Author Topic: DPOS chain based ratings and reputation  (Read 2836 times)

0 Members and 1 Guest are viewing this topic.

Offline Troglodactyl

  • Hero Member
  • *****
  • Posts: 960
    • View Profile
AirBnB is a forefront p2p home rental company that has too many regulatory challenges from the government. Making a DAC from a similar model might also end up risk of having to explain to the regulators how the business would operate in a legal manner.

These businesses operate internationally, and it's inefficient to try to manage them centrally and organize compliance with all the different law sets from the center.  It would be much more efficient to use a DAC for the core business model and let the individual service providers work out the legal compliance in their respective jurisdictions.

On further analysis and discussion with modprobe, it seems like it would be more flexible to use a reference system and likely a dedicated storage DAC for the review content.  In place of the reviews themselves, each business chain's transactions could support cross chain references including an identifier for the chain being referenced and the transaction ID to reference in that chain.

Modprobe suggested an escrow payment system for the storage DAC in which a service provider can prepay to reserve storage and provide the customer's public key.  The customer would then have a limited time to leave a review in that reserved space, or the escrow would be returned to the provider (minus transaction fee).

Offline jwiz168

  • Sr. Member
  • ****
  • Posts: 409
    • View Profile
AirBnB is a forefront p2p home rental company that has too many regulatory challenges from the government. Making a DAC from a similar model might also end up risk of having to explain to the regulators how the business would operate in a legal manner.

Offline Troglodactyl

  • Hero Member
  • *****
  • Posts: 960
    • View Profile
For context I was thinking about DACs to compete with Lyft, Uber, and AirBnB when I thought of this.  Both parties being able to leave reviews would certainly be desirable for such purposes.  I think reputation tracking for other purposes is inherently very different from trust for delegates, so I'm doubtful that direct delegate voting is really applicable.  I think a review cost is necessary to kill spam, but I think the actual trust model would have to be (somewhat) closer to the traditional web of trust system.


Offline JoeyD

Would it be possible to adapt the delegate voting mechanism and tailor it to fit such a reputationtracking blockchain? (DART, oh damn, I can hear the conspiracy theorists sharpening their pens and lighting up their interwebs already)

For e-bay- and silkroad-like ano-/pseudonymous marketplaces or trading in general some form of non-centralized corruption resistant reputation system for both customers and sellers might be a diserable thing to have.  Maybe another way of doing it would be to set it up like the insurance model, where the agents work like escrow agents and they keep track of reliability and risks on their own. If the escrow/trust-agents can be transparently monitored for corruption and hired/fired depending their performance, it might make things more practical.

If the separate reputation system is opt-in and not locked-in or obligatory then it might not be a bad thing. Higher web-cred on the DART chain would possibly allow for trading to happen with less delays and procedures as usual while still offering a bit of fraud-prevention to both parties involved.

EDIT
Not easy to come up with such a system I'd imagine, if it's even possible to do at all. It is one of the DAC applications I'm more interested in than the pure monetary ones. I'm not very happy with most current systems where gatekeepers get to decide peoples fate, but where the system is one-way only. If both (instead of one side only) parties could be judged objectively and transparently according to performance I could think of a whole range of applications that could benefit from that.
« Last Edit: May 13, 2014, 09:19:17 am by JoeyD »

Offline Troglodactyl

  • Hero Member
  • *****
  • Posts: 960
    • View Profile
This is a solid idea.   The price per-byte is already determined by selecting a fixed limit on total chain size and adjusting fees to control demand.   

The biggest challenge is lack of privacy.  This is certainly viable for many uses though.

This works if it's acceptable for feedback to disappear after the inactivity fee interval, but not if the goal is to establish reputations that are preserved as long as the ID remains registered.

Privacy will always conflict with reputation.  I don't think there's really a way to have trustworthy reputation without revealing a legitimate history establishing that reputation.  This is why it should always be optional.  Some tasks call for anonymity, others call for establishing a good name.

I think this kind of thing may be better done off chain as there is no reason for it to be stored forever on chain or propagated via the P2P network. 

I think it's as essential to preserve the integrity of reputations as the integrity of transaction history.  Where can this be stored off chain that can be trusted to be both reliable and unbiased?  Am I missing something here?

Offline legendface66

  • Full Member
  • ***
  • Posts: 69
    • View Profile
No I disagree... you could have a public DB of both signature dependency information and "receipts" that are "signed" (stamped?) by DACs AND all sorts of hierarchical key inference trickery. It doesn't need to be a DAC but it would effectively leech-mines many other DACs.
I am not sure about the long-term plans for keyhotee but from what I understand it doesn't lay a good enough foundation to be able to quantify trust and thus have effective credit DACs

That's unfortunate, I had envisioned Keyhotee as, in part, a reputation platform that may ultimately integrate an insurance history rating, credit score, ebay rating, etc. Maybe with the ability to grant limited permissions to parties concerned with specific reputation elements.

Offline Troglodactyl

  • Hero Member
  • *****
  • Posts: 960
    • View Profile
Leaving reviews is voluntary, so if you don't want a transaction linked to your persistent ID, you simply wouldn't sign with it and wouldn't leave a review.  If you want to establish a domain specific reputation (for the business model of the particular blockchain) for your domain specific identity (registered as a name in that blockchain), then you would have that option.  Each review only associates the persistent ID with the addresses involved in the associated transaction.

If people are trusted to store their own reviews, they can delete the negatives.  If the reviewer stores reviews, they have to be sought out from all previous customers by every potential customer.  Why bring in a corruptible and failure prone authority to store all reviews off chain if an on chain solution can be implemented with the added benefit of increased dividends?  Who would pay for off chain storage?

Offline toast

  • Hero Member
  • *****
  • Posts: 4001
    • View Profile
  • BitShares: nikolai
No I disagree... you could have a public DB of both signature dependency information and "receipts" that are "signed" (stamped?) by DACs AND all sorts of hierarchical key inference trickery. It doesn't need to be a DAC but it would effectively leech-mines many other DACs.
I am not sure about the long-term plans for keyhotee but from what I understand it doesn't lay a good enough foundation to be able to quantify trust and thus have effective credit DACs


@ newest post: clever key schemes + trust DAC allows for "private reputation"...
Do not use this post as information for making any important decisions. The only agreements I ever make are informal and non-binding. Take the same precautions as when dealing with a compromised account, scammer, sockpuppet, etc.

Offline bytemaster

For many DACs, it would be beneficial to have a built in ratings system to establish reliable reputation.  I propose that share transfer transactions allow the inclusion of additional transaction fee to purchase storage space in the blockchain exclusively reserved for a review of the receiver's performance by the sender.

Basically in addition to the standard transaction signatures by the sending addresses, an additional signature could be added from a DPOS registered name, attaching the transaction to that pseudonym ID and claiming the right to leave a review transaction signed by the chain registered ID after the business was concluded.

The reviews would have a different standard format specific to the needs of each chain's business model, and would persist in the chain tied to the reviewed ID.

Along with providing an indelible rating system, which increases the value of the platform, this would provide incentive for additional transaction fees, increasing dividends to shareholders.  Service providers would be motivated to provide discounts to those willing to review in order to build their reputations, and reviewers would not want to sacrifice their own reputations by providing fake reviews, or by failing to review after opting for review space in a payment.

I'd like to hear thoughts on both the economic model and the plausibility of implementing this with DPOS.  The biggest issue I see is finding a price for permanent blockchain storage space.

This is a solid idea.   The price per-byte is already determined by selecting a fixed limit on total chain size and adjusting fees to control demand.   

The biggest challenge is lack of privacy.  This is certainly viable for many uses though.

Privacy is in the design of blockchain and if ever it is being implemented, no additional fee will be charged.

Lack of privacy comes from reusing addresses or identities.   If anything links your name to your real world identity everyone can see all of your transactions.  Without being able to see all of your transactions the ratings are near meaningless. 
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline jwiz168

  • Sr. Member
  • ****
  • Posts: 409
    • View Profile
For many DACs, it would be beneficial to have a built in ratings system to establish reliable reputation.  I propose that share transfer transactions allow the inclusion of additional transaction fee to purchase storage space in the blockchain exclusively reserved for a review of the receiver's performance by the sender.

Basically in addition to the standard transaction signatures by the sending addresses, an additional signature could be added from a DPOS registered name, attaching the transaction to that pseudonym ID and claiming the right to leave a review transaction signed by the chain registered ID after the business was concluded.

The reviews would have a different standard format specific to the needs of each chain's business model, and would persist in the chain tied to the reviewed ID.

Along with providing an indelible rating system, which increases the value of the platform, this would provide incentive for additional transaction fees, increasing dividends to shareholders.  Service providers would be motivated to provide discounts to those willing to review in order to build their reputations, and reviewers would not want to sacrifice their own reputations by providing fake reviews, or by failing to review after opting for review space in a payment.

I'd like to hear thoughts on both the economic model and the plausibility of implementing this with DPOS.  The biggest issue I see is finding a price for permanent blockchain storage space.

This is a solid idea.   The price per-byte is already determined by selecting a fixed limit on total chain size and adjusting fees to control demand.   

The biggest challenge is lack of privacy.  This is certainly viable for many uses though.

Privacy is in the design of blockchain and if ever it is being implemented, no additional fee will be charged.

Offline bytemaster

I think this kind of thing may be better done off chain as there is no reason for it to be stored forever on chain or propagated via the P2P network. 

For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline toast

  • Hero Member
  • *****
  • Posts: 4001
    • View Profile
  • BitShares: nikolai
I like the idea that reputation can be determined by each user's personal weighted average from reputations in multiple DACs, and DACs can be made aware of other DAC's namespaces and piggyback off of their rep. I need to sit down and write my proposals for the .key and .tru namespaces...
Do not use this post as information for making any important decisions. The only agreements I ever make are informal and non-binding. Take the same precautions as when dealing with a compromised account, scammer, sockpuppet, etc.

Offline Count of La Mancha

  • Newbie
  • *
  • Posts: 15
    • View Profile
For many DACs, it would be beneficial to have a built in ratings system to establish reliable reputation.  I propose that share transfer transactions allow the inclusion of additional transaction fee to purchase storage space in the blockchain exclusively reserved for a review of the receiver's performance by the sender.

Basically in addition to the standard transaction signatures by the sending addresses, an additional signature could be added from a DPOS registered name, attaching the transaction to that pseudonym ID and claiming the right to leave a review transaction signed by the chain registered ID after the business was concluded.

The reviews would have a different standard format specific to the needs of each chain's business model, and would persist in the chain tied to the reviewed ID.

Along with providing an indelible rating system, which increases the value of the platform, this would provide incentive for additional transaction fees, increasing dividends to shareholders.  Service providers would be motivated to provide discounts to those willing to review in order to build their reputations, and reviewers would not want to sacrifice their own reputations by providing fake reviews, or by failing to review after opting for review space in a payment.

I'd like to hear thoughts on both the economic model and the plausibility of implementing this with DPOS.  The biggest issue I see is finding a price for permanent blockchain storage space.

This is a solid idea.   The price per-byte is already determined by selecting a fixed limit on total chain size and adjusting fees to control demand.   

The biggest challenge is lack of privacy.  This is certainly viable for many uses though.



Offline Troglodactyl

  • Hero Member
  • *****
  • Posts: 960
    • View Profile
For many DACs, it would be beneficial to have a built in ratings system to establish reliable reputation.  I propose that share transfer transactions allow the inclusion of additional transaction fee to purchase storage space in the blockchain exclusively reserved for a review of the receiver's performance by the sender.

Basically in addition to the standard transaction signatures by the sending addresses, an additional signature could be added from a DPOS registered name, attaching the transaction to that pseudonym ID and claiming the right to leave a review transaction signed by the chain registered ID after the business was concluded.

The reviews would have a different standard format specific to the needs of each chain's business model, and would persist in the chain tied to the reviewed ID.

Along with providing an indelible rating system, which increases the value of the platform, this would provide incentive for additional transaction fees, increasing dividends to shareholders.  Service providers would be motivated to provide discounts to those willing to review in order to build their reputations, and reviewers would not want to sacrifice their own reputations by providing fake reviews, or by failing to review after opting for review space in a payment.

I'd like to hear thoughts on both the economic model and the plausibility of implementing this with DPOS.  The biggest issue I see is finding a price for permanent blockchain storage space.