ichimunki wrote:

So maybe I don't have the advanced math skillz to comprehend this, but when I produce an MD5 hash from an input (salt or no) I don't get a string such as the one you indicate.

Using Digest::MD5 will not generate a string like that. See the link that mdillon referred to for a better explanation of how that string is created.

ichimunki also wrote:

Finally, remember this all comes down to the passwords. Are they computationally inconvenient? If not, I might simply use LWP to keep submitting until I find a match. Your CGI should prevent weak passwords. And as part of defense in depth I would (as I said) limit the number of invalid tries.

Without going too in-depth into our password policies, let me just say two things:

1. Over the objection of the users, we have implemented a moderately strong password policy. It's not as strong as I would like it, but no one is going to get away with using 'password' for a password.
2. You get seven tries to get your username/password combination correct. Each failure is logged. If you blow it seven times, you are locked out and the company needs to call us to unlock it. They are not even allowed to unlock failures. Of course, we investigate lockouts before unlocking.

Cheers,
Ovid

Join the Perlmonks Setiathome Group or just click on the the link and check out our stats.