Quote from: makomk on March 13, 2013, 12:27:40 AM

Quote from: Micon on March 12, 2013, 08:05:47 PM

Excuse my non-technical, conceptual-only questions / thoughts / checking if I have this right in my head:

1) There was no *permanent* "double spend" right? i.e. the transaction on the .8 fork would have eventually been orphaned / dumped / never credited to the address specified in the .8 block?

2) The risk here (and reason why SwC shut off all deposits / withdraws for a few hrs until this was resolved) was that during this very odd period a sophisticated user could use multiple 0-conf services that would all at the same time believe they were getting the same coins, but after the blockchain becomes 1 again only 1 of the transactions would actually be included in the "real" blockchain?

In theory this didn't just affect 0-confirmation services. The transaction on the 0.8 side of the double-spend got 15 confirmations on that fork before it was invalidated by the conflicting transaction from the 0.7 side of the fork, which is more confirmations than anyone requires.

So I think the 6-confirmation recommendation should be changed: tell the merchants to wait for as many as possible if time permits, or if the amount is large.