Bitcoin Forum

NaN.

After sitting on coins for 90d you make 1% per year mining stake blocks.  Stake blocks begin to be acquired after 30d.

http://ppcoin.org/static/ppcoin-paper.pdf

The paper can do a better job explaining it than anyone else

That's odd.  What are they encouraging by doing this?  Save more?

Actually, the paper seldom outlines any of the protocol specifications.  Rather, it just discusses the abstract theory behind the coin.

No, eventually PoW mining will end and blocks will only be mined by PoS.  PoW is intended to give the initial distribution of coins to the network.  The 1% per annum is a tiny incentive to keep people mining coins so stake blocks are still generated.  It's been something that bothered me about the chain from the get-go, since it's not a particularly big incentive to let your coins sit and do nothing for 30-90d +, but Sunny King thinks it will work.

It explained coinstake and why hashing power is not needed for network security just fine to me. For specifics on implementation, one can look at the code on github

That's interesting.  I think I get it; by ending PoW mining, people will no longer have to burn loads of electricity and burnt computers, and instead, all that they got through the mining (and selling) will then be used to "mine" the rest.  So do the coins get automatically added to your wallet based on how many you have at the time?  And if you move them at all, does the counter reset?  Very unusual.  And is it as secure as the PoW method?

The paper fails to explain a number of key characteristics of the chain, such as selection of the coin stake returns, the linear increase in coin stake return from 30 to 90d, why those timelines were selected, how the chain is designed to prevent PoS spamming, etc.  If you're one of the new devs, you should start talking about this stuff publicly.  As far as I'm concerned, the original PPC paper is far from technical and fails to address many of the potential vulnerabilities and shortcomings of the currency.  It's been my gripe since day one that the original paper borders on musings.

Correct.  PoS blocks are automatically mined starting with 30d of coin age unless you tell the daemon or qt client not to.  If you move them, it resets your coin age and you have to wait another 30d.  No one knows if it's as secure as the PoW method yet.

I know selection of what coins to spend will be handled by multiple wallets, so one would have a wallet for spending and another for saving up coinstake. Anyway, this is something I could certainly work on in the near future if you would find it helpful. Aside from the things you just listed, are there any other specific questions you have? If you have a list, feel free to PM me

Just like PoW mining there is some luck involved.  It is based off of how long they haven't been used for and also how much you have.  If you send it to another address it is reset.  But this doesn't exactly discourage spending cause its just 1% ppcoin per year.  And this method is quite secure also because it is hard to get 51%+ stake.  So instead of power protecting the network we have the coins protecting the network.

Sure thing.  I want to help as I have a fork in mind for PPC in the future, I just don't understand enough about it right now to really trust forking the code.

Oh!  That's interesting.  So you'd need to somehow get 51% of all the PPcoins available to have control over network.  That's a lot harder to accomplish, since, in BTC's case, you can coerce a group of miners into one pool and thus take over the network in this way, but you'll never coerce someone into sending all their coins.  However, if this managed to happen, it would be difficult to reverse the effect, wouldn't it?

Aside from that.  It would seem to me that you'd be better off putting some PPcoins in cold storage, and having a separate wallet in which you spend.  The cold storage coins slowly build up, and if you gotta break into it to get more funds, you have a few more than you had before.  1% really isn't a lot tho :P  It's almost pointless, lest you have quite a nice sum.

You can see that in the long run, PoS is actually more secure than PoW.  If the government want to destroy PoW, all they need is to gather some monster computers and they can out run all the other users.  But in the case of PoS, if they try to buy half of the coins, the price will skyrocket that even the government will find it difficult.

The other major problem is that for people with small sums, the amounts generated will eventually cost more in fees to spend than the stake outputs themselves, though SK states he also isn't concerned about this.

Also, stake isn't generated w/o being connected to the network, so cold wallets per say for stake generation are impossible.

One of the "problems" of the ppcoin POS scheme is the duality of the coins, acting both as currency and as stake. It is poorly studied, but the duality introduces a strong coupling between market price per coin and POS mining generation power, since increasing market price give an incentive to move coins out of stake. Thus an economic attack, in form of a short squeeze, may lead to a weakening of the network since people move coins out of stake. The opposite should be true for POW, since increasing market prices lead to an incentive to strengthen the network. The different dynamics comes from the reversibility between stake and coins. POS is still young and proposals like proof of burn can show ways out of this dilemma, e.g. by making the conversion from coins to stake non-reversible.

So it only counts while online?  Would it have to be online 100% of the time to ever hit the time required to gain the 1% a year?

This is not completely true. You can't mine PoS blocks if you're online, thats true, but you can still charge your stake while being offline. Then you might create some coins if you come online. Dublec, the owner of the bitparking exchange, found more than hundred PoS blocks in a row after bringing up the exchange's cold wallet. The new Version should fix this vulnerability, however it is not clear yet if it really can. Just one of several reasons why PPCoin still has central checkpointing.

I like the innovation in the proof of stake concept in PPCcoin very much, but in the current PPCoin implementation, as far as I understand it, I see some drawbacks.

I’m also thinking about improving the proof of stake concept, therefore I have some questions:

Why using coinage for securing the network? Doesn't this make it more vulnerable to burning coin age double spending attacks?
Isn't it better to just sign with the coins itself?
(The reward of coins could be still 1% per coin age)

Isn't it better if you can secure the network at once with your coins? So why someone has to wait X days?



Perhaps we can use a kind of consensus algorithm like ripple has and combine this with proof of Stake.
So the current ledger would be the ledger with the most "coin votes"

P.s:: A feature, which gives you the ability to give another address the "voting" rights for your coins for a certain time period would be nice. This could solve the cold storage problem.

It's a good question. Main reasons are:

1) Using new coins to sign block could cause problem for the block chain to recover from a chain fork. This is why 30-day minimum is chosen with the assumption that any chain fork that extends that long has to be bugs, attacks, or major network partitioning incidents, and would have been resolved via developer and user intervention.

2) Stake kernel is verified before block acceptance/connection to prevent a type of DoS attack on block-chain storage. This also requires a hardening point consistent among all nodes. (As of v0.3 this hardening point is about 21 days due to the 9-day modifier generation window)

I have explained in the design paper that we chose not to redesign bitcoin's data structures so the above design choices were dictated by this decision.

Note the coin-age weighting is capped at 90-day age since v0.2. So beyond 90 days it's effectively generated by coins instead of coin-age.

Can you elaborate on this?

Namely:
1) What stake kernel is and how it is verified, and what you mean by "block connection".
2) What a hardening point is, and what the 9-day modifier generation window is/means.

1) Stake kernel is defined in the design paper, it's the input 0 of stake transaction. As to block connection, bitcoin block is processed in 3 steps: check/orphan, acceptance, connection. Connection is when all transactions are connected and inputs verified.

2) Hardening point is where all nodes are assumed to have reached consensus on the part of block chain before this point. The stake modifier introduced in v0.3 is calculated over a roughly 9 day window. I have described this before in my weekly updates.