[creeble]

This is what x.com (aka payPal) did in the earliest days. You could submit a credit transaction to a credit card without an initial charge.

It worked, but it was strictly against Visa/MC policy, and they had to stop. By that time they had some critical mass and got ACH going quickly.

I don't know if it's the same thing with debit cards, but I suspect it is -- the transaction goes through Visa/MC, but settles via ACH once it's known that it's a debit card.

But Square might be big enough to change the policy. Especially since Visa is an investor?

[mey]

The debit networks don't universally allow for refunds, even for legitimate transactions. Additionally it is generally against their terms of use.

Either Square is flying under the radar or has gotten special permission to do this.

[bmcleod]

We've been looking at the payments infrastructure in New Zealand recently and have discovered that, unless you find whoever wrote something 30 years ago, most of the senior payments people have limited and ofter contradictory understanding of the capabilities of the underlying systems.

[jes5199]

Definately the US banking systems also still have the feeling that there's been an accretion of layers of technology over ancient systems. APIs that look like XML files wrapping fixed-field-length VAX-style data. It made me wonder - are there still mainframes under there? Or mainframe software being run in emulation? Or are the systems just patchworks maintaining the 1970s interchange formats because no one can stand to making breaking changes?

[recuter]

Well... that is horrifying.