the buyer hasn't any incentive to pay the seller

[n1c01a5]

for the arbitrable payment https://github.com/kleros/kleros-interaction/blob/master/contracts/standard/arbitration/ArbitrableTransaction.sol the buyer hasen't any incentive to pay the seller (except maybe to keep good relationship with him) in contrary because it costs gas. So the seller must do always raise a dispute and do a timeout request, 2 txs.
I propose to solve it the buyer makes a deposit and if he doesn't pay the seller in this case, he looses it.
An another one would be to make payment automatic if no challenge within a defined period of time with action callback (alarm clock).

[clesaege]

We can make a smart contract implementing that with a 1 contract- multiple transactions in order not to redeploy every times.

[n1c01a5]

@epiqueras proposes to add a timeoutTrade with a withdraw() function and a deposit to pay the gas to solve this issue
The withdraw() method can be called by the seller when the timeoutTrade is out and he gets the deposit gas from the buyer

[epiqueras]

I think we can get rid of pay() since the buyer will never have an incentive to pay before the timeout unless we penalize him. But, why penalize and make the code more complex if we can just have the seller withdraw, and have him take the gas he will spend on withdrawing into account when setting the price of the transaction.

This means we only need to remove pay(), add withdraw() onlySeller ifPastTimeout ifNotDisputed, and change the contract so it supports multiple transactions (no redeploy), and make the seller initialize transactions instead of the buyer, because he is the one who should decide the timeout.

struct Transaction {
  address seller;
  address buyer;
  uint256 amount;
  uint256 timeout;
  bool disputed;
}

Transaction[] transactions;

What do you think? @clesaege

[clesaege]

@epiqueras @n1c01a5 Being nice is sometimes an incentive, also for repeated business.
I don't think the buyer should deposit gas, that increases complexity.

For the other comments yeah, it's not really a change, it's a whole new contract :)

[epiqueras]

@clesaege So you are ok with the seller being able to withdraw the money after the timeout?

[clesaege]

@epiqueras If not disputed yeah.

[eburgos]

Hi, regarding this issue. Please review #32 . I'm curious about if I solved it or I messed it up 🤣

Regards,

[clesaege]

Yeah, here there is a timeout so if the payer does not do anything, the transaction is processed.

[clesaege]

In the multiple it works due to timeouts.

[epiqueras]

Yeah I put it in the spec.

…

On Fri, Jul 27, 2018 at 12:31 PM Clément Lesaege ***@***.***> wrote: In the multiple it works due to timeouts. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#27 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ASRQaCfjAJlGoUFXx1hW4N9V7dXLjoJfks5uK05sgaJpZM4Sv9Ro> .