This is why all these accounts get an email account that doesn't include my real name, too. There are so many sites that have these signup dialogs that require an email address -- why? Oh yeah, that's right, user data is the oil that will fuel the future economy.
- Password Recovery (this can be optional though, for my sites it usually is)
- 'Legit Users', sending an email and having them confirmed through a code in them gives a bit more confidence in the user
There isn't a day that goes by that I don't get an email intended for someone else, often including personal information, due to a mistyped email address. Whoever decided that email verification was a poor user experience needs to be hit in the head with a shovel after he digs the appropriate sized hole.
If you know Catherin (PA) let her know her round trip to vegas is confirmed
Carolyn's (NYC) open table reservation was canceled by the restaurant.
Different Carolyn (CA) needs to correct email address with the walnut creak school district.
Possible the same Catherine (PA), your financial advisor needs a distribution form filled out.
Clint (OH), your repair at Kay jewelers in Akron is done.
Ann (FL), Thank you for scheduling an online appointment with Conley Subaru
sigh, you get the idea
Edit: Oh, and I'm going to start canceling Open Table reservations that show up in my inbox.
What I really wish for is a link in emails that say "I am not the intended recipient of this letter." Normal mail works like that. You can ask the United States Postal Service to only deliver mail that has your name on it, and if you send back mail with someone else's name they will add that person's name to a database that says the mail is undeliverable.
I don't want to mark these organizations as spam, but that's basically what they are to me. I've started using the password reset to log in and change the email preferences so I don't get these emails anymore.
I even tried calling Apple to get them to cancel the account, but they wouldn't let me because I couldn't answer any of the security questions on the account!
The guy on the phone acknowledged that of course I can't answer them, because I didn't create them, bit their policy forbids them from doing anything without the security questions...
Quick Edit: I just read the Wikipedia page, and individuals cannot bring suit against spammers (using CAN-SPAM in any case). Which is a shame really, because spam is really really annoying.
I found this out the hard way when someone else had created an account under my email address and because some of my friends had uploaded their address book, my friends were following this stranger, perhaps thinking it was me. When I eventually did choose to sign up, I had to jump through a bunch of hoops just to be able to use my own email address.
Plus, this individual who keeps signing me up lives in Australia. I cannot imagine our system is so dysfunctional that both the United States and Australia would agree that a) this is a criminal offense, and that b) it would be worth going through the effort of extraditing me to face justice.
I mean, I haven't even received another password reset request from those websites I logged in to fix that issue, so I half imagine that person is using my email address as a plausible address for email they don't care about.
I understand what you mean, but I don't think I'm going to have to spend the rest of my life looking over my shoulder. The few personal emails I've gotten with financial details (gym membership bills, etc) I've responded letting them know about the mix-up and they've been more than happy to fix the issue. Feel free to say I told you so when I end up in a jail full of kangaroos though.
That's a solid argument for business cards with vCard QR codes as I see it. Or a pack of stickers with QR codes that you could slap on whatever documents you needed.
It's too bad QR code adoption is so horrible. I'm just as guilty of it as anyone else, I assume that any QR code is probably just some marketing thing.
But vCards seem like one of the very legit use-cases. Now that I am imagining using stickers and never filling out my contact info on a form again, I realize that's definitely a world I want to live in.
I think it's more likely if you use an initial in your email address. I have a fairly unusual last name, but I used my first initial to create my gmail address, so I regularly get stuff meant for a couple of other people.
Business cards or sheets of stickers are not going to solve this problem. The former is not universal, and you're not going to get enough adoption for the latter (do you really want to carry them around in case you need to put your email on a form?).
The only solution that might actually work is to put some kind of financial penalty on certain classes of email sender to incentivize reasonable levels of validation.
Yes. Or my name/address/phone number. It's essentially a machine-readable address label, the email is just one part of the vCard.
Do you people really not keep a book of stamps in your car? Such inconvenience!
> Do you people really not keep a book of stamps in your car?
No, why would I do that?
That's why you get send the verification email first, when creating the application user, and don't do anything until the email is verified.
The flow should be:
User: Please give me an account; my email address is [email protected]
Server: I have sent an email to [email protected]
Server → [email protected]: Please go to https://server.invalid/register?token=KDU6dG9rZW4xOWppbUBleG...
User: visits URL, completes registration
"Hi X. The £80 million for the new biomedical building is approved, please make the transfer ASAP."
It turned out I had the same unusual name as the most senior financial officer.
My last name is pretty uncommon, but there are at least a thousand of us or so in the US.
I probably get a dozen emails a month to people who are not me (not counting all the ones that end up in the spam folder after I'm on a damn list).
Susan. Stephen. Another Sam. So many damn S. Lastnames around.
On occasion I've tracked the people down. I've forwarded their emails if they seem important. And whenever I ask that they be more careful and memorize their own email address, they get defensive and tell me someone else made a mistake.
Buddy, this email was was an automated response from a computer after you signed up for something. No, it did not just randomly decide to send to my @gmail instead of your actual @yahoo address. YOU just don't know your own email.
If we didn't have a monoculture in email providing, this would be far less of an issue.
Not to the extent of the above, but for instance, last year I received someone's job offer in Colorado. I did respond to the sender to let them know of the mistake of course.
My most entertaining email yet was a group thread that kept track of a young missionary's progress as he traveled through South America. He's home now, so I don't get weekly updates anymore :( I almost miss it.
Seems like the simplest solution is to have some basic password requirements.
> Verifiers SHOULD NOT impose other composition rules (e.g., mixtures of different character types) on memorized secrets.
> When processing requests to establish and change memorized secrets, verifiers SHALL compare the prospective secrets against a list that contains values known to be commonly-used, expected, or compromised. For example, the list MAY include (but is not limited to):
> * Passwords obtained from previous breach corpuses.
> * Dictionary words.
> * Repetitive or sequential characters (e.g. ‘aaaaaa’, ‘1234abcd’).
> * Context specific words, such as the name of the service, the username, and derivatives thereof.
I suppose you could open a pull request :) https://github.com/usnistgov/800-63-3
Also, the big question that's missing from the article is: who's the enemy? If the enemy are Russian genius hackers, then certainly very long passwords and maybe other measures are in order.
But the enemy is not always remote. I just bought a new iPad and had to reset my password and so-called "security questions".
In my experience, there are only two kinds of security questions: ones I don't know the answer to myself, and ones everyone around me knows the answer to (what was your first job, what was your surname when you were a child, etc.)
The problem is, in the case of the iPad, your enemy is not some James Bond villain in a cave in Ukraine, it's your 13 year old child who wants to use your account to buy apps, or your spouse during a nasty separation. And they both know, or can guess, answers to "security questions".
What I do personally is type random things as the answers, and write them down somewhere. But that doesn't improve security, it weakens it.
Totally agree about the "security questions". They're just insecurity questions. Everyone knows what street I used to live on and what my previous phone numbers were.
"What is your mother's maiden name?" - "gHk899iL@"
So-called "security questions" are the dumbest security measure.
My solution is just to keep those answers in the "extra notes" section of my password manager.
Oh, and they referred to these security questions as "2FA."
https://www.united.com/ual/en/us/account/enroll/default
Derp.
I love the ones that ask about one's favourite sports team (not everyone likes sports), or first girl (or boy) friend (not everyone has dated) or first car (not everyone has ever owned a car).
It weakens it to the extent of your physical security. But in practice how many crackers are breaking in to your flat [aka condo] to find your written down security responses, unless you're famous - or a target some other way (politician?) - in which case you should have sufficient physical security to protect the written answers.
Having security questions allows users to use hard passwords without the problem of losing an account. It increases the risk of personal attacks, for sure, but I'd warrant the chance of personal attack [ie by someone who knows you (or has gone through your bins)] is minimal for most people vs. the chance of an automated online attack.
I think it's right to minimise the online attack surface at the risk of the offline one for most users.
Regretfully Apple has an annoying bug that the dialogboxes accepts 5-word phrases, but only remember the first 32characters, 3 words or something alike. Very stressfull when you’re trying to recover a password.
If you are willing to be brave about this then the best way of answering security questions is by typing a random string and NOT writing it down. You will never need that value again.
Today, years later, I still can't login to the damn payroll site because there isn't a procedure for just completely resetting the password state of the account and it wouldn't let you log in without those extra things.
Since that point I have recorded every annoying split-password (that's what the 'security questions' actually are, a forced split in the single real password)...
You do. A lot of users don't consider automation when it comes to people hacking their account. I've heard "Nobody will ever guess it though" a few times during my career.
What attack vector is a password with high entropy protecting against? The only one I can think of is an unreported database leak. The attacker may be able to more easily reverse the password hash and use your account on that particular service, at least until the leak becomes discovered and the service does a full password reset.
While it would be ideal to defend against that attack, I think it's far more important to stress that users have a different password for every website. The real attack vector that we see every day is a DB leak from a poorly secured website, passwords cracked, and then reused on important services. That's where the problem lies, not in the user's password entropy.
Jeff isn't wrong here, he's just attacking a relatively unimportant problem.
Very many people. And not all systems stop them from doing so.
And database leaks happen all the time too...
I'm no security expert, but I start to wonder if the "practical" entropy of a given password is actually much higher than the theoretical entropy numbers people derive.
Serious question: when is the last time anyone has had an account compromised because of a weak password?
Come to think of it, I also had an old account with a very weak password that I hadn't used in years mysteriously hacked one day. The language settings were changed to Russian when I went to reset my password. The fact that no other account of mine appeared to be compromised and I hadn't used the site in years certainly suggests brute forcing.
What is your definition of "weak"? I sometimes help my landlord fix their computer when they accumulate too much malware. One time I brought their tower home to do this and didn't get their Windows account password. Instead of calling them I decided to try a few things first, including their kids' names. Lo and behold, the second attempt was correct.
Brute forcing a weak password probably isn't what you have to worry about when it comes to targeted attacks against weak passwords, but social engineering is.
Weren't the celebrity iCloud account compromises due to Apple neglecting to rate-limit a certain authentication API endpoint, allowing the hacker to brute force passwords online?
Also, doesn't some of the current crop of IoT malware brute force device passwords as well (after trying the defaults)?
root ssh:notty 116.31.116.44 Fri Mar 10 22:27 - 22:27 (00:00)
root ssh:notty 116.31.116.44 Fri Mar 10 22:27 - 22:27 (00:00)
...
My personal server has 153,246 entries from this single IP, from 1 March until now.Login using the root account is disabled.
Do you run any ssh or email servers? If so, your logs will tell you already. I've got bots brute forcing passwords on my blogs, that nobody even read, but did have account functionality, with email sending. People made the point of adapting their bots for different captcha schemes just so they could send some email.
There's a difference between random script kiddies and targeted attacks. People do all sorts of wacky things.
Surely that's a cop-out? If the user getting their data stolen is not a big deal for you, why bother securing any system at all?
The reason service providers set password rules is because there are many, many people who do not understand what a good password is, or the risks associated with losing them. There are people out there using the same password for everything, unaware that the compromise of one service can spell doom for their financial life.
Developers really need to start taking responsibility for their code. Setting password rules, while arbitrary, are about protecting users. This may upset those who are tech-savvy and know what they are doing, but that really weighs nothing compared to the benefits it brings to those who are not.
It is weird that he goes on to suggest a complex system of rules literally right after saying rules are bad, and that's the point where he lost me too.
If users don't understand what a good password is, they are going to eventually get an account compromised somewhere, if not on your site, then on another. In cases where there's no liability back to you, then there's no sense in babying them. If you do have liability, then two-factor authentication is probably a better way to go.
I personally would hate it if a company forced me to use their generated passwords because I like to manage my own passwords.
However I take offense at the notion of "babying" users. We are all babied in some aspect of our lives: if it weren't for the active work and intervention of those in other fields, we wouldn't know whether the food we eat is safe, or if the roof isn't going to collapse during the night. I'd like developers to assume some responsibility for their users in turn.
But I concede that may be beside the point in the case of password rules.
It's not an illusion, 1q2w3e4r5t is indeed better than aaaaaaaaa, even if it's just the numbers interleaved with qwerty (and probably easy to brute force generate up to it).
>if I use "aaaaaaaaa" as a password on a website I know full well what I'm doing. So why even bother?
Because not everyone who uses "aaaaaaaaa" (or "1q2w3e4r5t" for that matter) knows "full well" that it's insecure?
It's a password that I postulate is more secure that the average and yet is only made of two different symbols and would've been rejected by the proposed "x amongst y" character policy.
It's a bit far fetched but not that much, you'll notice that the pattern is simply the first digits of pi, so it's fairly easy to remember. And the letters are "Hn", like hacker news.
That's my point, really. If you think you know better than the user how to pick a password, then just do it. Otherwise don't get in my way, you don't know how I generate my passwords.
Only they don't need to know how YOU generate your passwords. Just how most of their users generate their passwords. The suggestions and rules are not there to protect computer scientists from entering bad passwords...
I feel like the solution to everything in this thread is just to use zxcvbn and stop with the insane rules for things. In your two cases: the bank would disallow passwords below some limit while the blog would just show you a warning (in case you were ignorant of hacking enough to know that "aaaaaaaa" wasn't a good password), but let you use your awful password to spare you from having to remember it.
"COMMON PASSWORD: IN THE TOP 9635 MOST USED PASSWORDS
Your password is very commonly used. It would be cracked almost instantly[1]"
I do realise there's a difference between cracking a local system password vs a website password but always assume a site's database is going to be leaked at some point
[1] https://howsecureismypassword.net/ (Don't go putting real passwords in there..)
47 TREDECILLION YEARS to crack your password
Why not create even stronger passwords with Dashlane? It's free!
I didn't know that was a number!
In this case, we're talking 'better' but still bullshit [1]. Calling it 'better' makes no meaningful distinction whatsoever in practice.
[1] https://dl.dropboxusercontent.com/u/209/zxcvbn/test/index.ht...
We need to make passwords a much simpler thing to manage so that people don't create a bigger problem by putting it on post-its on their monitor. What I'm suggesting would take some effort but is where we need to go. Your username is your fingerprint. You won't forget that and it is very hard to fake/brute force. (Your phone already has one and a usb one for you computer would be trivial. A lot already have them.) Your password is always just 8 characters, alpha numeric only (62^8 is plenty big) and you must have 2FA on. It would be trivial for the phone manufacturer to add that at the OS level and it could be visible on your lock screen. Also, finally a good use for a smart watch. The rules are simple and you have removed the biggest security hole most companies have: people. Just my two cents....
So fingerprint gives you part 2 and password gives you part 1.
No need for an additional MFA device.
[1] http://www.barclays.co.uk/Helpsupport/UpgradetoPINsentry/P12...
[FWIW I expect the reverse situation is probably the same, this is just my anecdotal experience].
Yes! Please. Let users determine the value of their accounts for themselves. If I decide that say, my HN account isn't valuable and use a common password that is shared with many websites, I know the risk I'm taking and I'd rather take it than have to mark down a unique password for HN.
On sites I write I advocate for no password rules at all (hit enter is good enough for me, yes we can hash an empty string), have usually compromised on a simple length check.
Proper password storage is important, use the latest bcrypt/scrypt/argon2, don't cheap out there. You can cheap out on password rules. If you want to write a fancy entropy checker, run it client side and tell the user how secure their password is, let them choose whether they want that or not.
"So I send it to my mail account for when I am on the go... that account with the qwerty password..."
The point being, you can't really protect users from themselves so a bank can't build it's security model upon the assumption of safe user accounts.
I agree. Sometimes I'm just signing up to a site where I don't put any personal info up, or at least nothing more sensitive than my name. Why do I need high-level security there?
That's you. The average user thinks 'nobody will guess this password'. In other words don't assume they know about brute force attacks, rainbow tables, most common password lists and a host of other things that they haven't thought of or don't know even exist. They don't even know the complete group of actors that could potentially breach the account.
I've since had to change my random generator to avoid 3 of a kind, which actually makes it less random/secure i believe.. pruned some branches off of the search tree.. especially since i've now told everyone ;)
Nothing makes me immediately dismiss a site or piece of software faster than nannying me, especially from go.
There's no substitute for length. "dog" is a bad password. "doooooooooooooooog" is much better. The entropy is still the same, but the latter takes much longer to brute force and is just as easy to remember.
Of course one should not use passwords one can remember for anything critical, but many passwords are for services where compromise is not a big deal.
I don't want to trust a bank who knows my password.
> I had a bit of a sad when I realized that we were perfectly fine with users selecting a 10 character password that was literally "aaaaaaaaaa". In my opinion, the simplest way to do this is to ensure that there are at least (x) unique characters out of (y) total characters.
Isn't that exactly what you're complaining about with your arbitrary password restrictions to begin with?
I mean, I can imagine that a clueless user might have the illusion of safety if they're using something like "1q2w3e4r5t" but if I use "aaaaaaaaa" as a password on a website I know full well what I'm doing. So why even bother?
I think there are two possible ways to look at this problem from a service provider perspective:
- if the user getting their password stolen is a bad thing for you (i.e., you're a bank or something like that, and getting an account compromised will put you in trouble), then IMO the only satisfactory solution is to impose a password to the user. In effect these ridiculous password requirements are exactly that, except less convenient and secure. Cut to the chase and say "your bank password is Axei5aoc0i, write it down somewhere safe".
- if the user getting their password stolen is not a problem for you because it's not your responsibility to handle these issues (like a hacker news account for instance) then just let the user pick whatever they want and deal with the consequences. If they care enough about it they'll care enough to pick a decent password. At most if you really want to be friendly give an indication that a password might be weak, but please don't disallow it.