I have some experience here. If you are seriously suggesting these
measures, you might as well kill retail transactions altogether.

In practice, if a retail place starts to accept bitcoin they have a
similar situation as with cash, only that the fraud potential is much
lower. (e.g. 100-dollar bill for a sandwich might turn out fake later)
and the fraud frequency is also much lower.

0-conf concerns were never a problem in practice. except for 2-way atms
i have never heard of a problem that was caused by double spends.
while adding these measures is generally positive, requiring them means
excluding 99.9% of the potential users. so you might as well not do it.

RBF as implemented by F2Pool just flat out lowers Bitcoins utility
value. So it's a bad thing.

for any online or automated system, waiting for a handful of
confirmations was always recommended practice.

Am 19.06.2015 um 22:39 schrieb Matt Whitlock:
> Retail POS merchants probably should not be accepting vanilla Bitcoin
> payments, as Bitcoin alone does not (and cannot) guarantee the
> irreversibility of a transaction until it has been buried several
> blocks deep in the chain. Retail merchants should be requiring a
> co-signature from a mutually trusted co-signer that vows never to sign
> a double-spend.