[Cryptography] 9999 keys for this one iPhone

[Henry Baker]

At 12:07 PM 2/29/2016, Tom Mitchell wrote:
>0, 1, 2, 3, ....
>Programmers always confuse things ;-)
>Jokes aside you are correct I had assumed that 0000 was
>illegal based on a one time experience.  I tested it and 0000
>is a legal PIN.  If if 9999 fail then the 10000th is the
>answer and need not be tested.

(Fortran's 1-origin indexing set back software development by a full generation, and easily cost > $1 billion in SW errors.

Of course, why *math* people were using 1-origin indexing to begin with leaves me scratching my head.)

Why are you so sure that after you tested 0001-9999, then suddenly 0000 will open the iPhone?

Have you tested this theory?  Has Apple?

Your faith in SW development is misplaced.  I wish I had saved all the printouts from large companies with completely empty pages due to off-by-one errors; it would make a nice museum.

I doubt that Apple ever tested this particular 0001-9999, 0000 sequence, given the amount of time & work required.

Testing corner cases like -- I don't know -- ISIS-inspired murderer leaves iPhone that the FBI wants to open, e.g. -- is a little difficult to arrange, a priori; regression testing these cases for future SW updates produces a lot of orphans.

I would imagine that thorough testing for the next generation of "unhackable 2.0" iPhones will produce a huge number of dead iPhones; Apple may need to have a whole refurbishment line to desolder the CPU/enclave chips and replace them.