegates have signed there is no problem.

For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.