It is a good idea to pay someone for escrow services.

This is what sounds to me the reason smart contracts exist. But maybe that's just a different method for accomplishing the same thing?

Anybody have any idea what the main distinction [if any] would be here?

Well, smart contracts can't tell if the seller sent you your plushie.

Oh yeah, I understand that limitation. But if that's those are the terms you want to use: the receiver must register that they received their item then a smart contract just automates the processing of the agreement/escrow payments?

I mean, less action is required in order to respond, and less work required in generating the agreement? Otherwise a new program, or a program built on a framework (eg. how smart contracts theoretically work) utilizing bitcoin would be required each time a sale agreement is made.

I'm probably confounding this, so I'll try again: a smart contract would offer a format to build upon for such a market and escrow agreements with less work than writing something from scratch?

I feel like I'm missing something, but it's late. I've been thinking about this concept for a little while and can't find holes besides what you've described, or engineering errors like the Parity exploit... but it still feels like I'm blind to something.

Why not consult UPS tracking as an oracle for that case?

Because UPS tracking is limited to determining whether UPS delivered a package. It can't say who the package was sent from, who it was delivered to, or what was inside.

This is what Dr Sanchez on our team calls the "box of rocks" problem. You _could_ potentially use a lack of delivery to automatically refund the buyer or something but automatic payout to the vendor is harder.

yes but this is also a problem in the real world right? I could order anything, sign the package once it arrived, and then complain that the box was empty. With or without a smart contract, it is hard/impossible to prove I'm lying or not.

So I assume this case is mostly taken care of by insurance, not the platform fee? Or I guess that's what the platform fee is for?

IIRC OpenBazaar still has paid escrow- it's just cheaper than eBay's fees.

Only in the case of a disputed between the buyer and seller where arbitrage is needed.

Good point! By the way the term is arbitration- arbitrage is a different thing

arbitration

The escrow fees are only charged if there is a dispute too.

> The bitcoin fees would still exist but presumably would be lower than 2.9%

Potentially, but Bitcoin mining fees are currently sky high. I made a couple of purchases recently for about $60 worth of Bitcoin, and the mining fees ended up being about 3.5%. Of course, mining fees are pretty much unrelated to the size of the transfer, so for high dollar amounts the fee is negligible, but ends up being a huge amount for small $5-10-$20 purchases. Hopefully mining fees will go way down after "the great Bitcoin split" in a week or so.

That was the case when the network was clogged, which should be much better now. For a 250-byte (avg) transaction you should be fine with 0.0006 BTC ($1.68, 2.8% of $60).

https://bitcoinfees.21.co/

You're right, bitcoin became expensive for small purchases, which is what the whole recent debacle was about. Hopefully Segwit2X fixes this, at least for a while.

Do you know when this UASF stuff is happening, or if it has already happened, who won, or what looks like it's going to win? It's been a confusing few days...

> It's been a confusing few days...

It certainly has. Not sure what your knowledge is, but the main discussion re scalability in bitcoin recently has been about the activation of segregated witness (i.e. segwit)

There's still one hurdle to get over over the next few days called UESF (User Enforced Soft Fork) that miners have SAID they agree to. The method for this is a thing called BIP91, and it will orphan off blocks from other miners that don't signal they are ready for segwit. Segwit itself (BIP141) requires 95% of miners over the space of ~2000 blocks to signal segwit activation. BIP91 only builds on blocks that signal BIP141, which is how they get to 95%. Almost 100% of miners have already indicated that they will do this. It should start in a few hours?

The problem right now is that lots of people have said they're going to do something, but the question is whether the nodes that define and police consensus in bitcoin are capable of enforcing the BIP91 orphaning. It SHOULD go off without a hitch, but given bitcoin over the last six months, there are no guarantees. If all goes well, bip91 will start orphaning non bip141 blocks, the new ~2000 period of BIP141 activation testing will start, >95% of blocks will be BIP141, and segwit will activate in about three weeks. That will essentially allow bitcoin to double the transaction throughput.

2nd layer solutions and hard-forks sheduled for three months (that are never going to happen) are a discussion for another day.

I see, thank you! Pretty confusing goings-on, but your explanation makes them clear, thanks!

Now that BIP91 (SegWit2X/NYA) is locked in, we will be getting BIP141 (original SegWit) activated in a few weeks, so the BIP148 (UASF) fork should not happen.

Actually there are still transaction fees in the bitcoin network.

Yes, that's why I wrote "(The bitcoin fees would still exist...)"

Are you talking about something else?

which goes to the miners, yes?

Yes