Jump to content

NEWS
  • WELCOME TO CRYPTOCURRENCYTALK!
  • We've upgraded the website with a lot of new features!
  • 3 NEW THEMES! Click on the + Themes button above!
  • Notice about SOCIAL LINKS: UPDATE YOUR SOCIAL LINKS
  • New video section, add your videos: VIDEO SECTION
  • Advertising will be available shortly, hold tight.
  • If you have any issues with the new site, please submit a support ticket: SUPPORT

Should we Re-Base Gridcoin using this proposal, or should we stay with current codebase and fix it using the Roadmap 4.0 proposals?  

36 members have voted

  1. 1. Should we Re-Base Gridcoin using this proposal, or should we stay with current codebase and fix it using the Roadmap 4.0 proposals?

    • Re-Base Gridcoin using this proposal (a variation of Dash-Evolution/Bitcoin Core Nov 2017 with a Stakeminer)
      24
    • Stay with Current Code Base and fix it Piecemeal using Roadmap 4.0 Proposals
      9
    • Abstain (I do not wish to be involved)
      3


Recommended Posts

 

Should we Re-Base Gridcoin using Rob Halfords proposal, or should we stay with current codebase and fix it using Piecemeal/Brod proposals?

 

 

QUICK LINKS:

 

Minimum Escrow Requirement Poll:

GRIDCOIN III BLOCK DISTRIBUTION POLL:

 

 

 

 

 

 

 

All,

 

From a high level I've been testing cryptocurrency technologies in my lab over the last few months, on an information gathering campaign, and creating some internal mockups and doing budget expirimentation and looking for a way to break through the scalability barriers we have experienced in the Gridcoin client.

 

Finally, I have come to a better understanding of what I believe will actually work in production- and I did not want to finish writing a proposal until that testing was completed and it is now complete.

 

I would like to post a high level proposal to rebase Gridcoin for the future.  The intent here is to hold a public vote to see how much acceptance we have with this idea.

 

Specifically, what percent of the community backs This idea VS. going with other piecemeal solutions (those include fixing the current codebase, implementing Tomas Brod's proposals, and in general merging all of bitcoins security commits into the current existing codebase and then maintaining it).

 

So please, read this, and then look for a Poll that I create that will seek to understand the backing level of :  Either going with this particular Rebase, or going with Piecemeal fixes+Brods proposals. (Ill come back and edit this with a POLL link later).

 

I have had discussions with Denravonska (our lead dev) (I consider myself CTO for now, and him the lead dev), and iFoggz, CM, and Quez regarding this idea.  I checked the viability of this proposal from an IT standpoint and did the due dilligence in the background.  I have enough IT support from our back-end team committed that I know we can actually execute this plan if the community wants it.  The breakdown is:  Rob Halford (me) will push the core, Denravonska (Marco) will push the TCD system, and iFoggz will push the CPID associator system.  We think we can do this in 1 quarter.  So therefore time is of the essence for us to decide what path we are going.  If everyone is on board, this could be the future of Gridcoin delivered by March 2018!

 

Please see the PDF here:

http://gridcoin.us/gridcoinnew/dev/GridcoinIII_NonTechnical_Proposal.pdf

 

 

Warm Regards,

Rob

 

Share this post


Link to post
Share on other sites

@Rob Halförd - (Gridcoin) I like most of the things you described, but could you point us to some more detailed description of this PoSE algorithm? After a short googling I don't know if I don't understand it or don't like it (to me it looks like a flavour of Proof-of-Work and it is not the beneficial BOINC work here).

Share this post


Link to post
Share on other sites

I share @bzaborow's confusion/concern about the PoSE seeming a lot more PoW-heavy.  In addition, this looks like it will completely disenfranchise smaller investors (400,000 GRC is a pretty high bar -- at most 1000 people can ever have that stake, and realistically that number is a lot lower) both in terms of voting and in terms of any sort of rewards for holding beyond whatever happens to the price. The latter is probably less concerning than the former; "Gridcoin 2" was cool insofar as it sort of gave everyone a voice if they cared to use it and could be a future governance utility for things far beyond the chain itself, and this new proposal will completely strip that.

 

I may be able to get behind this if we can get some reassurances that it's not going to destroy the premise that Gridcoin is more energy-efficient than the competition to focus more computing power on distributed computing, and if we can at least talk about ways to avoid completely disenfranchising the 99.9% of people who aren't ever going to come remotely close to having 400,000 GRC.

 

 

Share this post


Link to post
Share on other sites
1 hour ago, nateonthenet said:

I share @bzaborow's confusion/concern about the PoSE seeming a lot more PoW-heavy.  In addition, this looks like it will completely disenfranchise smaller investors (400,000 GRC is a pretty high bar -- at most 1000 people can ever have that stake, and realistically that number is a lot lower) both in terms of voting and in terms of any sort of rewards for holding beyond whatever happens to the price. The latter is probably less concerning than the former; "Gridcoin 2" was cool insofar as it sort of gave everyone a voice if they cared to use it and could be a future governance utility for things far beyond the chain itself, and this new proposal will completely strip that.

 

I may be able to get behind this if we can get some reassurances that it's not going to destroy the premise that Gridcoin is more energy-efficient than the competition to focus more computing power on distributed computing, and if we can at least talk about ways to avoid completely disenfranchising the 99.9% of people who aren't ever going to come remotely close to having 400,000 GRC.

 

 

Regarding the 400,000 minimum, I was trying to set it high enough so that our Neural Network becomes a minority that represents Gridcoin.  Im totally open to creating a vote for the escrow level!  I think we can make tranches in the vote, such as 50-99k, 100-200k, 201-300k, 401-500k and hold a vote for this.  It should still remain a relatively high # (IE 100k) so that it is a challenge to "hold a leased seat" as part of the voting board, but I agree - it should be up to the community.

 

Regarding POW heavy, no absolutely dont want that to be the case.  To explain POSE a little more, there are two parts.  Proof-of-Service is a requirement set forth requiring a neural node to have an external IP, and stay online, otherwise the node risks falling out of the payment queue.  The queue is enforced using a program, the name is not yet determined, something to the effect of Gridcoin PoSE Bulldog, it is something that runs in linux alongside gridcoinresearchd, that checks in with other neural nodes every hour and votes for your node to be paid.  If you do not vote, your payment falls behind in the queue , thereby enforcing Pose.

 

The Mining is not intended to be PoW.  It is intended to be a CPU algorithm that runs slow, and this is only to keep the blocks moving forward.  Note that the miner only gets between 1-10% of each block, so it is already hardly beneficial to buy mining devices to mine gridcoin.  With the monetary dynamics of a lower reward for higher nonces, it will become more beneficial for the user to mine slower (to receive a higher reward) than to throw more mining power at Gridcoin.  I believe this system will work.  In essence, it is a CPU miner with a variable scale reward per block of 1-10% depending on how high the nonce is for the solution.  (With a tweak to X11 to break compatibility and to rename the algorithm POSE).  This is very green, and on par with the stakeminer we have now.  If it fails for some reason, our IT can port in the POSv3 utxout mining hash using the new "entire balance" idea, but Im trying to keep things simple and new, and move away from POS, and move to POSE for a fresh start.  But from a high level, it is intended to be more efficient than a stakeminer.  I would like to see 1% CPU usage in the new wallet, and 99% for boinc :).

 

Rob

 

Share this post


Link to post
Share on other sites
2 hours ago, Rob Halförd - (Gridcoin) said:

Regarding the 400,000 minimum, I was trying to set it high enough so that our Neural Network becomes a minority that represents Gridcoin.  Im totally open to creating a vote for the escrow level!  I think we can make tranches in the vote, such as 50-99k, 100-200k, 201-300k, 401-500k and hold a vote for this.  It should still remain a relatively high # (IE 100k) so that it is a challenge to "hold a leased seat" as part of the voting board, but I agree - it should be up to the community.

 

Regarding POW heavy, no absolutely dont want that to be the case.  To explain POSE a little more, there are two parts.  Proof-of-Service is a requirement set forth requiring a neural node to have an external IP, and stay online, otherwise the node risks falling out of the payment queue.  The queue is enforced using a program, the name is not yet determined, something to the effect of Gridcoin PoSE Bulldog, it is something that runs in linux alongside gridcoinresearchd, that checks in with other neural nodes every hour and votes for your node to be paid.  If you do not vote, your payment falls behind in the queue , thereby enforcing Pose.

 

The Mining is not intended to be PoW.  It is intended to be a CPU algorithm that runs slow, and this is only to keep the blocks moving forward.  Note that the miner only gets between 1-10% of each block, so it is already hardly beneficial to buy mining devices to mine gridcoin.  With the monetary dynamics of a lower reward for higher nonces, it will become more beneficial for the user to mine slower (to receive a higher reward) than to throw more mining power at Gridcoin.  I believe this system will work.  In essence, it is a CPU miner with a variable scale reward per block of 1-10% depending on how high the nonce is for the solution.  (With a tweak to X11 to break compatibility and to rename the algorithm POSE).  This is very green, and on par with the stakeminer we have now.  If it fails for some reason, our IT can port in the POSv3 utxout mining hash using the new "entire balance" idea, but Im trying to keep things simple and new, and move away from POS, and move to POSE for a fresh start.  But from a high level, it is intended to be more efficient than a stakeminer.  I would like to see 1% CPU usage in the new wallet, and 99% for boinc :).

 

Rob

 

The proposal for PoSE sounds insecure and near unachievable. How would you ensure that the network is not periodically attacked and certain nodes hindered from the voting process? In Proof of Work this is trivially evaded, since one miner would take much longer to find a single block than many miners, which means that you'd constantly have to attack the entire network over weeks for it to succeed - an impossible scenario given the resources this would take.

Mining is supposed to secure the blockchain. Target nonces are made purposefully hard to have competition between miners on who gets to solve the next block. It is an arms race by design and works well like this. If you invert this scheme by giving higher monetary rewards for lower nonces, you remove the incentive for innocent miners to do proof of work (as you stated above). An attacker however would not care about this small reduced reward, if this would mean, he could just increment nonces faster to be able to successfully mine several blocks in succession and execute a successful double spend. You even implemented a similar scheme before in pre V8 gridcoin and it was successfully exploited on livenet by doing more proof of work.

I cannot see how this would survive an attacker, and therefore reject this proposal in its current state.

Edited by TheCharlatan

The Charlatan

Share this post


Link to post
Share on other sites

 

Quote

Our daily superblock pays every researcher via a superblock payment drop.

 

How will this affect superblock size?

Quote

(I consider myself CTO for now, and him the lead dev)

 

 

Please explain.  Are we not open-source development?

Quote

 

Re-Base Gridcoin using Rob Halfords Proposal!

Stay with Current Code Base and fix it Piecemeal and by using Brods proposals

 

 

Would appreciate non-biased phrasing in the client poll. = )

Quote


One of our developers, iFoggz has created a CPID association solution that will allow us to associate a CPID in a scalable way without generating keypairs.

 

 

How does iFoggz proposed association solution compare to CM's and Brod's? 

 

-------------

 

Personally, I am here for Proof of Research -- a blockchain owned by those who contribute to a distributed computing network which focuses on science, technology, and data analysis.  A blockchain in which at least half of the blocks are made by those who contribute to this distributed computing network.  Masternodes seem to contradict the original purpose of Gridcoin.  It seems to operate in a way that would greatly benefit extremely early adopters and financially well off entities, while limiting access for new users and those who contribute processing power to the network.  Investors are great, but should they own the blockchain?  I feel like I'm missing something.  Am I missing something?

 

Again, I am here to see those who contribute to the distributed computing network benefit the most and hold the most power.  In other words, if I were a large scale research facility, would I make blocks based on the computational power I contribute to BOINC and would my contributions to BOINC grant me a seat on the voting board, weighted based on my contributions relative to other contributors?  Or would I have to allocate tens or hundreds of thousands of dollars of my already tight budget in order to "earn" my position on the voting board?  (It currently costs USD $585,000 to lease a DASH masternode).  How do masternodes seek to advance the primary goal of Gridcoin?  Do masternodes limit the potential of a well financed entity buying control of Gridcoin while contributing nothing to the computing network?

 

In its current state and presentation I would support the "piecemeal" improvements, but I look forward to being enlightened.

 

As a side note, the name "Neural Network" is misleading and I think it ought to go.  

 

Some suggestions: 

 

Data Storage Network DSN

Data Retrieval Network DRN

Data Collection Network DCN

Research Statistics Network RSN

Research Distribution Network RDN

Share this post


Link to post
Share on other sites

Hey Rob, Long time no see - sorry I've not had the time to respond to your emails, but I replied quite negatively to your proposals and email snippets in the private OPs slack private channel - you really aught to be more visibly present in Slack rather than just emails (which can be a bit of an echo/isolation chamber).

 

In terms of calling yourself the CTO - there technically isn't a Gridcoin company thus there is no such role, parts of your email on this topic which I found quite disconcerting/disturbing especially the 'firing/taking-over' terminology/intentions being thrown about by yourself - not confidence inspiring given the increased power you'll have in a masternode environment compared to the power you yield in the current network. The github repo ownership could be viewed as a CTO role, except the community has already begun working around this centralized repo for non-core-client work: https://github.com/gridcoin-community

 

I don't believe that a forum poll is appropriate for making this decision, it doesn't involve the Gridcoin network (coins/magnitude) and is vulnerable to being manipulated very easily. Forum polls back in 2014/2015 were rushed in the panic move from classic to research and led to a low quality economic policies being implemented (30* reduction, etc). The precedent for serious development polls has been 2-3month duration 'both por/pos' type within the Gridcoin network. I have thus voted to abstain from this poll.

 

Speaking of the forum poll, it's unfortunately biased in that it only provides the choice between your proposal & maintaining the current codebase - there have been alternative proposals for the last few months: https://github.com/gridcoin-community/Gridcoin-Tasks/issues/183 (Please accept your invite to the Gridcoin-Community organization)  updated post: https://steemit.com/gridcoin/@jringo/gridcoin-research-4-0-2018-roadmap-progress-report#@cm-steem/re-jringo-gridcoin-research-4-0-2018-roadmap-progress-report-20171111t133914323z (being somewhat active on steemit would be cool too)

 

 

Quote

 

Sharedropping onto a new codebase

 

If we're considering a snapshot/sharedrop onto a new code base (to remove plain-text email (and bloated scrypt block) data from blockchain, to shortcut to many requested features, and to improve stability of the network) my choice would have to be graphene.

 

I do not like the idea of forking Dash, masternodes would entrench power to the wealthy, as you'd have to pay a large sum of GRC to run a masternode where as with graphene you could have no investment required to get voted in as a witness/committee-member.

 

Graphene Links:

http://www.digiter.io/btsdocs/index.html (WIP Updated docs wiki)

http://docs.bitshares.org/

https://github.com/cryptonomex/graphene

https://github.com/bitshares/bitshares-core

https://github.com/steemit/steem

https://github.com/PBSA/peerplays

 

Disadvantages of graphene

  • New codebase thus potentially large developer learning curve.
  • A large amount of current code wouldn't be compatible & would require being ported to c++.
  • This is a mutual disadvantage when rebasing to any codebase & we're already talking about rearchitecting most of Gridcoin & snapshot/sharedropping.
  • This proposal is not a short term development undergoing, it's highly plausible that the porting of the GRC features to c++ will prove difficult & time consuming. This disadvantage applies to all codebase change proposals.

 

Advantages of graphene

  • Security
    • Uplift to a maintained codebase (though that applies to any new codebase we switch to)
    • Low stake weight concern would be gone - only elected witnesses would produce blocks & if unreliable they can be voted out by the network. Block producers would be trusted community members, not every token holder on the network.
    • Witnesses could implement their own external BOINC statistics gathering scripts (with any preferred language), rather than a single solution being built into the client.
    • A single solution built into the client is a disaster waiting to happen - what if a single whitelisted project is compromised and a malicious xml file is placed to crash nodes on the network, or worse?
  • Performance
    • 10k+ transactions per second (TPS) proven on Bitshares, compared to our TPS in the single/double digit range.
    • We don't need 10k TPS (given our small tx volume), we can easily tweak graphene's settings to be less intensive (larger block times, smaller witness node specs required, etc).
    • Reduced quantity of users downloading the stats files each day (as opposed to hundreds to potentially thousands currently) eliminating DDOS concern without needing central proxy server.
    • Witnesses could publish BOINC statistics at will, similar to the current BTS price feed mechanism, negating the need for a daily superblock if we were to average the credit published by each witness when issuing rewards (perhaps at a scheduled time, or more frequent for users with more GRC).
  • Development shortcuts - No need to continue development on current GRC issues such as :
    • POS Staking mechanisms : Replaced by DPOS
    • Hardening/decentralizing voting : Secure voting implemented, we'd need to implement POR weight though.
    • Beacon deletion : Replace with CM/Brod's beacon proposals - providing 100% control of beacon to users.
    • Decentralizing whitelist maintenance responsibilities : The elected committee can vote to change network parameters (such as the whitelist).
    • Witness/Delegates : Graphene uses DPOS, which brings elected witnesses/committee
    • Manual reward claiming
      • If we were to implement a similar distribution mechanism to steem (issuing additional tokens to users based on mag instead of votes) the distribution mechanism would be similar to the older cryptolottery system without the risk of manipulation.
    • No initial investment requirement would ease recruitment of new users, potentially enabling automated BOINC + GRC client wizard setups.
    • The longer we wait, the more features BTS/Steem have which we can inherit.
  • Existing tech
    • Graphene has an easily configurable UI & web wallet interfaces (which anyone can host). We could have BOINC project/team wallet themes to improve community engagement & many web wallet mirrors.
    •   Open source block explorer
    •   Python, C#, Java, (etc...) graphene libraries exist.
    •   If a fork of BTS: User issued assets, market pegged assets, inbuilt exchange. Imagine projects launching their own tokens on top of our network? We could scale this back, removing MPAs for example so as to not step on BTS' feet.


----------------------

 

Directly responding to your proposal document:

 

Quote

Bitcoin has a bright future, with SegWit, Larger blocks, HTLC trading, and is the most popular by market cap.

 

Segwit and larger blocks are not being combined in the one codebase, it has split into Bitcoin Core (segwit) and Bitcoin Cash (Larger blocks - superior). Likewise, a large market cap does not mean technology is superior - there are many scamcoins worth far more than Gridcoin currently is worth, that does not mean they're better than GRC. Segwit isn't desirable technology neither, haha!

Quote

 

In addition, the blockchain balances and wallets are more compatible, so I propose we stay with Bitcoin.

 

Snapshotting a bitcoin codebase coin distribution for claiming coins via private key import has been proven possible by multiple Graphene projects who each created a custom genesis.json file which contained their userbase's public keys (donation addresses | snapshot addresses) & allocated balance. Generating this file would be a compute intensive process, however its use after generated is simple for kickstarting a new blockchain's initial distribution. It's the same process which would be required for any codebase snapshot/sharedrop, we cannot repeat the classic -> researcher burning process as it was an unnecessarily centralized issuance which resulted in a significant sum of funds going unclaimed (foundation funds).

The wallets are more compatible, sure - moving to graphene would result in a loss of current UI, but it would yield an easily customizable UI: https://steemit.com/gridcoin/@cm-steem/gridcoin-s-customised-bitshares-web-wallet-dex

 

 

Quote

Moving on to available technologies, while evaluating different successful models over the entire spectrum, I have come to the conclusion that the most compatible and advanced feature list that would give us the ability to support the future codebase is a rebase of Dash Evolution.


What other cryptocurrency toolkits/codebases did you evaluate?

Do you have notes of feature comparisons?

Do you not think that privacy focused cryptocurrencies are going to be heavily regulated in the near future?

What about graphene, dragonchain, hyperledger, other proposals, etc?

 

I'm concerned that your proposal for moving to Dash involved very few community members, the only masternode thread outwith here has zero replies in the development subforum & the community based alternative codebase threads/github-issues lack your involvement.

 

Quote

Their Masternodes, branded as our Neural Network Nodes, would allow us to transition our investors from proof-of-stake to instead, leasing Neural Network Nodes for an ROI revenue stream, yet our researchers would be paid through the scalable superblocks.


If we do end up going down the route of masternodes, throwing away their technologies brand recognition doesn't sound beneficial - likewise we shouldn't be using the terminology "Nerual Network" because it isn't a neural network, "Gridcoin statistics mechanism" or something functional/realistic would be more honest/accurate.

 

Graphene also funds witness positions at a rate controlled by the committee, I don't think that having to fork up hundreds of thousands of GRC to provide critical network functionality is going to yield many participants - the majority of the Gridcoin userbase does not have the 100k required to create a poll nevermind 400k to destroy for a network role. Masternodes entrench power to the wealthy, as participants have to pay a large sum of GRC for the right to run a masternode where as with Graphene there is not an up front investment required to become a witness/committee member and those who are voted in will have had to campaign to the community & hopefully be the most trusted entities from the community (Benefit being you can vote them out with DPOS, where as you're potentially stuck with masternode operators permanently).

 

Rather than paying all users via large superblocks, would it not be better to investigate the 'Manual Reward Claim (MRC)' proposal suggested by the community? Steemit has the ability to access your rewards with the click of a button, coincidentally a Graphene crypto.

 

Quote

 

This means we could expand to allow secure payments to the entire Boinc network and break our current barrier levels.

 

If MRC was/is made possible, this would apply to multiple codebase proposals.

 

Quote

One of our developers, iFoggz has created a CPID association solution that will allow us to associate a CPID in a scalable way without generating keypairs.

 

Can you provide any documentation on this proposal?

 

I don't see it alongside the other legitimate CPID/Beacon improvement proposals: https://github.com/gridcoin/Gridcoin-Research/wiki/DEV-CPID-Ownership

 

My beacon proposal: https://github.com/gridcoin/Gridcoin-Research/wiki/DEV-CPID-Ownership#cms-proposal

TL;DR:

  • CPID is not a good unique identifier, it can merge/split and be maliciously/accidentally changed.
  • Verifying userID ownership is an improvement over CPID verification, as userID is an unique id which never changes on an individual BOINC project basis (no matter what happens to your BOINC account).
  • BOINC web code done: https://github.com/BOINC/boinc/pull/2134/ https://github.com/BOINC/boinc/issues/2118
    • 2 new PHP pages, only WCG and perhaps Einstein would be incompatible due to homebrew BOINC web server codebase implementations.
    • For projects which don't immediately implement the above PR, we could fall back on CPID onwership proof if an userID which shares the same CPID is successfully validated.
  • No need for a centralized entity moderating/maintaining the beacon system - users could create a new key to reclaim ownership of their BOINC account.
  • Inter/Intra crypto replay protection (can't steal registration here and register elsewhere, likewise replaying my beacon on the GRC network wouldn't work).
  • Cache project's public openssl key & verify userID ownership without contacting the BOINC project once (massive reduction in scraping projects).
  • Requires:
    • Peer review
    • Project admins to perform minimum project maintenance (2 new PHP pages & new openssl keypair).
    • Once implemented in BOINC projects, implementation within Gridcoin/Any-Crypto.

 

Quote

Distinct Features in Gridcoin III:
Masternodes branded as Neural Network (Nodes)

 

NN isn't a desirable branding (it isn't a neural network) & branding not a feature.

 

Quote

  (The Optional) Leasing of a Neural Network node offers a revenue stream for the end user (~400,000 GRC to lease a neural node) pays the investor 30% of each block, giving an ROI of approx. 2-3 years depending on participation count.
  60% of each block goes to Research (equaling payments of 50,000 GRC per day), with 30% going to Neural Network Leases, and 1-10% to the Miner who Mined the block (see PoSE mining algorithm).

 

This is a substantial economic policy change proposal which should not be taken lightly.

 

How long does 'leasing' last for?

 

What happens to the paid funds (burn or centralized fund?)

 

Will neural network leasers be forced to use the one neural network statistics gathering implementation (security risk) or will third party scripts be possible?

 

Why bring mining/pow back into the equation after a long period of POS?

 

What happens to the foundation funds in your proposal?

 

Quote

 

No interest payments

 

Graphene too would eliminate interest payments on the core GRC asset.

 

Quote

 

PoSE (Proof-Of-Service) requires Neural nodes to have external IPs and be online to receive Neural Lease payments, however Researchers do not have this limitation.

 

Graphene doesn't require this, witnesses are largely not publicly visible - users connect to the network via their light/web wallets which interact with full nodes which don't participate in the consensus mechanism.

 

Quote

Internal Green PoSE CPU Miner for each block eliminates Proof Of Stake, with an ASIC resistant algorithm that compensates the miner with lower payments for more work, so as to decrease miner competition levels. (The block payment to the miner is 1-10% depending on work performed, with 1% being Most Work performed, 10% being Least work performed to solve the block). This is a variation of X11 called POSE (Proof-of-service) that breaks ASIC compatibility.

 

Reintroducing proof of work is a huge step back technologically; POW is responsible for significant amounts of pollution, it will negatively impact BOINC computations (mind we saw a huge increase in computing power pointed at BOINC after we left Scrpyt for POS?). Any future proposed codebase should not involve proof of work (IMO).

 

Quote

Secure voting by Neural Nodes (IE voting requires Neural Node lease)

 

I completely object to this, we already see almost nobody making polls because of the local 100k balance requirement, now imagine poll participation if the required balance was 4 times the current amount? It'd be pushing many voices entirely out of the picture within the Gridcoin network, further empowering the rich in the GRC network.

 

Graphene has coin based voting for committee and witness members; the committee then have the abiltiy to vote amongst themselves the network rules (changing witness payouts, block timings) and with the gridcoin graphene proposal would be responsible for maintaining the list of whitelisted BOINC projects in a majority vote basis manner. How we'd go about implementing casual polls or community project polls (for the committee to act upon) would require some original development - see: https://github.com/FollowMyVote/StakeWeightedVoting
 

Quote

 

Secure Superblocks using Dashs budget system for daily research payments

 

One cannot evaluate the security of a system prior to its implementation, as far as recent history goes security is not your strong point (not try to be a dick, rather making an accurate observation given the last 3-5 months history).

 

Quote

InstantSend allowing Instant Payments
2.5 Min block times to reduce future BlockChain size


Graphene has 2-3 second block times (which can be tweaked to reduce blockchain bloat) point is that fast payments aren't unique to dash & TBH not a problem right now in Gridcoin.

 

Quote

More efficient core memory structure (IE all history and payment calculations stored in Neural Network)

Applicable to any C++ ported NN replacement, though is it neccessary to track history before the previous payout if we're able to achieve 100% accuracy of payments?

 

Quote

New unlimited scalability to allow us to expand CPID reach for the entire boinc network (handling all CPIDs in boincstats), and also for research payments: Our daily superblock pays every researcher via a superblock payment drop, therefore no Pools are required.

'Unlimited scalability' is likewise achievable with the userID ownership beacon proposal, a reduced NN participation rate and the suggested manual reward claim functionality. I'm not massively confident in daily superblock payouts given recent historical insecurity of superblocks and the old cryptolottery system (which aimed to reward many users in the one block a couple years back).

 

Quote

Unlimited scaling using the iFoggz CPID associator system. (No need for research keypairs).

Repeating myself here, but CPID isn't a perfect identifier - userID beacon proposal > CPID. I'd love to see some public documentation/proposals to peer review this new CPID associator system.

Quote

 

The new architecture minimizes the possibility of Security Issues.

 

New codebase wise yes, but any changes we make to the dash codebase invalidate this claim until we have a professional security audit - like the recent security audit made entirely public except performed internally/privately at our discretion.

The proposed short/concentrated development (and closed participation planning) period likewise does not inspire confidence in the security of the proposed network.

 

Quote

Denravonskas Total Credit Delta class feeds the new neural network, instead of the current Web Magnitude gatherer.

Sounds good, moving away from RAC based magnitude towards total credit delta is a great proposal for any codebase.

Quote

 

More bitcoin compatibility as we have latest bitcoin script library.

 

Certainly a legitimate advantage, there are many bitcoin based libraries which the current gridcoin codebase is entirely incompatible with. That said, there are several Graphene based libraries which we can benefit from too.


 

Edited by C.M

^ Smash that upvote button! ;D

Follow me on:

Github   Twitter   Steemit   SoundCloud (Hangouts)

 

Share this post


Link to post
Share on other sites
2 hours ago, TheCharlatan said:

The proposal for PoSE sounds insecure and near unachievable. How would you ensure that the network is not periodically attacked and certain nodes hindered from the voting process? In Proof of Work this is trivially evaded, since one miner would take much longer to find a single block than many miners, which means that you'd constantly have to attack the entire network over weeks for it to succeed - an impossible scenario given the resources this would take.

Mining is supposed to secure the blockchain. Target nonces are made purposefully hard to have competition between miners on who gets to solve the next block. It is an arms race by design and works well like this. If you invert this scheme by giving higher monetary rewards for lower nonces, you remove the incentive for innocent miners to do proof of work (as you stated above). An attacker however would not care about this small reduced reward, if this would mean, he could just increment nonces faster to be able to successfully mine several blocks in succession and execute a successful double spend. You even implemented a similar scheme before in pre V8 gridcoin and it was successfully exploited on livenet by doing more proof of work.

I cannot see how this would survive an attacker, and therefore reject this proposal in its current state.

PoSE is time tested on the other masternode based coins and works well (its called Sentinel in dash),  (Ive had it running on my network for a couple months now and testing it on a crypto project for some crypto friends) and know how the inner workings work.  Since all PoSE votes are stored in a local database, each neural node is actually aware of everyone elses votes, by tracking something called a gobject.  Its similar to a txid.  If its your turn to vote for yourself to be paid your neural ROI, you must be online to do it.  When you send your Original neural escrow (IE the 400,000 GRC), we know your IP, TXOUT, and VOUT and you are registered permanently (until you pull out as a leased node).  When you vote, you must vote from your escrows Public IP.  If you dont vote, you dont get paid your neural lease revenue.  If you do vote, the other neural nodes see the vote and chalk it up as a Yes and allow you in the payment queue.  Other nodes pay you, so you actually dont pay yourself.  So its 100% secure and does prove youve been online from your static external IP.  Btw, in order to run a neural node in the new system, you have to have a static public IP and run Grid POSE bulldog along with gridcoind.

 

PoSE mining does not reduce the security of the blockchain, since difficulty still exists.  If one were to increment nonces faster and go for the lower reward, the diff would rise and they would be unable to continue doing that on the next block. 

I think we only need two rules to implement POSE mining:  a sliding 1-9% reward based on the nonce value (lower is higher payout meaning you can solve the block by waiting for seconds to tick by), and a decreasing difficulty ADJUSTMENT [to the current difficulty] for each minute elapsed from pblocklast->pblockcurrent.  Since we have 2.5 min block targets, the block would be in a range of 0% - 90% easier to solve per minute that ticks by, increasing in easiness by 10% per minute.  This pblockcurrent-pblocklast diff adjustment parameter is a hard parameter - it is Not a security problem to do this, we just have to enforce block max elapsed to be 15 mins max in the future.  These two rules should implement POSE quite well.  So far I have not heard anything alarming here.

 

 

Share this post


Link to post
Share on other sites
1 hour ago, jringo said:

 

 

How will this affect superblock size?

-> I was thinking one superblock per 24 hours, for every recipient owed over 5 grc.  We can ask Marco to accrue smaller balances.

 

1 hour ago, jringo said:

Please explain.  Are we not open-source development?

->  I do not mind being called Founder of Gridcoin (instead of CTO), until I hand over the Github to someone else and hand over the foundation funds to our next entity, I really do not need to be "CTO".  I switched my title to CTO when I found myself making payroll and trying to foster a new Lead Dev to replace me.

 

 

 

1 hour ago, jringo said:

Would appreciate non-biased phrasing in the client poll. = )

-> Thats a good point, I guess there are a lot more proposals that Tomas'.  If I can still edit it I can expand it.

 

 

How does iFoggz proposed association solution compare to CM's and Brod's? 

-> iFoggz can you please chime in.  My only succinct answer is its secure and has unlimied scalability and we can do it today.  We know how it works and iFoggz has been writing it at night and its coming together.

 

-------------

 

Personally, I am here for Proof of Research -- a blockchain owned by those who contribute to a distributed computing network which focuses on science, technology, and data analysis.  A blockchain in which at least half of the blocks are made by those who contribute to this distributed computing network.  Masternodes seem to contradict the original purpose of Gridcoin.  It seems to operate in a way that would greatly benefit extremely early adopters and financially well off entities, while limiting access for new users and those who contribute processing power to the network.  Investors are great, but should they own the blockchain?  I feel like I'm missing something.  Am I missing something?

-> Wait, the new solution would be mined by everyone, random miners, neural network participants and researchers.  The key is 60% of every block would make the 50,000 daily GRC payments.  The whole idea of this 30% system is to provide an avenue for investors to buy and hold GRC, and to put GRC to work to earn them ROI if they have a balance sitting.   I dont mind it if the community wants a 25,000 GRC investment (thats only $800).  Its a way for people to invest it and not withdraw it and go to other coins for ROI.

 

 

1 hour ago, jringo said:

Again, I am here to see those who contribute to the distributed computing network benefit the most and hold the most power.  In other words, if I were a large scale research facility, would I make blocks based on the computational power I contribute to BOINC and would my contributions to BOINC grant me a seat on the voting board, weighted based on my contributions relative to other contributors?  Or would I have to allocate tens or hundreds of thousands of dollars of my already tight budget in order to "earn" my position on the voting board?  (It currently costs USD $585,000 to lease a DASH masternode).  How do masternodes seek to advance the primary goal of Gridcoin?  Do masternodes limit the potential of a well financed entity buying control of Gridcoin while contributing nothing to the computing network?

-> Masternodes provide the POSE of the network, while Researchers can do as they please and not have to have external IPs.  Neural Nodes tend to make GRC more valuable, as people would have a propensity to stay in.

 

In its current state and presentation I would support the "piecemeal" improvements, but I look forward to being enlightened.

-> I understand.

 

As a side note, the name "Neural Network" is misleading and I think it ought to go.  

-> I want to see the name Neural Network be a cornerstone in Gridcoin, and want to see the network grow into what it was meant to be - an array of nodes that shares files among each other, gaining intelligence, and housing the boinc/research history of each CPID for historical reports.  It can grow into the Neural Network.  Dont be brainwashed by people who fall short of the original vision.  Maybe in the future it will run a smart contract and truly be neural.

 

1 hour ago, jringo said:

 

Share this post


Link to post
Share on other sites

 

Regarding holding the vote here instead of the client, I didn't want anyone who is currently bashing gridcoin in the slack channel to say the vote was hacked.  So this channel will need to work!

 

Ill leave the vote up long enough to get at least 100 users votes, and we will take it from there.  If we have a supermajority and over 100 votes, then I assume its a valid vote that reflects the feelings of the Gridcoin community.

 

JRingo: I cant edit the title, which is good, but for all others:  Just consider option #2, the No as :  Explore any other proposals, either keeping the current codebase and fixing it piecemeal, going with Brods proposals, Or going with Quez/JRingo/CMs potential proposals.

 

I created a vote to help us decide what the escrow requirement should be for the Neural Network Nodes:

 

 

Thanks everyone.

 

 

Share this post


Link to post
Share on other sites

I think there is some confusion. 

 

1. Calling one proposal "piecemeal" which keeps the same codebase, which it is not by my understanding, while exclaiming the other proposal -- that is the bias I am concerned about.  Production polls must be language neutral.  You do bring up a good point, however: there are 3 proposals under public development that I can think of off the top of my head, excluding this one.


2. 

Quote

I didn't want anyone who is currently bashing gridcoin in the slack channel to say the vote was hacked.  

 

What bashing in the slack channel?  I do not understand.  The slack channel is beyond cordial, and infinitely more productive than cryptocurrencytalk forums.

Quote

So this channel will need to work!

 

This is meant to be a production poll?!  Hosting a production poll on a highly inactive forum is inappropriate, restrictive, and easier to manipulate than a client based poll.  Would you consider hosting a poll on the client after ample time for community feedback and after your masternode proposal is detailed at length?  We still need economic effects, governance effects, and possible negative consequences, at least.

 

3. How will this proposal affect superblock size?  How will the superblock size be affected if there are 1000 addresses receiving GRC from the superblock once per day?  What about 10,000?  100,000?  Does this scale?

 

4.

Quote

The whole idea of this 30% system is to provide an avenue for investors to buy and hold GRC, and to put GRC to work to earn them ROI if they have a balance sitting.

 

How does this help the primary goal of Gridcoin?  In its current state, it sounds like a way to inflate the price of Gridcoin while doing little or nothing to encourage its principal use-case/utility: Data Analysis through a distributed computing network with open results and open access.  There are ways to encourage holding GRC other than what is proposed here -- ways which seek to advance the principal use-case and encourage holding as a secondary effect. 

 

4.1. Do masternodes remove a major reason to hold onto your GRC - voting?

 

5.

Quote

Masternodes provide the POSE of the network, while Researchers can do as they please and not have to have external IPs.  Neural Nodes tend to make GRC more valuable, as people would have a propensity to stay in.

 

By my current understanding, masternodes centralize control of development with early adopters and financially wealthy entities.  This actively excludes the main target audience of the product from influencing the product's development.  To me, the target audience of Gridcoin is researchers, both by the literal definition and by our definition (people who contribute their Idle Processing Potential to BOINC).

 

6. Neural Network is hype.  If at some point the mechanism becomes a true neural network, we can always change the name back.  As it stands now, it is a buzz word that makes it difficult (impossible) to pitch Gridcoin to reputable research facilities, universities, and their researchers.  They know better.


7. While I do not support a switch to graphene at this time, I would like to 2nd many of the questions posited by @C.M

 

8. Why are we going back to PoW?  This seems like it could be a disaster.

 

9. 

Quote

Dont be brainwashed by people who fall short of the original vision.

 

What is the original vision for Gridcoin?

 

10. Do you have documentation regarding all the tests  you have done in your lab?

 

11. Have we given up on Proof of Research?

 

The process regarding this proposal -- the lack of information, lack of discussion, lack of outreach, presentation, and using this forum as a deciding vote -- does not seem beneficial, particularly while we have a voting mechanism built into an open source algorithm.

This makes me fear for my admittedly sizeable investment and more importantly, the future of Gridcoin.  Please do what you can to tranquilize this fear.

 

Thank you.

Edited by jringo

Share this post


Link to post
Share on other sites
2 hours ago, Rob Halförd - (Gridcoin) said:

Regarding holding the vote here instead of the client, I didn't want anyone who is currently bashing gridcoin in the slack channel to say the vote was hacked.  So this channel will need to work!

 

What on earth are you talking about?  Are we completely abandoning our established poll mechanism for some perceived "bashing" that does not exist?  WTF?

 

Edit:

 

Quote

The process regarding this proposal -- the lack of information, lack of discussion, lack of outreach, presentation, and using this forum as a deciding vote -- does not seem beneficial, particularly while we have a voting mechanism built into an open source algorithm.

This makes me fear for my admittedly sizeable investment and more importantly, the future of Gridcoin.  Please do what you can to tranquilize this fear.

 

I share @jringo's concern.

Edited by barton26

Share this post


Link to post
Share on other sites
4 minutes ago, barton26 said:

 

What on earth are you talking about?  Are we completely abandoning our established poll mechanism for some perceived "bashing" that does not exist?  WTF?

What Im talking about is Tomas Brods bashing comments about the existing qt-clients poll system.  Ill put a vote in the client now that mirrors this one, this way we can vote on both.

 

 

 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


  • Recently Browsing   0 members

    No registered users viewing this page.

×

Important Information

By using CRYPTOCURRENCYTALK.COM, you agree to our Terms of Use.