Saturday, December 9. 2006Password shadowing: Pimp My PlastiqueTrackbacks
Trackback specific URI for this entry
No Trackbacks
Comments
Display comments as
(Linear | Threaded)
Hi!
I have some bad experience with it, when using the polyester 1.0 style: please see my postings here: http://forums.gentoo.org/viewtopic-t-525707.html
I would be nice if you implement the behaviour which can be found in Lotus Notes client.
When you enter a password in Lotus Notes client, the program shows a constant number of [weird] and easy to recognize pictograms instead of bullets. Each typed character changes the whole set of pictograms to something new, and the exact set depends on previously typed characters. So when you type the entire password you may say confidently if the picture is "usual" and thus you haven't mistyped. Since the number of pictograms is always constant and they changed without any evident rule, it's quite hard for an evil one to guess how may characters in your password, what they are etc.
In theory, the pictograms leak information by changing on every keypress. If you have 20 or even just 5 "checksum" images, you can use brute force on a much smaller set if you have, say, a video camera recording each change. If you only show the checksum image after the first 5 characters, it still effectively reduces your password to 5 characters plus a bit of brute force. You could stop the brute forcing by using a random seed unique to the system or network so you can't reproduce it at home.
Of course, in real life, you could probably just as easily watch their fingers on the keyboard, or one of many low-tech methods, so the issue is pretty much moot, especially on something like a desktop computer |
QuicksearchMy Books on Qt 4Calendar
ArchivesKategorienStatisticsLetzter Artikel: 2009-07-03 11:16
118 Artikel wurden geschrieben
227 Kommentare wurden abgegeben
832 visitor(s) this month
PLUGIN_EVENT_STATISTICS_TEXT_DAYVISITORS
5 Besucher online
Blog AdministrationShow tagged entries |