3 ms·

> Reversability and “right to deletion” will trump trustless because most of the actual market participants want the former and not the latter. Does reversibil

> Reversability and “right to deletion” will trump trustless because most of the actual market participants want the former and not the latter. Does reversibility actually matter anywhere outside of consumer merchant payments? If you send a wire transfer, it can't be reversed. If you write a check, you can't reverse it once it's been cashed or deposited. You can ask for a stop payment before a check has cleared, but you can also be accused of check fraud if you ask for a stop payment under the wrong circumstances. Outside of the payments domain: if you buy stock on the stock market, you can't reverse a filled order. You can sell your stock, but you need to sell to a new buyer at a new market price. Even in the modern fintech space, not all transactions are reversible. A Zelle transaction can't be reversed, for instance: https://money.stackexchange.com/questions/86920/can-a-zelle-bank-transfer-be-reversed-or-denied-after-credit-has-been-added https://money.stackexchange.com/questions/86920/can-a-zelle-... All this is to say that true reversibility is only available in certain cases. It is something that financial institutions offer when they know that the funds recipient is a merchant in a captive relationship with a financial institution, where the merchant will most likely be able to fund the reversal by offsetting future unsettled payments that are yet to be reimbursed to the merchant. It works only because merchants are dependent on financial institutions, and do not have the negotiating power to demand instantaneous, final settlement. I very much doubt that financial institutions will want to (or will even be able to) extend reversibility to other transaction types just so as to deploy it as a strategy to prevent open blockchains from becoming predominant.

codefreeordie 4y ago
I can't find the article right now, but even wire transfers can be reversed under the right circumstances. Reversibility is actually necessary in ~everything under the right circumstances. There is a surprisingly high volume of wire transfer hold-harmless agreement traffic (this is banks agreeing to reverse wire transfers in exchange for transferring the liability if the transfer was not actually bad/fraudulent)
theptip 4y ago
> Does reversibility actually matter anywhere outside of consumer merchant payments? Yes. For example, the stock markets can reverse a day's trading when it's decreed that there was a flash crash. And in contrast to smart contracts, all business contracts are in some sense reversible in extremis, in that you can go to court and consider the agreed upon intent if there was a typo in the paper, and resolve ambiguities after the fact based on precedent, instead of contract bugs resulting in all of your money getting stolen. I think we probably agree in that reversibility isn't important in _most cases_ (the mainline happy path), but I think it's really important when you do need it. > extend reversibility to other transaction types just so as to deploy it as a strategy to prevent open blockchains from becoming predominant That's not quite the claim I was getting at. Perhaps if I'd said "tighter control, for example being able to implement reversibility and right to deletion” it would be more precise. The core of my claim is not that those two are the killer features; the feature that will cause decentralized options to lose in the end is more generally the control that a centralized entity has over the end-to-end experience. Those two are just prominent examples of the kind of feature that are difficult or impossible to implement on the blockchain, which are actually strongly desirable for the market makers. Another example would be KYC, which is hard to solve at the protocol level. My overall prognostication here was really that the existing big players will pick and choose the bits of the blockchain systems that are actually innovative and useful (I'd highlight smart-contract execution engines and true API-based transaction instructions for example) while dropping the "distributed" bit to maintain tighter control and enable broader consumer adoption. Analogously to how Google et. al. have transitioned the internet from a distributed/federated "everyone hosts their own server and communicates over protocols" world to a centralized "everyone uses hosted services for everything" world. And consumers seem to prefer the latter, if the FAANG valuations are anything to measure this by!
Adrox 4y ago
I work in the stock market, there is dozens of rollbacks a day. Any transaction can be rolled back in the end of the day by the exchange, or the price/size adjusted. Most of the times, the retail user just doesn’t see it. There is 0 need for non-rollback transactions. Think as a Database, which businesses really want a append-only database? Its so much more work to get it right…
- legutierr 4y ago
  Very interesting. I didn't realize. Can just any transaction be rolled back, or are there restrictions? Do both parties need to agree to the rollback, or can a rollback be involuntarily imposed on one party by another party? I imagine there are some circumstances where after a disadvantageous trade one party might claim to have fat-fingered the order, just to try to avoid the repercussions of a mistake that was made with full consent and awareness. I imagine there are cases where if a counter-party were to agree to a such a reversal, it would represent a loss for that counter-party. How often are trades like this reversed, where one party clearly benefits from the reversal at the expense of the other party, and where the other party would not agree to the reversal if it were in their power to do so? How are these situations arbitrated, if only one of the two parties wants to reverse the trade, and the other party objects?

codefreeordie 4y ago

I can't find the article right now, but even wire transfers can be reversed under the right circumstances. Reversibility is actually necessary in ~everything under the right circumstances. There is a surprisingly high volume of wire transfer hold-harmless agreement traffic (this is banks agreeing to reverse wire transfers in exchange for transferring the liability if the transfer was not actually bad/fraudulent)

theptip 4y ago

> Does reversibility actually matter anywhere outside of consumer merchant payments? Yes. For example, the stock markets can reverse a day's trading when it's decreed that there was a flash crash. And in contrast to smart contracts, all business contracts are in some sense reversible in extremis, in that you can go to court and consider the agreed upon intent if there was a typo in the paper, and resolve ambiguities after the fact based on precedent, instead of contract bugs resulting in all of your money getting stolen. I think we probably agree in that reversibility isn't important in _most cases_ (the mainline happy path), but I think it's really important when you do need it. > extend reversibility to other transaction types just so as to deploy it as a strategy to prevent open blockchains from becoming predominant That's not quite the claim I was getting at. Perhaps if I'd said "tighter control, for example being able to implement reversibility and right to deletion” it would be more precise. The core of my claim is not that those two are the killer features; the feature that will cause decentralized options to lose in the end is more generally the control that a centralized entity has over the end-to-end experience. Those two are just prominent examples of the kind of feature that are difficult or impossible to implement on the blockchain, which are actually strongly desirable for the market makers. Another example would be KYC, which is hard to solve at the protocol level. My overall prognostication here was really that the existing big players will pick and choose the bits of the blockchain systems that are actually innovative and useful (I'd highlight smart-contract execution engines and true API-based transaction instructions for example) while dropping the "distributed" bit to maintain tighter control and enable broader consumer adoption. Analogously to how Google et. al. have transitioned the internet from a distributed/federated "everyone hosts their own server and communicates over protocols" world to a centralized "everyone uses hosted services for everything" world. And consumers seem to prefer the latter, if the FAANG valuations are anything to measure this by!

Adrox 4y ago

I work in the stock market, there is dozens of rollbacks a day. Any transaction can be rolled back in the end of the day by the exchange, or the price/size adjusted. Most of the times, the retail user just doesn’t see it. There is 0 need for non-rollback transactions. Think as a Database, which businesses really want a append-only database? Its so much more work to get it right…

legutierr 4y ago
Very interesting. I didn't realize. Can just any transaction be rolled back, or are there restrictions? Do both parties need to agree to the rollback, or can a rollback be involuntarily imposed on one party by another party? I imagine there are some circumstances where after a disadvantageous trade one party might claim to have fat-fingered the order, just to try to avoid the repercussions of a mistake that was made with full consent and awareness. I imagine there are cases where if a counter-party were to agree to a such a reversal, it would represent a loss for that counter-party. How often are trades like this reversed, where one party clearly benefits from the reversal at the expense of the other party, and where the other party would not agree to the reversal if it were in their power to do so? How are these situations arbitrated, if only one of the two parties wants to reverse the trade, and the other party objects?