A Guide to Using Passwords Without Distraction
filterjoe.com
filterjoe.com
Password managers are great for some folks. Especially when paired with a USB security dongle or other secondary authentication. One of the author's other posts has some good comparison: http://www.filterjoe.com/2010/05/14/which-password-manager/
But overall, I've been using a 'passphrase system' and it has worked conveniently. Creating a passphrase could be formulaic like: MyPassphrase4[gmail.com]!
Easy to remember (generally, just look at the URL you're logging into and plug into your formula) and rather secure by conventional standards.
http://www.filterjoe.com/2010/12/06/a-base-phrase-approach-t...
However, I suspect that very few people are willing and/or able to do this. Look at what people actually do, and how it fails: http://www.filterjoe.com/2010/05/14/the-usual-way-to-manage-.... About half the population uses the same 2 or 3 passwords on all sites. A third use the same password on every site. This is according to a Sophos study: (http://www.thetechherald.com/article.php/200911/3184/Interne...)
So at least 80% of the population has terrible password security. Most of them will never change unless they are offered a very simple and convenient alternative.
Password Managers like lastpass, 1Password, RoboForm, and KeePass are currently the easiest way to manage passwords in a relatively secure manner.
As you say, there is the theoretical single point of failure. But in all of my research, I have not been able to find a single instance of a password database (from one of these 4 market share leaders) being compromised when protected by a strong master password. On the other hand, password theft occurs millions of times per year. As you likely know, those who reuse passwords often then get more serious accounts hacked when a password for an inconsequential account gets captured.
I would love to see the tech community rally around this simple solution for the masses, just like they did for virus checkers on Windows systems a decade ago.