Andreas Wachowski (7) [Avatar] Offline
#1
Not a priority, just a suggestion:

Table 2.2 is meant to show that "changing the input just a little bit" always produces a (very) different hash. However it always uses the same string pattern Hello<n>!, with n being an integer (except for the refreshing outliers Grumpy Cat and your entire music collection smilie). I would suggest to use, in addition, various other classes of changes. It feels to me that the higher the integer (and hence the longer the string), the less similarity the two strings have, so the original premise doesn't hold anymore.

Further examples might be:

Adding a character: Hello1!
Removing a character: Helo1!
Exchanging two characters: Hello!1
Visually similar characters: Hel1o!
Uppercase/lowercase variation: HEllo1!
"Special" characters: Hello1ยก
Kalle Rosenbaum (10) [Avatar] Offline
#2
Hi Andreas!

Thanks for your suggestion. Yes, the table might benefit from your suggestion. However, I still want to make it clear that we make >10000 tries while not making the table too big. Maybe just one or two other classes of changes?

On the other hand, explicitly listing all 10000 tries + extra classes, would quickly help us reach our target of 250 pages smilie

/Kalle
Andreas Wachowski (7) [Avatar] Offline
#3
Yes, I agree it makes sense to keep the long sequence in. So as you said, two more such concrete examples at the start before starting the sequence might convey better that the idea is not to literally take the numbers 1 to 10000 (there's always https://github.com/minimaxir/big-list-of-naughty-strings if readers want to fill in the blanks smilie).

Kalle Rosenbaum (10) [Avatar] Offline
#4
Thanks

Ok. BTW, the target is not 250 pages, but 350 pages.

/Kalle