|
raddevus wrote: making passwords be limited to a certain length
Which almost invariably means they're not storing them properly. If they were hashing them, the stored value would always be the same length, so there'd be no need for any meaningful limit on the password length.
Even worse are the banks which ask for specific characters from your password. Again, they claim this is to increase your security by preventing key-loggers / shoulder-surfers from getting your whole password. The fact that it means they're storing your password in plain-text, or using reversible encryption, seems to pass them by. 
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Great post and you are absolutely correct. 
Richard Deeming wrote: Which almost invariably means they're not storing them properly
It is amazing how uninformed many of the sites and developers are about these issues.
It's scary. And, it comes as no surprise when Yahoo! has 50 million accounts hijacked.
They're one of the ones who limit password length. Ugh!
|
|
|
|
|
Obviously, if you want user to define a new password for a web site, he should not be able to reuse an existing password, so it really does make sense to prevent pasting password.
And this is even more obvious for a confirm password box. If the user mistype its first password, the confirm box help ensure that he has entered the same password twice reducing the chance of an error. Well, in that case, a site should probably disable all copy operations : copy, cut and paste.
This is not 100% full proof as it would fails if wrong keyboard is selected or if caps lock is active...
If it is a pain to type a password twice, then it would be a pain in the future to retype that password whenever you have to.
And for discouraging people to select insecure password, usually there is a minimal length (often 8 characters) and rules like having at least one digit, one characters, one uppercase character and a symbol...
Thus, if fact, I would that the problem is that you don't really understand security issues as otherwise, you would not complain about having to type a password twice...
Well, if you need to fill a form with many fields (like 10 fields or more) and the validation fails (say the site want phone numbers using 000-111-2222 format and you used (000) 111-2222 instead, or haven't filled a required field), then having to retype the password then begin to be somewhat painful...
Although it is possible to make improvements to make the site more user friendly, you don't always want to take much more time to develop a page (or multiple pages) for marginal benefit.
Philippe Mori
|
|
|
|
|
Philippe Mori wrote: Obviously, if you want user to define a new password for a web site, he should not be able to reuse an existing password, so it really does make sense to prevent pasting password.
It's not about the ability to re-use a password. People who do that (and there are many) probably do it from memory.
Paste is required because most of us use password generators these days so we have a nice, thoroughly random 20 character password each time we sign up to something.
So having generated a key along the lines of "Rx87Htv01pUWxb2WqkLLp" - to have to type it in twice (on a single screen machine, as it happened) was something of a PITA. To then find out that I'm expected to type it in manually each time I want to log in ...
Philippe Mori wrote: the problem is that you don't really understand security issues
Well, maybe I don't, but I do know that 8 characters is stupidly short for a password and that people who make up passwords rather than generate them are going to be a whole lot easier to hack than people who use Guids or lengthy random strings. "pa55w0rd" is not a very good password!
|
|
|
|
|
For those who trust password managers, then enabling Paste is a good compromise... User is still unable to copy a mistyped password in the first box...
When filling a form, often I mistype my password in one box so they mismatch so I really find that the idea is useful...
Philippe Mori
|
|
|
|
|
You'd only need to type it twice if the password-editbox is hiding what you are typing, which is hardly usefull if you are the only one in the room.
Philippe Mori wrote: Obviously, if you want user to define a new password for a web site, he should not be able to reuse an existing password, so it really does make sense to prevent pasting password.
Nonsense.
Bastard Programmer from Hell 
If you can't read my code, try converting it here[^]
|
|
|
|
|
Password are always hidden on Windows. Newer UI usually have a way to show password. Password being visible is only useful as long as you read what you type...
The problem is that often you thing you have written right so you won't even bother to read what you have wrote.
In my opinion UI like the iPad where one see the password while writing it (last character) make it a bit easier for someone to see your password that to see which letters you type...
Philippe Mori
|
|
|
|
|
Philippe Mori wrote: Password are always hidden on Windows. Newer UI usually have a way to show password. Password being visible is only useful as long as you read what you type...
Not "always", and there have been versions where you had the option to show or hide the password while typing.
Philippe Mori wrote: The problem is that often you thing you have written right so you won't even bother to read what you have wrote.
If the password is hidden then checking it for typo's is not possible. That is why the second textbox come to be.
Not because we assume that the user makes a typo in each entry; otherwise you'd have the same two textboxes for your accountname 
Bastard Programmer from Hell
If you can't read my code, try converting it here[^]
|
|
|
|
|
Most site use an email to identify the user so obviously, if you make a mistake, you won't receive the confirmation mail and it would create an orphan account...
Obviously, one should do much less typing error on its own name... and he might be able to update it afterward.
Even if it is possible to show the password, you would generally have 2 password box anyway. And you often have a confirmation for the email which is always shown.
Philippe Mori
|
|
|
|
|
Philippe Mori wrote: Most site use an email to identify the user so obviously, if you make a mistake, you won't receive the confirmation mail and it would create an orphan account...
..this started a bit before the wide-spread use of email.
Philippe Mori wrote: Obviously, one should do much less typing error on its own name...
You're right, that must have been the reason for the second textbox, silly me. It's not like people can be expected to jot down something important in a single time. So, my bank should ask me to insert amounts twice? And should ask each accountnumber twice?
You're making stuff up here.
Philippe Mori wrote: Even if it is possible to show the password, you would generally have 2 password box anyway.
If you can read the bloody password, then there's no need for a second textbox. It is merely there in case the characters are hidden, which has not always been the default.
Bastard Programmer from Hell
If you can't read my code, try converting it here[^]
|
|
|
|
|
Philippe Mori wrote: he should not be able to reuse an existing password
Let me guess - are you the guy behind the Password has already been used by another user message? 
Philippe Mori wrote: This is not 100% full proof
Neither is it fool-proof. 
(Clearly the spelling of the word fool-proof is not fool-proof.)
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Now I understand why the spelling corrector was not accepting fullproof... English is not my first language so I was thinking "full" like completly proof instead of idiot proof...
Philippe Mori
|
|
|
|
|
Not an answer... but this reminds me of when I once had to book some ferry tickets from an internet café. I happened to notice, that site was super-user-friendly. They had auto complete on the credit card number and the ccv :b.
... such stuff as dreams are made on
|
|
|
|
|
IMHO there is no logic for it since as you mentioned, it makes it a real PITA for those of us with password managers. Those developers should be flogged!
"Go forth into the source" - Neal Morse
|
|
|
|
|
Flogging's too good for 'em, but I like where you're coming from.
|
|
|
|
|
How a developer who does not use a password manager would have known of such issues.
I am a developer and I don't trust much passwords manager so I never used one... (except the one in Windows for network drives) or individual site "remember my password" on some sites.
I would never have though that a password manager would have rely on paste...
In fact, copy and paste a password has not been allowed in many cases for so long that I haven't tried to copy a password into the confirm box since many years... if I have ever tried it. And obviously, I would have never tried the effect of pasting a password on web site I have developed. It does whatever the browser does by default for password field and I am not even sure of what is the default.
Philippe Mori
|
|
|
|
|
The thing is, the default properties of a password field are already handled correctly by the browser...allow paste, but not copy/cut. I can't think of a single reason why pasting should not be allowed for a password field.
Philippe Mori wrote: I would never have though that a password manager would have rely on paste...
Well, that's the real beauty of it to me. I wrote my own password manager, like I'm sure a lot of others here have done. I haven't had to type a username or password in years for most of the websites being managed. My process goes like this:
0: start the password manager and login
1: click the desired website from a list
2: click a button to copy the username to the clipboard
3: launch the site and paste in username
4: click a button to copy the password to the clipboard
5: paste into the password field and login. Done!
I refuse to let any browser remember my login credentials for any website, though they continue to ask.
Everything's stored in a password protected sqlce database, which works great since I can share it between multiple computers, and even use it on my laptop when away from the office.
"Go forth into the source" - Neal Morse
|
|
|
|
|
PeejayAdams wrote: a dev team making some really, really bad UX decisions?
As a developer who does UX, I can tell you it's simple. We're all f***ers.
Software Zen: delete this;
|
|
|
|
|
It's similar to how they set passwords to expire every 60 days forcing people to write down passwords and stick it on their monitors. Security through wrongly assumed obscurity.
|
|
|
|
|
"I don't use a password manager, so no-one needs one!"
Don't tell me you haven't worked with that guy.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
I once made a typo in my password that then allowed me to copy and paste the erred password into the confirm box. I had to beg IT to reset it for me.
|
|
|
|
|
I can tell you what it is: Internal or external security audit has roasted the dev team and they had to make it "more secure" while making it less user friendly at the same time. Happend to us!
|
|
|
|
|
This is the conundrum faced when personal responsibility is rejected in favor of having someone else handhold us though processes constantly.
Well meaning coders attempt to prevent someone from copying a incorrectly entered password in the first field into the verification field with these sort of measures. Why? To protect us from ourselves! If we were, to stupid to do such a thing then I guess we'd deserve not knowing what we entered for the password and having to reset it later right (at least that is the way I feel about it).
I agree, most users a going to fall on either side of that scenario where they might copy the bad password into the verification field. Instead the more novice user will actually type both fields content manually; where as the more advanced user will be working from a password generator or create a complex password, copy it to a safe then to the verification field.
It is one of the massive mistakes of our world to think that we can code correct the human flaw. We cannot. We can only provide for a means for them to resolve their error after the fact with a reset. To do anything else only frustrates the bulk of the user base.
|
|
|
|
|
What's in the clipboard? JavaScript / Flash can access the clipboard.
|
|
|
|
|
I would assume it's just about making sure the "confirm password" box does its job. Sometimes the clipboard isn't reliable (think screen sharing tools, this bites me all the time when a coworker and I are both looking at the same customer server). Sometimes you may think you hit ctrl-c but you really didn't for whatever reason, and now your password is whatever was sitting in your clipboard. Since they probably hide the password field you won't know what happened and your first interaction with their site will be the password reset page.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
