Topic: Communist Bitshares Wealth Redistribution IS THEFT! - page 18. (Read 28375 times)

It's really a case of quality vs. quantity: We would rather have 101 quality nodes than 100,001 unknown nodes. Especially when adding nodes increases operating costs linearly.   It's that simple.

• We added a node reputation component and the ability to detect node misbehavior and know who misbehaved.
• We added a way to uniformly distribute signing across nodes so that node influence is strictly limited.
• We added diligent discrimination.  People have to work hard to become a node and aren't likely to give that up when they are certain to get caught.

Decentralization occurs in the stage where delegates are selected by all owners.
Decentralization occurs in the stage where delegates are removed by all owners.

With nodes of this quality, it is our considered opinion that 101 is enough (especially with up to another hundred hot spares standing by to swap out any failures without missing a beat.)

Another big factor is network analyzability:  Our biggest fear is that there might be some latent attack vector or bug that our best efforts could not detect.  By engineering a very simple, easy to predict system where owners can observe node behavior and hold node operators accountable, it got a WHOLE lot easier to assure ourselves that the network was secure.  This is under-appreciated, but priceless.

You've got to get well known and trusted to get elected.  It's very hard to get well known and trusted in many different identities.  And even if you did, you can't get away with misbehavior without being detected.

You may say, "The idea is to be trustless!"  We say, all existing systems have trust placed somewhere.    
We merely make it explicit how it is being placed, give you control -- and decentralize that!

Having "just enough" trusted nodes determined by proof of stake gives BitShares other advantages, like a way for owners to directly influence how the network is run, what development and marketing initiatives are preferred, and if and when to upgrade to a new version of software.  It gives BitShares a way to fund itself by allowing employees to work for equity.  This is a long term competitive advantage we hope no one else copies any time soon!  

All of these considerations went together to form a design sweet spot that lets BitShares break new ground in many areas.  

Perhaps, anyway I'll leave it to people that understand it better than me to discuss.

Worst case of selective blindness I've ever seen... 

Unfortunately this doesn't work well in the real world, the majority think it's an SEP, someone else's problem/responsibility. This incentive structure leads to very low participation and probably only a few whales forging but trying to appear decentralized.

To prevent this centralization of power you might want to consider maintaining a good level of decentralization via a 101 delegate system.

Not when they are limited to 101, vulnerable to Sybil attack and multiple delegates are allowed to be controlled by one individual.

The law of diminishing returns. Why is 6 confirmations enough for a Bitcoin transaction to be deemed secure? It's because the difference from 6 to 7 or more confirmation doesn't add a significant amount more security. The same can be said with nodes, Unless you are making and exponential jump in nodes, the difference becomes mute. How much reliable is 99.999% vs 99.9999% in overall security?

Yes, I saw it.  

Correct me if I'm wrong,
but pools don't really centralize
Bitcoin's Proof-of-Work consensus mechanism.

Here's why:

In order to attack the blockchain,
you have to build a chain several
blocks long onto the end of the
current blockchain.

If miners in the pool are all working
just to solve one block at the end
of the current chain, how would the
pool accomplish this?

EDIT: Apparently at the current
time, mostly pools feed the block
headers to miners, although there
are several ways for full node miners
to prevent 51% attacks by pools.

You forge to protect your existing investment.

You missed this bit. No additional cost, many extra nodes. An encrypted messaging app or similar is more likely to take off though before this is finished development. Whatever the market decides.

Please explain how limiting the number of nodes in any sense promotes decentralization.

Ok so there's no profit incentive. Atm NXT is relying on goodwill essentially but in future NXT businesses will be running nodes as they already have servers. So NXT is designed to centralise around businesses without the people that own NXT really participating.

I think you missed the first paragraph - cost of a node is negligible to a business. They already have servers running 24/7 for any payment processors they run. Nxt could also be bolted on to MMO games to run in the background of say skyrim. Bytemasters post is a strawman.

Forging is not designed for making profits, it is for securing the block chain only. This was very hard for people to accept.

Have you seen Bytemater's latest blog post?  

http://bytemaster.bitshares.org/article/2015/01/09/How-to-Measure-the-Decentralization-of-Bitcoin

Bytemaster talks about decentralization of NXT too in this new post. I think the argument is that assuming NXT has a similar distribution of stake to Bitcoin then DPOS by BitShares is more decentralized.
As far as I know NXT has potentially a worse distribution than Bitcoin, so DPOS would be far more decentralized.

I'm an answer questions asked guy. Call me old fashioned...

^^^crypto creates un-bridaled arrogance for those with very meager achievements 

As you can imagine, the number of opportunities to talk about something like BitShares is overwhelming.  It could take all our time if we let it, and that would be fun.


So we try to keep Bytemaster locked in his office, writing code, doing interviews, and working on general purpose articles that only he can write.  We do unleash him on Fridays to answer free-form questions from a world-wide live audience.  (He also generally takes time to scan every post at bitsharestalk.org and occasionally can't resist joining the discussion.  Fortunately this has only happened 7965 times to date.)

Then it's up to the rest of us to pitch in and help out.  Our strongest technical types generally hang out at bitsharestalk.org where we can get the necessary concentration of brainpower to thoroughly address questions.  If you really want to get proper attention on a juicy technical issue, that's the place to go to discuss it.  This has the added advantage of organizing all the collectively-derived consensus answers in one place, rather than scattering them across the internet where most will never find them.

Some of us also like to reach out and visit other forums to learn what the hot issues are and try to answer the easy questions on the spot.  The biggest service we can do it that capacity is to help people find where their questions are being discussed by posting links to articles and threads.  We also refer tough questions to Bytemaster who then tries to respond in one of his general purpose articles where more people can see it.  He has written three articles in response to this thread since it went live.  The latest, just out, is:


Then there's me.  I seldom try to delve into individual technical questions.  I'm a big picture guy.  My main objective is to help people see the forest without getting hung up on individual trees.  

That was the purpose of the "side step" in my previous post here.  If I tried to answer it here, it would only be the product of one poor old aching brain - not good enough.  (And then we would be wasting everybody's time arguing about why Stan's answer isn't good enough.)


Now, the BitShares dev team is focused with getting on with polishing, hardening, and extending the BitShares product.  Others can continue to research individual issues and adding to the general knowledge base - for use in future products.

But there are ways to get answers to your questions at all levels.   We try to point people to the right venue to get satisfaction. The rest is up to those who really want to know.

fuckin' propaganda 

I don't have time to produce
an algorithm, sorry.

And I am not familiar with
the details of the Bitshares
implementation -- although
that appears irrelevant to
the specious points in the article.

If you control stake, or
delegates, in this case,
you simply keep trying to double
spend until successful. 

With "normal" PoS, the cost
would be the price of the stake,
although you can always sell it
after (or even before) you attack,
giving you a net 0 cost.
 
DPos seems even worse in the
sense that you're actually making
money as a delegate once you get
your security deposit back...
and you can then attempt your
double spends or reorgs.