I will give a last shot in this thread, I should probably not even bother as the dishonesty and way to construct my point as a pos argument make it hard to have a discussion, but ..


I need a remote control to access my tv , does it mean my tv had permission ?  Nope. Anyone can grab the remote and charge channel.

I need a transistor to listen the radio , does this mean the radio has permission ? Nope, anyone can grab the transistor and listen the radio.


There is investment cost to make profits out of it. Doesnt mean it's permission based.


But my point is not even this.


The main point is the OP seem to have the same thinking bias of people who only see coins as way to make profits.

And they think along the line of "I do something that help the network or is good for the coin and i deserve a reward"

Which is incorrect thinking.

What matter is that the procuded proof included in the block is sufficient to make sure it will be accepted as the next block by all nodes .

The only need for a reward is this.

Because the way pow work, there is cost and risk involved to produce this proof, so to make the game interesting and it can get enough security with high market cap, it need a reward.

The amount of work and cost to produce valid block with valid tx is marginal.

The amount of work needed to produce the proof that this block is the next block is huge.

With pos, the amount of work and risk involved with producing this proof is minimal. So it's much less based on reward / risk economic model of pow.

And if the amount of work and risk to produce the proof is low, then it become more like private network, or permission based system, where its a state associated with the user in system that determine his authority. This is the case for POS.

But again I dont have any coin, im not here to speak fud or argument pro or against any coin in particular.

I personally think even is most  case having trusted nodes and private network is much better than global pow consencus. For many raesons.

So again it's weird how my point get constructed as an argument against trusted node or private networks. Because it's not my point at all.

Anyway have a great day 

I asked my girl friend yesterday, she is indonesian and work with chinese ambassy and she know a bit of japonese i think.


What she told me that satoshis is a word related to taoist concept of balance of opposite (male/female, yin/yan) etc, and nakamoto mean 'the central source of'.

http://www.ancestry.com/name-origin?surname=nakamoto

Nakamoto Name Meaning Japanese: ‘central origin’ or ‘(one who lives) in the middle’; found mostly in the Ryukyu islands.

http://www.behindthename.com/name/satoshi/submitted

Given Name SATOSHI
GENDER: Feminine & Masculine
USAGE: Japanese
PRONOUNCED: sa-TOH-shee
OTHER FORMS: 覚 Japanese
CONTRIBUTOR: Spirited Sarah on 3/12/2011
LAST EDITOR: Spirited Sarah on 5/11/2011   [revision history]
Meaning & History
Means "wisdom" or "sense" in Japanese


So litterally satoshis nakamoto would mean ' the central source of balance '. (Cause wisdom with taoism is related to duality of symmetric opposite)

Because with the kanji also name can be translated to a meaning.

And i'm pretty sure it's not hard to find connection between nash theories of equilibrium from opposite agencies and taoist philosophy of harmonious whole from opposite and central balance.

In the idea, it's not because you name a coin 'archimedes' or 'descartes' that your name is necessarily this

It more look like a philosophic concept based on tao and/or nash game theories, involving surely persons from different field.

It's almost sure to me it involve at least two persons from at least mathematics and IT background.

Because to get into implementing a game theory model on a P2P network like this, need necessarily two persons. And it's obvious the code is not involved with game theory math. But in itself it doesn't mean the value and parameters are not computed from a math model issued from game theory.


And i'm being more and more convinced that there might be game theory model directly involved with the code of bitcoin now.

The two more important value for a node on POW is basically block reward and mining difficulty, that's the two things that matter, and it's exactly the same parameters you would find in game theory model (risk = difficulty = 1 / rateof ( good nonce) ).

Or it doesn't look like something that would not be very hard to express in term of nash game theories.

I'm not implying POS or centralisaton, or trusted node is a bad thing

But with POW, anyone who is entirely exterior to the network can use ASIC or other to gain control over the network, or distrupt it.

With POS only people who have coins can do this.

Mining doesn't require any form of identification or interaction on the network prior to be able to mine a block, POS does.

With POS the power over the blockchain can be changed arbitrarily by coin holders, aka everyone can give his coin to someone, or if the POS is computed from a separated token separate from the coin, it can become exactly like a permission system, except permission cannot be revoked once granted unless you have the private key to the addr. A POS coin could easily be made to work exactly like a private network.

With POW anyone can always get ASIC or mining hardware and gain control over the blockchain, no matter who has the coin, even if there might be economics argument against this, it's still more permissionless, even a total collusion of all the network user or operator don't have more power than anonymous unidentified miners.


It's not mainly about pos/pow argument, but on how the consensus is reached on the network, which is the main point of the proof of XX that has to be valid for all the network, and the system of reward either it's from staking or mining stem directly from this problematic of trustlessness , with POS you automatically rely on stake holders, and there is not other form of power on the network.

Would you really trust a POS coin where all stake holder are anonymous hiding over tor ?

Ultimately the logic resolve to trusting stake holders, which is why it's also easier from computing power perspective because the proof of validity of block do not come from this. The security come from the fact that you rely on stake holder to keep the chain valid for everyone.

With pow, as there is a whole economic of its own with it, with it's own reward/risk, you only have to know that the person is complying to the pow protocol, and participate in the mining economic with the reward/risk associated with it, no matter who they are, if they have interest in the coin at all, they could just be mining for a random coin on a multi pool, but their economic interest is always to produce block that will be accepted by the user of the network, no matter if they are user of the network themselves.

It's why it's completely permissionless. Either POW centralized or not in the end doesn't matter because it still mean the miner are participating in the reward/risk economy imposed by the protocol, and you don't have to know anything else about them, and they don't have to be user of the network at all for their proof to be meaningful for all the node on the network.

The whole point of reward for proof is to emit blocks with a proof that is enough for all node to accept the block as valid. I'm just comparing the two system to see what the reward is really for in the two case, and it's a different economic and trust logic behind.

But i'm not saying POS are bad or worst, it's just to compare the logic of the two system of proof and economic of reward/risk involved with it.


With POS it's all about trusting stake holders, but it works well too, i'm not saying it's bad or 'less secure' than pow or anything, this logic works, but the logic is not based on the idea of permissionless / trustlessness at core.

As far as i know anyway most POS coin turn to low reward after relatively short time, because anyway in the end the reward/risk economy become useless, as the risk involved with staking a block is low, and the need for reward is low too because stake holder are naturally incited to mine valid blocks even if they don't win anything out of it, and the risk involved with staking is low.

The logic is less into the reward/risk game involved with proving anything through the network protocol level, and there is not a lot of new coin emission either. So it become more like a private network. Even if permission can still be moved around by any stake holder.


But i'm just pointing at what the reward are for in the two schema, and why what matter with it is to be able to produce blocks that will be accepted by everyone, that's the only thing it's made for.

What matter is what the miner or staker prove by submiting this blocks and getting the reward from it.

It's not only about inflation/coin emission or the coin being deflationary or not, it's about what the staker/miner prove by submiting this block to the network in sort that it will be accepted by everyone.





Let say, with simple collecting of tx free, any node can process any tx randomly from the network, without any proof of work, and each node can produce different blocks with different tx, then what ?

How do you say which node has to the good next block for all the other nodes ?

How do you now from 10 differents block with different tx which tx are the good ones ?



The problem is not validating the transactions, but that there is naturally easily a way for all node to know which block will be good between different ones produced by different nodes.

It just occured to me now how reward seeking is completely the base variable for game theory algorithm  

https://sites.google.com/a/nau.edu/game-theory/about/philosophy

In Economics, Game Theory models the behavior of individuals as if they are participating in a game. Much like any other game, they are playing to receive some sort of payoff or benefit. The goal of the game is to attain the highest reward for themselves by using any strategies available to them. Risk dominance and payoff dominance are two related refinements of the Nash Equilibrium solution concept in game theory defined by John Harsanyi and Reinhard Selten (Risk Dominance, 2013).

So it's still possibly someone familiar with game theory and mathematics. I was looking for links from mathematics jargon in the code.


Now I can completely see the equation with risk taking = computing a hash, reward = coin emission for the block miner, and how the thing is tied together with the proba/risk as work, and how low risk taking lead to seek for consensus on mutual benefice, and lead to an equilibrium.

Where the force of the market aka speculators/whales are still separated from decision power by risk taking of computing hash to win the reward, and they wont manipulate directly the network even if they own large part of the data it hold.

And it still make in sort everyone still will keep relaying the good transactions, either speculators/whales, miners, with different risk taking (buying the coin /  computing a proba hash (difficulty = 1/rateof (xx)), and different reward (coin emission for miner, high coin value for whales), well it still need more thinking on it to get it completely, but the similarities in thinking with game theory math start to appear to me  

Im sure a good mathematician could pull out the 2x2 matrixes with coefficient being reward & difficulty (aka risk), weighted on if the risk is computing hash or buying coin, and deduce the good parameters for it to reach equlibrium on consencus. ( !   fatal genius )

It's a bit more twisted than this because miner also get benefits from high price of the coin,  and also you would expect pow difficulty to get higher as trading volume & market cap increase due to higher coin value. And the inflation rate of the block reward also cross the speculators reward (and maybe involve a risk for them).

I really wonder if the parameters are pulled out of mathlab now  

Im sure it's a very simple 2d stuff in game theory now with difficulty/reward with trading and mining  



But it's hard to get how he went from the math theory to the code, or from buisness problematic to math and to the code.

If he went from math and ideal concept to the application code, there is definately a break in the chain somewhere and at least 2 persons involved.

But satoshis nakomoto looks more like a project name based on concept of balance and source of balance, involving person from different expertise, and probably inspired by Nash in some part.

You see for example  libsecp256k1 only by the name you know it's made by a mathematician

But it has been integrated after, as a library with math functions.

Nah, it's not this

It's very typical of engineer who are all indoctrinated with making profits with startup, and their thinking is getting the lowest time of developpement for maximum profit, developped with clear timing in mind to reach the spec and whitepapper.

This lead to always tend to use well approved mainstream libraries, and where code reusability , packaging of code is more important, to be able to develop complex application with many different aspect with low time.

The c++ code is pathologically this thinking, of encapsulate / integrate / stick some boost duct tape on top of it, it does the job for the buisness model, it's good.

The code wont be modified / upgraded by anyone, no one has to modify it, the open source is more for trustless aspect, than for collaborative developpement.

But there are nicely packaged app for linux/win/mac and some relatively good security, and it allow to duct tape things together in a way that it can still work together without crash, but there is no "holistic approach", there is low coupling everywhere, some .h file are included in almost all files, it's still very monolithic.

The things that duct taped together are already existing framework like qt, boost, openssl, database engine, and that sort of things.

Not code made by bitcoin developers.

In my analysis, there is definately contradictory aspect to bitcoins. When I do analysis, I try to grasp at person intention, what they spend time on, what they care about , and what are their motivation and state of mind.

Because you get on one side still a super smart concept, well polished, well thought, very deep etc.

And on the other side, the code who seem to be really rushed.


On one side something that definately involve Nash like math. I have read the last post of iamnotback im convinced Nash theories and bitcoin are connected.

But to write blockchain problem, mathematician would express as something like a set of nodes, and matrixes / tensors applied on them to get some kind of statistic informations or whatever.

It's generally easy to spot when code is made out of math theories, because the variables and functions and organized with mathematics thinking.

Here you get code who originally used openssl crypto, and most of the work on the code is glueing different part of framework together.

Basically, it's not math code.

It's not code made to be easily upgraded, or developped collaboratively GPL style.

When people want to launch gpl projects, they install bugzilla, and tracking tools, doxygen,  and lot of other things to manage collaborative developpement on large scale. No such thing with btc right.

More or less the base was laid once for all, from the first time, with method that remind completely of software industry, one whitepapper, one shot release made once for all, and not thought to be developped GPL style. Not too much documentation. Something targeted at end user more than to developpers.


And ive been tweaking with my share of blockchain code, on bitcore, monero and blackcoin, and they are incredibly hard to modify safely. There are threads everywhere, even you could say the whole thing would only make sense if someone wanted to make the threading such that it's very hard to change anything because  of all the different concurent access from different things.

it's so messed up you would almost think there is some game theory in it

But the whole way the code is produced really look like something out of a company or start up. Or from people who think in term of commercial software, not mathematician or gpl guys.

From the post attributed to him, you see he still had a plan in mind, but more something based on trading/gambling shared profits and personal profits rather than something really made out of mathematics theories.



If I had to make a bet, id say there was on one side some sort of think tank crunching on the concept of trustless security, decentralized currency, and distributed database is also huge interest for IT due to the economic impact, and adding some twist of crypto on top of distributed database is not necessarily super new either, there are things like hibernate who already deal with this sort of problem, and the infrastructure of blockchain is a distributed transactional database.

The part in itself with distributed database, and signed operation/transaction is not totally new.

But what is very smart with btc is that it's completely made to take in account interest game, speculation, reward, and is completely sound commercially.

It's clear to come up with something like this, need to have at least heard of the kind of math of Nash, and the problematics involved, with decentralized authority, conflict of interest, to get something that is successful and reliable.

But to me it seem it's more like someone like smart buisness man from IT world, from the world of startup and software companies, who would have been introduced to game theory, maybe in the context of gambling, or probability based games, but fundamentally someone from IT industry, and maybe he came into Nash throught online gambling industry.

Yes POS is not permission less , unlike pow you need to have coin to participate, it's less anonymous, and a step away from the principle of decentralisation, permission less etc


But the thing is the important thing is not the reward, as i have seen even pos reward is a bit silly, because stake holder should be naturally incitated to keep the network functioning, only to keep the value of the coin, and the energy cost are marginal.

The whole point of reward and pow become useless in the context of trusted node.

Like with multichain and private chain, the security is not based on pow.

The energy cost of processing the transaction and producing the blocks in itself is marginal. Any regular user should have enough interest to keep the network working by itself, if it was only about processing tx.

The real cost is making sure all the network will accept the same block in permission less / decentralized manner. Basically to resolve orphan blocks and double spend, or cases where two different valid block are produced.

When you rely on trusted node, a common computer can process hundreds of tx / sec, and there is no need for PoW, only proof of identity of the trusted node to show the tx are valid and will be accepted by everyone.

And in the absolute, reward could be detached from coin emission, aka via tx fees or other.

If there was a way to determinstically determine in all the case which block or tx should be selected in case of conflict, there might not even be need for PoW at all.

There is difference between learning a language and pulling out commercial app within deadlines

Did you ever worked in software industry when if you follow what they teach at school it need 2 month and you have one week to roll it up , and then send complaint to customer support ?

Bitcore it look like this sort of code

Depend the criteria, if criteria is time spent vs commercial benefits it's not the same than scientists or academics

Me it was more the opposite with math teacher lmao you got the good result but i dont understand a shit, it's not like we taught and your writing sucks took me 1h to read it so you got F

Me after class ok, i will not come back to your exam :p

Learning c++ basics at school is a thing, making a portable app with boost/qt/openssl , multithread , database etc in reasonable time with enough security etc it's another thing

I think even at any rate a good mathematician noob in code wanting to do this would have used java. Some math nerds do java applets

But a crap like bitcore with c++, net, crypto, thread,qt etc need people who have experience building commercial software

Orlov says this too  


https://en.m.wikipedia.org/wiki/Dmitry_Orlov_(writer)

Exactly my thinking, posted similar though in another thread

Doesnt look like code made by institutes or scientists, or mathematician but from software industry like engineers

Nash mostly did coding for mathematica, and more looking like mathematician code with arrays matrixes and operations, with groups, subset, and math concept, mathematician tend to see programs as linear system, or group theory and program libs or scripts for mathematica, not portable c++ app with boost & openssl & qt.

But maybe he hired someone from software industry with small budget instead of doing it himself with fortran or mathematica, or going throught institutes or universities.

But you see here i see a fundemental problem with this approach

I will try to explain shortly lol

But the whole idea of immutable truth comes from parmenides thinking, saying only truth  ( or god ) is eternal / immutable, and it has direct influence on idealism, but in any case all material things are mutable.

The closest things are gold or diamond, not because of federal reserve or wall street, but if you put a gold bar in the bottom of the ocean it is still there in 1000 years. That's immutability

Bitcoin immutable .. ha ha let me laugh , it's not ever sure it will still be there in 1 year

Immutability can have resonance in alchemy or gnostism (aka immortal soul etc), but it's more interesting with alchemy as it study how time affect things, and also lead to algorithm, and thinking you can represent the equivalent of time transformation with algorithm.

In the context of computer, the effect of time is a core processing an instruction from a program, it's how time will effect the system represented/maintained by this computer program, with Turing law you can come to the conclusion you could get a completely deterministic representation of time using a computer program based on mathematical / philosophy theory and Turing law.

The problem is when you want to have 'real time' simulation, aka based on hardware or system timer, like for real time physic simulation, you will never have exactly the same result after 30 sec on a multi tasking system.

Even for very simple case using algorithm that are supposed to be deterministic with turing law (or if computer can't compute any physics, what are they good for ?), once you introduce even a single external variable like a timer in the algorithm, it stop to be really deterministic.

It's in part also why i wanted to make my micro kernel, to do real time physic simulation having all the core and memory dedicated to the program, even disabling completely interrupts, virtual paging, and that sort of things to have more linear execution of programs, and to have something more deterministic and closer to Turing machine.

But running on the average busy windows or linux, even plain real time physics applying plain deterministic linear algebra based on real time timer (aka coming from source external to the program) is not deterministic.

With something running on top of a world wide P2P network, thinking having some kind of distributed pow and synchronization over network packet will give something very immutable is very idealistic.

Something based on a model of state based P2P network can never be immutable on the whole network scale. Impossible. Or it need quantum entanglement thing

There is no science to describe immutability of distributed P2P network.

Even on the fundamental 'real time' program execution level, which is how the effect of time on the system is represented, there is no way to apply any deterministic to theory to a global program running on millions of P2P node.

Not even speaking on the economic theory and various theorical problem to solve with how to compute ratio of exchange against different things, either they can be stock options, bonds, other currencies, or direct goods or services, in real time, and keeping the whole system deterministic on the global P2P network scale.

Not even possible 

And indexing the stability/immutability on market law, well it never give any stability at all. Markets are even more unstable than multithread programs 

It's not so much about reward for node who produce blocks, but about proof that enough nodes on the network validated it aka solution to byzantine general problem. Otherwise , it's back to trusted nodes, centralized authority, etc, and the cost of operarion are then minimal.

Cia did this many times 

It's maybe bleak Outlook but seen this sort of things happening many time, hard not to see the pattern  

But still need to have faith in positive outcome lol even when things seems hopeless and sucks, it's the best time to focus on positive outcome and find good solution  

I watched ares yesterday also  ( http://m.imdb.com/title/tt4216902/ ) might taint my view lol

Reminded me of this with big labs giving product to boxer to rig the match out of big interest and no matter how good the boxer is, if he is not sponsored by the good lab he has no chance , and all is around economics interest for big corps and speculators while  the boxers get all the side effect of drugs, with all corruption lol

But even italian football has good players, and can be fun to watch, but the whole betting side is rigged, and in cahoot with sponsors etc etc

But it make people think they can become rich and famous being good at football, but in fact it's just about following sponsor interest lol

Same with betting, you can  try to bet based on sport speculation, and make some win, but if you know the mafia you can get rich very quick lol

Seem the same with crypto valuation a bit, it is more indexed on knowing whales and miners moves, than on the technical or utility value of the coin.


It's always the same thing in the end https://youtu.be/Dw4rtERX0zw

Maybe segwit can improve micro economic aspect, and least move some of the current power/weight repartition, and make it more stable but hard to say

The only point is see where my approach is different is with the GC vs ref count thing, but i think the two are orthogonal.

I want ref counting to avoid pointer ownsership and it's mostly used for short lived object, or permanent objects, i don't really make intermediate.

And you want GC to have macro management of memory with determined life time.



The thing is for me most GC language (like java or js) are incredibly inefficient with memory, and GC sucks for general purpose, and for most simple webapp, the case are not so complex.

The only kind of GC that really make sense to me is in video game engine, or game console SDK, because there is a high level abstraction of objects like levels, bsp tree and some high level abstraction of the objects and the execution flow inside of the engine.

So it allow for good management of objects life time in automatic manner because all execution context and the object it use are well defined from application level.

And mostly the whole GC is hard coded into another sdk, and hard coded into the game program.



Ultimately for me GC only make sense in the context of higher level paradigm to define object life time in relation with explicit life time boundary based on the application execution.


General purpose GC sucks and they just end up not freeing the memory.

Trying using any java or javascript app that is a bit intensive with opengl, multi media or such for long time and it will all end up with a big mess. Unless the user flush the GC manually.

Android cheat a lot on this with fake multi tasking and only the memory for one app has to be really loaded in memory at once, so it's not so bad when you have 40 app and tabs running, only one is in physical memory at one time, so even if the GC sucks, it doesn't matter too much.




But i would be all for defining high level application paradigm when it make sense to have fast GC when object life time boundaries can be determined easily at compile time or when it's possible to determine at a certain point of the program that no reference to a certain number of marked objects will not be used anymore anywhere and they can be safely sweeped.



But other than this, for general purpose it sucks, and for short lived objects it sucks.



As the way i program the blockchain node doesn't use much caching in memory, and deal mostly with short lived objects who will be coming from the network and either stored or sent back to the network, it's not too hard to manage object life time.



And with the reference counter, it still allow for lot of freedom when passing reference to other functions, even for short lived objects, or when life time doesn't matter, or can't be really known at the moment the object is instanciated.



Even if it's not extremely efficient to deal with complex cases, with the lockless internal allocator and memory pool it's still fast to allocate and free memory, and can easily create specific pools for object that have determined life time boundaries.