That’s why I’m asking for William George’s opinion.
Because at some point, it either becomes insecure or confusing. I don’t see why we would ever want to allow more than two mistakes. I mean, if you have three or more mistakes, then you either haven’t made any effort, in which case you kind of deserve losing your deposit, or you’re dyslexic, in which case you should not be writing down the address yourself.
I totally agree, but unfortunately, there is no lack of empirical data showing that often, vouchers don’t care to check.
So I reiterate my point that implementing this proposal does no harm. And since you’re making a slippery slope argument, you’ll need to justify why you think people would want to continually weaken the standard. Since I would expect the number of challenges for address errors to go down exponentially with each allowed errors, I don’t think the community will have much appetite for going further than allowing 2 errors. Unless I messed up the math again (I edited my first post; the security is actually better than I thought), allowing for 2 errors would weaken the security of the address by at most 19 bits, so down to 141 bits, and once again there’s no way an attack on that would make any economical sense in the context of PoH even if it were possible. To put things into perspective, based on the current hash rate (128EHash/s), all existing bitcoin miners put together running for 100 years wouldn’t even be able to crack 90 bits of SHA-256, and computing a hash is a simpler operation than deriving a public key and then computing its hash, which is what an attacker would need to do here.