Topic: Research for decentralized Online Market (Ebay/Amazon-like) (Read 744 times)

Thank you for the input so far.

I am working on the theoretical concept now which I will publish here as a first draft as soon as I am clear about the implications.

If you have any other ideas, concerns, tips, etc, just reply here - no matter how stupid they may sound.

Greets

I suggest not to use various blockchains but instead use Bitmessage's chans for listings. Bitmessage requires that the one who sends a message to a chan performs proof-of-work. This will deter spammers from posting garbage.

You definitely don't want to use fiat currencies for a decentralized market. Use bitcoin.
There is no need for the buyer to send the BTC to the escrow to hold on to. Bitcoin has a feature called multisig address which allows using escrow without entrusting escrow with your BTC.

What I described here is what probably will end up getting implemented on the OpenTransactions platform. I heard FellowTraveler (the main developer of OpenTransactions) speak about decentralized markets using Bitmessage.

There are a number of challenges faced when designing an escrow system.  The main ones are 'the seller not sending the goods' and 'the buyer not releasing the money after receiving the goods'.  It is therefore necessary to ensure both parties have disincentives for not fulfilling their obligation.
An example might be the buyer pays twice the amount of the purchase price into escrow and receives the extra portion (equivalent to the purchase price) back when he/she releases the money to the seller.  The seller would also be required to deposit the purchase price into escrow, this would be returned to the seller when the buyer agrees that the transaction is satisfactory and complete.  In this way the seller would lose money equivalent to the purchase price if he/she didn't ship the goods, and the buyer would pay twice the purchase price if he/she received the goods and refused to release the money to the seller.  Now both parties have an incentive to complete the transaction and a disincentive to try and cheat the other person.  This would allow a peer to peer escrow system to work well for most people, without having to worry about being cheated.  There are still challenges however, what happens if the buyer dies, he/she will not be able to release the money, or the seller dies and is unable to ship the goods.  This would create a situation where both parties money becomes stuck, and unable to be returned.  You could have a time limit set to automatically return funds held in escrow if neither the buyer nor the seller object within a given time frame, but this would require a trusted 3rd party or network node.

It is good to see people creating new services and things to support Bitcoin and other crypto currencies.

Yup I know it can. Some sort of trust score/feedback needs to be in place. So that you know not to trust more that $30 worth of bitcoin to someone without any past score.

The problem with escrow release on delivery that is you can ship an empty box, a brick, etc. Until tracking includes an x-ray of the package... escrows can be scammed like BitMit was in its final days (empty box, flat rate USPS priority).

You can close the escrow when tracking number says it was delivered. This can be automatized. For most carriers. The rest are out of luck :-)

You still need some level of customer service, but it can be decentralized. It can be in a form of taking a 1% transaction fee, and then distributing the total amount of collected fees automatically between volunteer customer service (based on activity) and volunteer developers (based on GitHub activity), etc.

Hi,

I am currently researching the possibility and acceptance of an independent open-source and decentralized online market place working with crypto currencies for my master thesis.

The goal is to have an alternative to ebay/amazon.
Please note: This is primarily a research and might or might not be be actually realized.
This will basically depend on the difficulty and whether all problems can be solved.


Currently I am still at the very beginning and need lot's of input to proceed.
What I was thinking so far:

1. TECHNICAL THOUGHTS:
---------------------
In order to not have participant download tons of gigabyte of blockchains, I was thinking about dividing it into different blockchains.

1.1.) There will be one blockchain that holds all active listings, one blockchain which holds the reliability of buyers and sellers and a third blockchain which will hold all past transactions.

1.1.1) Active Listing Blockchain
This one will only always have the active listings and cut off past ones, so you will not have to download a lot of stuff to get started.

1.1.2) Reliability Blockchain
Can optionally be downloaded in order to see how reliable a user has traded so far.

1.1.3) Past transactions
Can optionally be downloaded to see all past buys/sells.

It should also be possible to see all listings/ratings (blockchains) via a server (just like blockchain.info does for bitcoin). Trading from a server-side platform can also be possible, but everything must/will be in synch with the decentralized blockchain.

1.2.) To use the platform I was thinking about a client that will synchronize the blockchain(s) and prepare the data on a local webserver to the user.


2. IMPLICATIONS:
---------------------
There are still some questions that arise:

2.1.) Buying/Selling
There must be something that will function as a virtual escrow.
If I buy something on that market and transfer the money into the system, it must be holded there and only released once I have successfully received my goods. Then my private key can unlock the "kept" money and it will be transfered to the Seller.

The problem I currently see: What will happen If I never release the money, even If I have successfully received the goods.
There will be no central instance where the Seller can complain to.

2.2.) Currencies
Here I am not sure yet how to include all the crypto-currencies or if I have to use an independent crypto for buying/selling.


Anyway, this is everything I have so far, so please post any ideas and critics if you wish.
Not only about the technical stuff, but also about the acceptance about this theoretical approach.

Thank you