This is not something I would like to do. But it might become the only way to save yacoin if situation gets worse.

I told that to my woman, so she will go easier on me now.

Current chain "caretaker" mines approximately million YAC every ten 10 days. If he doesn't sell, he might gather enough funds in next couple of months to compete with higher (pure) PoW hashrate by later also minting PoS blocks and consequently raising his chain's trust.


Correct - by all means honest branch should have more trust with PoS blocks included already.
But if an attacker can generate longer PoW-only chain, that one wins.
Chaintrust fix in development will mostly address issue Balthazar exposed. But fix for ignoring PoS blocks issue will make it harder/impossible for attacker to create longer PoW-only chain.


I have a pretty good feeling about what is happening now:

We build PoW+PoS+PoW chain with a trust of 10+11+20=41 for example.
Minutes later attacker broadcasts PoW+PoW+PoW+PoW+PoW (10+10+10+10+10=50) and kicks our blocks out.

But the real problem is that PoW difficulty in our branch get's adjusted towards 2 minute spacing when PoS block arives. On the other hand, attacker's blocks are stil one minute apart, so he is mining on lower difficulty, has more time and is able to overtake us eventually (before our POW difficulty readjusts back). It might be that in addition to having more hashing power.
Just one more thing to consider while refactoring GetBlockTrust method.



Every transaction sent must be eventually included in a block by a miner (PoS or PoW miner).
If there is only one miner, he decides which transactions land in the chain. Small change in code and any or all optional (not coinbase or coinstake) transactions can be left out.
This is extremely ungrateful situation, but if everything mined now get's directed towards exchanges, he, she or they must let transactions through (not all though).

If for some reason this would stop, we can conclude there is malevolence behind and attacker does not really care for any loss incured due transaction system "halt".
In that case we should temporary shutdown and possibly restart from past point to alleviate attacker's influence over large currency supply portion later.


We need to fix yacoin. If that goes well, we will get more attention and miners.

It's not about creating bogus transactions. Whoever mines all blocks can opt to exclude all transactions (except coinbase/coinstake). He has the power to stop people transacting.


And what about 49% of honest PoW miners, who do not get their block excepted? We are paying for hashing power, and getting 0 coins. Yacoin was all about fair distribution. Nothing fair now.


It has become privately mined CPU coin. All earnings go to one person/group. Not a single PoS block in a chain yesterday.



There is.


Agreed 100%. I would even sugest temporary network shutdown until this is fixed.



That is actually very positive information. I did not check YAC/CNY and thought attacker is hoarding.
Can we be sure it is him/they who sells?

I think both of you bring up valid points (insults excluded).

Price dropping until it is unprofitable to mine certainly sounds promising. But I see no large volume on Cryptsy or BTER since he took over. As long as he sits on so much mining power, he is able to control network.

And all the folks who quit mining now, made it possible for him to drop hashrate to just above 50%, which raised reward, made him richer and consequently he grows more in PoS power.

I am mining full time with 100% loss and intend to keep doing it. Unfortunately I only have a tiny portion of hashpower.

I haven't checked what happens to transactions now, but having them included in blocks can be considered an act of mercy.

What is happening here is not exploitation of chaintrust vulnerability  - issue that Balthazar warned about (that renders whole system unusable anyway). This is simple resource domination.

Transition to PoS only would mean throwing away main yacoin feature. I would not have any incentive to participate in development afterwards. I am in only because of scrypt-jane. Perhaps if we can combine that ...

But I also think there is a high possibility that yacoin is being run in the ground right now.
I am surprised how easily one person/group can destroy a crypto with pure mining power.

Ron, I share your frustrations regarding code readability. I guess you don't know where that 25 came from either:)
To find out what is going on, I help myself with wikis, forums, documentation sites, blogs, you name it ...

I am not sure it would be even possible to put all this information in code, but I agree 100% it should be more readable.
Perhaps one day someone brave will create libyacoin with design principles taken from libbitcoin



I misread your message. Sorry. I know now.


That would be most welcome for all of us, and also for future yacoin developers. Perhaps I will put something together ....

We have block rewards decreasing with increased difficulty right now. That is how it works.
Rewards do not matter here. They are a side effect of fix. Chain that does not ignore PoS blocks will be most rewarded.


Did you not see my reply (in your messages):







And while we are at confusion. I haven't received an answer to one of my previous questions. I wonder if anyone knows it:



As for the rest, soon I will put up my draft for fixing remaining security issue - small PoS blocks. In part it will resonate with what you write here.

I wasn't thinking about throwing away that 1 million + blocks. I don't know how you got that idea.
But you shouldn't count on them, if chaintrust issue remains intact.



My primary goal is to fix critical issues. Another goal is to keep PoS blocks in the chain and have them spaced apart at 1 minute and if we need to occasionally increase PoW block spacing in order to achieve that -  I have no problems with that.


Yes, if blocks get to close, we still need to raise difficulty.


We've had two simple solutions that did not work. I wouldn't want another one like that. Especially if it requires a hard fork.
Complexity should be addressed with revisions, testing and simulation. But if you have a simple solution and it doesn't come with major shortcomings, I would love to hear about it.


Chain trust is sum of GetBlockTrust() values for each branch. Branch with highest chaintrust is the main one. Follow first link pointing to peercoin wiki in document mentioned.
NVC's solution is not so strong as I'd wish. They give hybrid chain more trust (and perhaps higher rewards too). Sairon also acknowledged that some time ago.

Here is the idea:
Let's say every 10th block on average is PoS. Once there are 9 PoW blocks one after another, PoW difficulty readjusts to two minute target. If another PoW block comes difficulty adjusts to 3 minutes. And so on. Beave reminded me (good catch Beave) that higher difficulty also affects PoW reward, that means whoever will insist on PoW only chain will have to work harder and also make less profit.
I will need to make refactoring of code - whenever that happens nActualSpacing in GetNextTargetRequired() will need to be set to "min( 60,nActualSpacing )", simulating blocks were max 1 minute apart. We would not want difficulty to drop because PoW blocks were intentionally pushed apart.
Perhaps even GetBlockTrust() will need to take that into account - if one mines PoW only chain too long it's difficulty will skyrocket, but his chaintrust must not.

EDIT - answer2:


That get's called at startup. You also need to count PoS blocks while being online and minting or receiving new blocks. And what about that big painfull REORGANIZE when all PoS blocks get trashed?

Here are two drafts/outlines. Ask if something is not understandable.
And feel free to criticize or better suggest corrections and improvements:

fix trust
fix ignoring PoS blocks

And to those who have hijacked the chain: I am having thoughts about having block range validations hardcoded.
Are you aware that everything you PoW mine can be made nonspendable with next release. You greedy assholes. What are you, a bankers?

Again, did you check yacoin daemon debug.log?

You are right about 2 minutes. Perhaps I have got it wrong here:

Another problem would be if rogue PoW miner would combine such PoS blocks with his private chain in order to bring his chaintrust higher (trust value of PoW block following PoS block get's doubled). That would make it possible to mine under lower PoW difficulty and use a lot of small stake PoS blocks to help build up chaintrust against primary chain mined at higher PoW difficulty but containing lower number of PoS blocks.


Perhaps each PoS block in honest branch makes next PoW difficulty higher (by extending to 2 minute block time) and that makes it possible for private solo miner to eventually catch up since he is still mining on 1 minut block time?
Could that be one of main reasons why main chain can't run away easily?

I did not thank you yet for those videos - very helpful indeed. I am one of windows guys (but I strive to make portable code), and I must confess successfully compiling yacoin daemon under Visual Studio was a special event.
I hope the same will be possible with version 4.5.

For such errors just grep daemon's debug.log for "ERROR". It seems (15 posts down from the top) meaningful detailed info should be there.

I am not sure, but perhaps you have your wallet locked?
I believe it needs to be unlocked in order to submit blocks from rpc.

Yacoin needs several issues fixed. I've already mentioned them to Joe and Groko, ane some bits I've posted here. My hope is to get larger audience onboard.

• Flawed chaintrust calculation
• PoS blocks with small stake posing security risk
• PoW miners ignoring PoS blocks to maximize profits (also happening right now)

During past couple of months I've been analyzing the code, brainstorming and putting together blueprints for a solution.
It can not be done without hard fork, because of both -  first and last issue. Fix for small PoS blocks can be rolled out incrementally.
Currently I am analyzing fix for chaintrust issue, because it underlies solution for others.
I will let you know about details in a day or two.

This will not prevent miner with 51% from capturing the chain, but fix for last issue will force PoW miners to include PoS blocks and fix for first issue should make PoS blocks get higher trust value. Consequently block reorganizations should not get deep as now and stakeholders should get their power back.

That would make it harder to put together currency supply projections.
Perhaps something like this could serve as a template.

Looking at code it seems that block reward calculation goes like the this:


but wiki and other sites say it is like this:


Here for example
https://yacoin.org/tech.html
http://coinwik.org/YaCoin
http://coinwiki.info/en/Yacoin


I can not figure out where did 25 came from. Can someone explain this or at least point me in right direction please?

These are bits values from both branches from blocks 1103105 and onward:

1e02141d(01:55:49)   1e02141f   1e021430   1e021418   1e0213fa
1e02141d(01:58:32)   1e021469   1e021478   1e02146f   1e021496

Second branch, that later overtakes the first one has higher bits value (block found 3min after first branch). That gets expanded to higher target value and thus lower difficulty and not the other war around like I presumed before. That could make it possible to catch up with first chain without having more resources.

Regarding those 13 disconnects:

Here the chain splits. Following both branches ( links after "Next Block" ), I believe one branch mostly awards familiar coinmine.pl address (YESJtirJQJsDRTVS67UuuPZfeoJ6YWojwt), while the other (it's coinbase rewards 15Lxek3tUvxzDhjX1dtC3Znz6xXi62HtvF) is competing branch that refuses to mine on top of block of first one, but it does not seem to withold blocks until later.
Two steps later It can be seen that it also bumps up difficulty (perhaps because of different times), so it is able to overwrite best branch inspite of that one having POS blocks included.

I don't know why, but a lot is happening - 62 reorganizations in last 12hrs.
13 blocks disconnected at one time.

Tx fee's function is to prevent blockchain bloat/spam attacks. Lowering it would mean removing this protection.

The only way it seems sensible to have it dynamic (and less than 0.01) is to somehow tie it to block reward.
More miners, consequently higher difficulty and lower reward would indicate possible higher price. In that case it would make sense to lower mandatory tx fee.

As it is now, at such low price, tx fee could even be 0.1 yac.
I admit it is challenging to estimate price from difficulty, especially considering N factor.
Besides, I think tx fee is perhaps even more an economical than security issue. And we've got enough of those to fix.