On 10/24/12 11:37 AM, Adam Cameron wrote:
> Yeah, Windows does that with a WLAN key, and cellphones often do this with
> password fields *as you're typing them in*.
>
> That said, it should only be possible for values that originate from the
> current user typing into the field. You definitely should *not *be pulling
> actual passwords back from [anywhere], and then give people the option to
> see them. Indeed the password should never exist in clear text, it should
> be hashed, and the hash saved.
Great in theory, not really doable in "real life" when you're a "glue"
type of application framework.
> There should be no way to retrieve an already-entered password. It should
> be possible to reset a password without knowing what the old password was,
> perhaps, but even that should be something that's very secure, and not
> accessible in the normal course of events.
>
You're forgetting that we are middle men to some extent, with databases,
APIs, etc. so on and so forth. We store things encrypted, and security
is an onion, etc., and if you have concerns -- now this is freaking
awesome -- you can audit the code yourself! Or even hire someone to
audit it! Yay open source! :) (seriously tho- this is *huge* if you
are all 'bout security)
In general I agree with you 100%. One-way is best.
And look how freaking awesome Jordan is-- this particular issue was
already addressed! I remember him saying "I want to address this" with
a prior release... and then he did! Yay Jordan! I just wish he
wouldn't go all Hulk stylez on folks all the time. ;)
:Denny
--
Railo Technologies:
getrailo.com Professional Open Source
Skype: valliantster
(505)510.1336 de...@getrailo.com
GnuPG-FP: DDEB 16E1 EF43 DCFD 0AEE 5CD0 964B B7B0 1C22 CB62