Slashdot Mirror
Using Images as Passwords
TekkenLaw writes "According to this news on Reuters, MS is looking at images rather than plain old text for enhancing security. The key - images, which tend to make more of an impression on people than strings of text characters. This is especially interesting in context of the crappy passwords story that ran on Slashdot that ran few days back."
So when you call support to get your lost password, will they ask you what
your mothers maiden hair color was?
did they not run this same story a couple weeks ago?
"you sonofabitch i didn't know!"
a friend of mine has a cool USB device that reads his thumb print, and he uses that to unlock his Windoze box.
"I would say that 99 per cent of what my father has written about his own life is false." - L. Ron Hubbard Jr.
Johnny Mnemonic
Whoa!
a string of characters as a password how am I going to remember exactly which points and which sequence of points/graphics to click???
I don't get it - call me flummoxed.
Users simply remember exactly where on the images they clicked and in what order.
If this really down to the pixel level as the story says, then this is not simple it is impossible.
Even having it sensitive to with 10 pixels say is going to be difficult with the pictures they used. Most country flags consist of large blocks of colour. To have a chance of reproducing a password, people are going to have to pick points near edges and corners - similar to not using uppercase and punctuation in passwords.
I remember this freeware app awile ago that would remember your passwords for you (it was'nt gator). Instead of you putting in a password it would show you a picture of a bedroom or somthing. Then to access your passwords you would click on series of objects in the room. It worked quite nicely.
Hacker Media
for pr0n site access ;-)
It can be seen here
AfterDark for Mac OS used to have a feature like this; you could select an image, and you would have to click on a certain part of it, optionally holding down a control-key combo, to unlock the screen saver, rather than type a password.
You have to click on a number of pictures in the right area in the right order. This is easier to remember than a bunch of keys in order (which is what happens after you've typed your password enough times)?
I'll stick to my text-based pass *phrases* while the Wal-Mart XP crowd sits and clicks on images like a 3 year old, thank you very much...
Can you guess which points a typical person would click on that image of a face? That's right - Eye, eye, nostril, mouth.
People don't select lousy passwords ONLY because they are lazy. They also select them because they don't think there is a credible threat to their accounts. They don't BELIEVE in hackers who would target them.
Without an increase in paranoia among average people, I don't see how a user-selected secret will ever provide security.
Novel idea, but I can see a lot of practical problems arising. For example, how do you determine how much room for error there is in clicking on certain parts of an image? Someone might choose to click on the sky, then a boat for their password. Will positions be based on something like +-5 pixels from where you originally clicked, or something smarter like using a magic-wand kind of algorithm? Also, what about people who are blind, or visually impaired? How will people sitting down at a computer figure this system out when they are presented with a picture? If you wish to share your password with someone remotely, how do you do it? (e.g. your mom forgets the password to the family computer and calls you up). Don't get me wrong, it's a novel idea, but I can see a lot of issues coming out of this.
slashdot!=valid HTML
Welcome to Microsoft Windows .NET 2005
In order to log in, please choose the One who you will truly worship, for He is the Supreme leader.
[ LINUS TORVALDS ] [ BILL GATES ] [ ROB MALDA ] [ LARRY WALL ]
Note: According to the EULA you agreed to unknowingly, choosing the wrong password could result in death and/or excommunication.
qslack.com
Check out this Slashdot story from last December, and the Real User site with "passfaces", which have been around for a long time.
"Thank you for participating in the required MS Passport sign-up verification to get your latest reinstall of XP2005 to work. We're sorry, but the image of a closed fist lifting the middle finger has already been taken. Others you may want to consider: You lifting your middle finger while wearing gloves; you lifting your middle finger while wearing a Cracker Jack ring..."
--------
Bleah! Heh heh heh... BLEAH BLEAH!!! Ha ha ha ha...
From the news story: "Even with such a system, people would still be susceptible to "shoulder surfing," in which someone watches a computer user type in their password."
Users would have to be fools to "click" their password unless they are positively alone in the room. The current standard at least has masked text on screen, and the order of keys on the keyboard is VERY difficult to track even when the user is moderately good at typing.
Let's not forget that in the case of the new photo passwords, with 50% of users you would only have to know the "Lenny Bruce sequence" in their Playboy passphotos: T'n'A
~zecg.
The first article can be seen here
Next up will be the "Tapping System" where folks will rap out "Haircut & A Shave" on their desk to log in.
What other quirks of human nature will next be put to use trying to identify folks? The "Mictation Flex Rate"? The "Eyebrow Lift/Tongue Roll"? How about the "Tell the Same Stupid Joke" one; I've had co-workers who've been able to do those hundreds of times over & over without a single variation.
Or just teach folks how to use good paswords, put in some really good acceptance tests, and make it clear that if security is compromised by their poor password choice they'll be held responsable, same as leaving the door to the safe open.
Nahhh, there's gotta be a technolgy fix...
I don't read ACs: If a post isn't worth so much as a nom de plume to its author then I wont bother either.
This is kindof interesting. A couple of things spring immediately to mind.
First, presentation of the image will (may) vary in different situations. The visual presentation of a password is pretty irrelevant: as long as you can understand and input the right symbols the font, colour size etc. in which they are presented isn't relevant. On the other hand an image must look substantially like the crib image. Sounds obvious, but consider differences in resolution, colour depth etc. You can divide the image into regions (a grid, perhaps) but ultimately there will be a limit to the resolution of the grid that you can rely on (not to mention input errors limiting the viable grid resolution.) To get more possible regions, you'd need a plain bigger image to get around the input resolution issue. All of which complicates the implementation (of course, you could break each image down semantically somehow, but that sounds like a further adventure altogether.)
And, after all that, prople may turn out to have pattern preferences that are "as crappy" as poorly chosen passwords? Always use a photo of your daughter and click on both eyes and outline her cute smile? Ooops. Use your country flag and click where regions of colour meet?
the BSD !
with Patrick Stewart in where each day he had to put these strange Chinese characters into a picture otherwise the government secrets he knew would be e-mailed to hundreds of newspapers. Can anyone remember what it was called? As for thumbprint technology the hardware is still very expensive. :o\
Video Game cheats, hints a
And i think that i'm not the only one, i'm not THAT much of a freak
Then the government can check to see where you like clicking pictures.
Did you use the Iraqi flag as your password?
Are you clicking on suggestive areas of that picutre of Natalie Portman?
I much prefer just having a city-wide network of surveillance cameras to verify my identity at all times.(/sarcasm)
Read Lostbrain's Oscar Predicitions!
tcd004
Did everyone forget about this already?
Peace, Love, Linux
Sounds interesting, though I'm not sure i see how much a difference this would make. What's the difference between remembering certian details in the image you selected as your password vs remembering a text password?? You *Still* have to remember something. I've been very fond of the fingerprint scanning system and other simular devices that allow you to access your data without having to *remember* anything.
:)
But in any case, where there's picture passwords there's bound to be some strange tech support calls. I'd use one of those "magic eye" pictures where you have to make your eyes blurry, and cross them funky so you can see the hidden image. That way if he's a real *STUPID* tech, he'll look like it too.
...anyway...
A penny for my thoughts? Here's my two cents. I got ripped off.
A Penny for my thoughts? Here's my two cents. I got ripped off!
Back in the day, before The Matrix, Keanu made another sci-fi show predicated on just this idea. It was called Johnny Mnemonic!
a keyboard. It would be easy to remember where to click, because I could remember it as a string of alphanumeric characters. I think this technology has promise.
Evil is the money of root.
Say, I need to log in from different stations, do I have to carry my password on a floppy/cd/whatever? What if someone steals my storing device? How do you log using an image via telnet? =)
MS figured out that it can gather more than just boring ol' text information... It can gather images or sounds, or almost anything.
How about DNA security, where you sign your contract in blood!!!???
Why does that sound familiar?
If an image is 1280x1024 and is sensative to a 10x10 pixel area, that gives the user a grid of 128x102 to click in. A total of 13,056 clickable squares. If the user's password was 5 clicks long, that would give them 379,359,275,350,832,971,776 possible passwords. Is my math correct?
I'll use that guy from goat.cx... That'll keep people out of my computer
So now you have to remember the order in which you click on an image? Maybe that's easier for some people, but certainly not for me. I have one password that I've used for the past 15 years or so. It's 8 characters (9 if I need to mix numbers with it), and it appears completely random.
I've been using it for 15 years an nobody has ever hacked it. All you have to do is have one of these and remember it. Almost anyone can remember a single 8-10 digit password, if that's all they use. Just make one and stick with it. Maybe you'll need to change it every couple of years, but even so, once you have it down, it's pretty easy to remember.
Is it hack-proof? Of course not. Not even close, but for most applications where a password is needed, it's more than sufficient. I doubt anyone will take the time to try to hack my hotmail account when there are so many that can easily be dictionary attacked. I'll always be the last one someone tries to hack because it will take too long to hack mine, compared to most.
Just my personal opinion. Obviously for some things, you simply need real encryption, but for most online stuff, a single 8 character/digit password is fine.
"This is especially interesting in context of the crappy passwords story that ran on Slashdot that ran few days back."
:)
And it is even more interesting in context of the the the using images as passwords story that ran on Slashdot that ran [sic] a few days back.
how about some modern art..
I visualize a blue circle on a which background.
Or a white line on a black background..
With images instead of passwords, the new Windows(TM)(R)(C) will now be fully average-monkey compatible.
People who like this sort of sig will find this the sort of sig they like.
Pictoral Passwords (using abstract art)
(It isn't karma whoring when you're already at 50.)
"And like that
Well, I've got this idea quite a few years ago, but honestly, did you ever try to login with someone watching? And its much easier to watch the monitor than your keyboard. And at least I can type my twenty something passwords reallllly fast and have some intentional typos in them, but - man - how can you click on pictures without someone seeing the pointer moving over the right pictures....
If programs would be read like poetry, most programmers would be Vogons.
Well, isn't it possible to still try and hack someones password simply by brute force? Someone could just emulate all combinations of the mouse click on all pixels of the picture. Also... how the heck is someone going to memorize pixel locations better than strings????
-=-=- I don't suck... you blow. -=-=-
Actually, this might not be as stupid as it sounds.
If the keyboard picture had the keys swapped to new positions every time, it would be impossible for anything but a camera to deduct the password (as opposed to the vulnerability that makes snooping passwords possible because of the timing between keytrokes).
Of course, it would still be vulnerable to attacks from the person standing behind you...
Don't make fun of my speling, english is my 2nd language...
I've seen something like that. You could coose an image (the more complicate, the better) and define some points, which you have to remember. To login, you have to click the points you selected before, with more or less accuracy in a predefined order.
Life sucks.
Monkey, Sheep, Sheep, Monkey, HORSE. you HAVE to remember horse! Because if you don't.... You'll have to click on all the images! or... or could just click clippy for help. . .
Instead of just clicking an image, how bout a setup where choosing an image is only the first step. Once the image is displayed, a user then has to click or move the mouse around the image in a certain manner. Much, much more secure then just choosing an image.
Did n't the password in that movie was images?
Okay, so lemme guess: you picked a mental pattern on your keyboard to repeat as your password, and you use that password anywhere?
Getting your password would be trivial by shoulder surfing, and once it's obtained, every account you have is wide open. Yipee skipee!
That, and if you used your magic password on a system with an unscrupulous operator, that operator now has the key to every other account you own.
There's damn good reasons why you're told not to reuse passwords.
Lotus Note on the Mac (I've never seen or used the Windows version) has a little something kinda like this in their password dialog.
As you type in your password, small images in a 2 x 2 layout change according to what you've typed. Even though the password text is bulleted out, you eventually come to recognize the 'correct' four images and know when you've misyped your password before hitting Enter. IMHO, this is the best feature of Notes, which otherwise sucks-- Lotus might not have been the first to use this idea, but it's the first place I've seen it.
And now I'd like to complain about the increasing retard-ification of our society. How can people be unable to choose a few non-obvious passwords (hell, just some random sequences of alphanumeric characters will do) and remember them with a mnemonic device? Why must we create an authentication system geared to the stupid so they can easily exist among us? Maybe they'd smarten up if they chose "password" as their password and had their checking account cleaned out for the third time as a result.
Of course, I should have seen this coming when McDonald's started using cash registers that had photos of the food on the keys and spit out the customers' change automatically, without the operator having to overtax his/her brain thinking about how a quarter, a dime, a nickel and three pennies have to combine forces to make 43 cents.
~Philly
Sweet, now I can log in and enter my password without even touching the keyboard!
Not surprising that MS would come up with this knowing their track record with security...
Consider anyone standing behing you while you select the appropriate login. They are bound to see the images you are selecting as your login much more clearly then the key combination you would have typed.
-- bartman
Can you imagine having an emergency in our future-tech age?
"No Bill, it's Black Guy, Asian Guy, Samoan Woman, Black Guy with the scar, White Guy with glasses! Hurry up before the Holodeck explodes!"
the most widely used passpic
Seems like you'd have to be really careful not to exclude the color blind. And the actually blind. Or just those with bad vision, or really poor visual memories.
It seems that a visual password would make it much easier for someone across the room to see and learn. One would have a hard time looking at my keyboard if they were behind me, but the whole reason any password login puts bullets on screen is so someone looking at the screen can't see it. Does this system use a mouse or is there some way to pick out the pictures using a keyboard with no on screen indicator? Of course, if that's the case, then this system may not be as idiot proof as they hope.
Picture this:
The Mona Lisa by Leonardo Di Vinci
The Scream by Edvard Munch
A picture of David by Michelangelo
A picture of a not quite so cute dog with a caption underneith it that says, "Fluffy".
I wonder which one of those would be the password. Hmmmmm.
Today's is now 25 years after the Punk explosion in England (1977) so I believe it would be a bad idea to ask today's 25'ers about which color their mother could have painted her hairs
Trolling using another account since 2005.
Problem is that this has nothing at all to do with how you actually pull out that memory. I mean, having this strong kinesthetics allows you to keep that password in your head, but it does nothing for pulling it out (unless you always use the same password... more on this later)
What triggers that memory really has to be one of four things: A sound, an image, a phrase (written), or a touch. That's not true, at least with me (functional keyed-retreival) but most people at least fall into those four.
This is a cue that your mind uses to pull out those memories at the appropriate moment. The feedback starts and you can whip out your password completely automatically, right?
Some "realistic solutions" to these problems include: biometrics - which don't require any memory, single login - which limit the number of cues needed, asymmetric key - which relies on math, etc, etc.
I say "realistic" because people have used them and they do work. They don't affect that memory pathway in and of itself, but instead rely on more durable pathways (e.g. outside of the person :) )
Unrealistic methods? Pictorial passwords. Besides the obvious that they're useless to the blind, many (dare I say most? nah, I couldn't find those numbers) people lack a visual eidetic. This means that they're very easy to confuse with similar images - because they cannot be used as triggers for their memory- They simply cannot remember seeing that.
Surely, they can remember the memory of seeing, or the act, maybe if they described it to themselves (common: turning a visual cue into an audio one, but this is time consuming and rarely works for long) - point being, it pushes way too much emphesis on only one cue.
With our current method, I gain some visual cues; input fields on the left, on the right, a popup, etc. I also gain some functional cues (mail related? do I know these people? am I these people? was this just a test?)
I then turn all these cues into the blinding flash of realization that sends my fingertips into a frenzy typing out the appropriate login and password for wherever I'm at. (except on slashdot, i'm a wuss... i use cookies :D)
My cues may not be the same as everyone elses' but everyone does have cues. I think that changing the focus of what we remember is less important than changing the cues by which we do remember.
So when you call support to get your lost password, will they ask you what your mothers maiden hair color was?
:)
... or the size of my "structrual beam", which of course i can't exaggerate because the wood just isnt big enough
ok. it's... the password is ONE - as in ONE foot...
my blog
Just because it is a mouse, doesn't mean it can't be snooped. Mice and keyboards both use serial communications and can be captured by many means.
The Microsoft Mouse(tm) protocol sends out a three byte sequence to signal a mouse movement. The current from the wires of a serial mouse can be picked up remotely with a good antenna that can sense the large RS232 voltage transitions at a slow 1200 baud. From another room, you could track mouse activity just as with a keyboard.
Oh well, it's not like we haven't seen this before
people are just clicking on key points in a picture.
To me it seems that is not much different from anything else, you have a picture of a face, there are probably 5 or 6 key points, the eyes, mouth, ears, and possible the top and bottom of the head. People would key in to the same features of the picture, after that, it just become an order of what is clicked, and people would tend to be predictable about that, forming geometric patterns, like going in a clockwize or counter clockwise pattern.
In the end, things might be safer in the short term, but it jsut means that the hackers jsut need to read up on a new set of psychology books, Once they got that down, you are back to where you started
"So when you call support to get your lost password, will they ask you what your mothers maiden hair color was?"
Har, har, har, I know what color hair YOUR mom has!!!
AC (1) CmdrTaco (0)
Ok guys, here's how you can use the power of visual identification and still have a cryptographically secure system. All of this and it's implementable RIGHT NOW with current tools on a standard linux distro.
.pngs, .jpgs, a mix of verious types or whatnot. All that matters is there's quite a few of them on the machine. I'm going to use the /usr/kde/2/share/icons/hicolor/48x48/ directory. This directory contains 5 subdirectories with a total of :
/ fip180-1.txt)
3 24
D 7F
:
./ after all. :)
1. Take a directory full of images, it doesn't matter if they are
find . -name '*.png' | wc -l
297
pictures. Given this, we can do som basic combinatorics (permutations of these standard pictures) for any value of 297 choose n. Using the permutation of (297 3) gives us 25,934,040 possiblilties (remember the order of choosing pictures is unique). It gets even nicer at 4 (7,624,607,760). Why am I bothering with this? Let me show you a snippet of python code:
# requires python 2.x
import sha,sys
print sha.new(sys.stdin.read()).hexdigest()
This little beauty will compute the hex-digest of the Secure hashing algorithm (http://csrc.nist.gov/publications/fips/fips180-1
.
All you have to do to use this program is the following:
$ cat apps/kedit.png filesystems/zip.png mimetypes/widget_doc.png | hex_sha.py
066686143327A8A582E5F5333A98D6C3F1426
or, if you prefer:
$ cat apps/kedit.png mimetypes/widget_doc.png filesystems/zip.png | hex_sha.py
2C35BA8998BAAEA70008AE41E31F923142A48
Obviously, order matters. Starting from this simple building block I'm sure it woulndn't be too hard to have kdm/gdm/xdm use this alternate method. There are c libraries available (openssl) which accomplish the same feat.
In short, this can be implimented in a weekend by a skilled hacker. One could even see crative ways of assigning short characters to each picture so that clicking isn't necessary. Something along of the lines of:
Actions == A
aPps == P
Devices == D
Filesystems == F
Mimetypes == M
And each subdirectory use the same method as well. So instead of catting those three files via the CLI, I could opt to type
PE == aPps/kEdit.png
MW == Mimetypes/Widget_doc.png
FZ == Filesystems/Zip.png
So I could type PEMWFZ (case shouldn't matter as we're indexing through a series of directories/files) and get my first catted line above. The second line would be PEFZMW.
The weaknesses in the algorithm described above lie in the strengths of SHA and the number of choices (I'm using 3). Since SHA's collision space is larger than (297 3) The weakness lies in the permutation. As I showed above, it's pretty damn big. Make it 4 (and all pw's become 8 characters).
Hardest part is the passwords are still gibberishlike. Or are they? Each grouping is paired in twos naturally. The password in ones's mind isn't PEMWFZ, it's PE, MW, FZ. If one can visualize the picture with the grouping then there is a direct visual association. This would appleal to most hacker-types. And the non-techies can even just opt to scroll through the pictures clicking on the 3 (or 4) that comprise the password. There could even be an option displaying the shortcut keys as the pictures are being clicked in case the person can't remember one of the mnemonic groupings. This must be done in absolute secrecry should the should-surfers wander by.
You guys get the idea. I'm just spewing ideas about this topic.:)
(And to others about this "dumbing-down" passwords; I think my hacker/non-hacker solution above compliments both types nicely. It also gives rise to REAL passwords without having to memorize `a09GD3hz'. A compliment of pictures and shortcut blocks works well within the human mind -- try it if you don't believe me. On top of this, it eliminates the possiblity of people choosing 'god', 'stud' 'master' and other such obvious passwords.)
Feel free to flame my constructive brainstorming. This is
Hope we won't all have to go through that to unlock our computers.
today is spelling optional day.
I think most people use one or a set of very few passwords, as well as usernames. It's inevitable, but has it's problems.
I'm sure If the Slashdot crew wanted to, they could use the usernames and passwords from here to log in to thousands of peoples Ebay, Amazon and Paypal accounts. Anyone that puts up a site that requires a username and a password could do the same.
I'm not trolling, but what good would this do if the MS users are still forced to "sign up" to some Microsoft "controlled" thing like MS Passport? We should have enough proof by now that MS are totally unable to keep something (except possibly the source-code to NT, but the computers holding it aren't connected to the 'net) safe.
It would be like installing a keycard + keypad controlled locking mechanism on a shed made out of plywood - with a swing-door in the back.
I wonder if mouse gesturing (ala Black and White) would make a good password protection system?
:)
I guess you could enforce a certain complexity to the password (no mouse up, mouse down).
This would have the great advantage that it would be tremendously difficult to teach to someone else...
Just a flawed thought. Find the flaws...
It signals movement, not position. So you have to guess the sensitivity and acceleration (although there aren't many combinations).
To counteract this, the mouse driver can put the pointer at a random spot when it starts, or apply a little randomness to the movement, especially when the mouse is moving quickly and the user won't notice a few pixels deviation
Besides, the click locations would have to be stored in terms of percentages to allow for scaling the image for display on different devices with different resolutions and still accepting the user's "password." Add in a tolerance factor since the user probably won't click the exact same spot, and look...if I display all your images so they're really tiny I can click wherever I want and login!
Reading through this thread, there are lots of valid issues brought up. I would agree that this concept alone would either be just as difficult as passwords (assuming the resolution of where you clicked was tight) or just as insecure as a bad password (assuming fairly forgiving resolution).
BUT, a simple pictorial password combined with a simple alphanumberic password could be very secure as well as easy to use. Far greater than the sum of either used individually.
I used to work at a large bank which employed this kind of multi-level security. A mag card got you into offices, a mag card plus a numeric keypad got you into medium security areas (teller lines, etc.). The higher security the area, the more techniques were added (retina scan, knowing your mother's maiden name, manager's name or department name, etc.). Basically, each aspect is individually attackable (stealing the mag-card, dictionary attacks, shoulder-surfing, password sniffing, etc.), but you have to know all of them to get access. Each obstacle in the way added a large measure of unpredictability and hence security.
I could even see this being used in a "telnet" (ehem, ssh) like scenario where a traditional userid and password are the first level, then some quiz (arranging shapes or colors in a specific sequence for example) is the second level. Each would be easy to remember, combined it would be very difficult to guess both (or several).
Basically, I think there is a great amount of promise in this kind of research. Yeah, you can shoot down each method as flawed, but combine a few of the methods and you can get some very powerful and easy to use security.
The only athletic sport I ever mastered was backgammon - Douglas William Jerrold
> If the keyboard picture had the
> keys swapped to newpositions every time
http://use.e-gold.com
click the SRK button to see such a keyboard.
SRK = Secure Randomized Keypad
I suppose this defeats the purpose of using asterisks to conceal your password in a place where people could be watching...
Imagine a Beowulf cluster of these!
Detecting acceleration of the mouse is not an issue when the amount of movement is encoded in the sequence. Also, the initial position of the mouse is fixed upon boot.
Its easy to scan and parse where the user is going to be. After all, this is done in software anyway! It makes no difference if it is done on the host computer or a remote spying box.
byte: contents:
0 1 L R Y7 Y6 X7 X6
1 0 X5 X4 X3 X2 X1 X0
2 0 Y5 Y4 Y3 Y2 Y1 Y0
OK, most people have quite a few accounts and some of these accounts are used infrequently. And in a large number of these cases it is considered more important to be able to access the account than to protect it. Quite frankly, I could care less if some one accesses my voice mail ... it would be inconvenient if I have to call verizon to reset my password.
The registers make the servers more efficient. Also why doesn't anyone complain about grocery stores and barcodes. Why don't we give them a sheet of paper and a pen, so what if it takes them 15-30 minutes to ring up a customer ?
Woe be on to them, all who rise against poor people, shall perish in a the end. Buju Banton
How would this system work for remote login ? The terminal has to download the whole image series from the server ? The image series has to be standardized ? Both seem flawed IMHO.
"Users simply remember exactly where on the images they clicked and in what order."
How is that better simpler and more secure? 99% of the people will simply click on the middle of the picture, and boom you're through. Of course then there might be instances where you have to click a minimum of 5 places, so suddenly everyone is clicking on each corner and then once in the middle.
Personally, I'd just as soon stick to my text passwords. I don't find my passwords hard to remember, as I utilize a seqeuence of rules to generate the password. That way I can choose a word (I usually like titles of Books/Movies/Albums/Songs) and run it through my little set of rules to product a string of characters that bear little resemblance to the original word, but is still easy for me to remember, because I don't have to remember the actual password, just the methodology to get to it.
If by pictures for passwords, they had meant that you supplied (uploaded) a special image of your own personal creation, and then that image is authenticated using an algorithm that generates a key by the values of the pixels in the picture, and then matches it via a public/private ssh key authorization manner; that, I think would be pretty slick.
Well, I'll quit rambling now. I just don't see how clicking on parts of a picture is easier to remember or more secure than typing in a string of text.
RFC2119
"I don't think you can create a password that is easily memorizable that is 20 characters long," Kirovski said.
:)
It is not difficult to create an easy to remember, 20 char password. Emphasizing pass phrases instead of the traditional 6ish char passwords would be a much better solution than this bizarre click on the image thing. besides, how would it work with SSH
I was driving to work in heavy traffic, and this guy drove up next to me and kept gesturing at my car. I think those gestures were to try to login to my car via visual passwords.
He did not try very hard, for he used the same gesture over and over with very little variation. You Americans are not very creative at image password hacking it seems.
Table-ized A.I.
Now I've seen it all!
Appended to the end of comments I post? 120 chars?!
In keeping with Microsoft's tradition of rarely doing its own innovation...
l #DEJAVU
Many years ago somebody was selling Automatic Teller Machines that used this approach instead of numeric PINs. I wish I had a reference but this was way pre-Web (1970s).
Also, this was discussed at Usenix 2000 and CrypTec 99 - see:
http://paris.cs.berkeley.edu/~perrig/projects.htm
and on Slashdot on Dec 28, 2001
The only good weather is bad weather.
"That's right Mr. Johnson, you password is boob, nipple, tongue and lips."
skuzzywhores.com now has downloadable pass-pictures of your favorite screen sluts, from Anal Ashley to Luscious Lydia! Why not have some fun with your security? Download 'em now!
Read the EFF's Fair Use FAQ
Be careful of those super-intellegent dolphin hackers, man! ;-)
8==8 Bones 8==8
So who else but me thinks that the photos used in Johnny Mnemonic as a password will somehow become common in the world of image passwords?
I guess if we used images as passwords, there'd be a smaller possibility of having the same password as someone else.
If we're talking about taking pictures yourself of something you cherish to use as a password, no two people are going to have the same picture.
If we're talking about capturing an image from a TV show or movie, again, not too many people will have the same exact image. Since there are normally 30 frames a second, the pixel arrangement in the picture can change up to 30 times each second. Therefore, images can sometimes be similar to the naked eye, but when it comes to it, their pixels will be arranged in just a slightly different way, depending on the action of the movie/show being captured.
Anyways, that's my 2 cents.
burnboy
Do you think it would be hard to guess where someone was clicking?
First of all, that one was different (this requires you to click in very particular places in the pictures, not just on the right pictures), and secondly most of the comments on that were "This is stupid" and all the downsides. This idea has even more downsides than that.
Visit me on #weirdness on the Galaxynet.
...intelligent to contribute, then just post the damn story!
"So when you call support to get your lost password, will they ask you what your mothers maiden hair color was?"
if pictures are easier to remember, won't it make it easier for someone looking over your shoulder to see, and remember, your passwd??
Market droid: Our research indicates that our users aren't being humiliated nearly enough.
Pointy haired boss: Why don't we make them play pin the tail on the donkey before they can use the system?
Engineer: I suppose I could work it into the login sequence.
And now I'd like to complain about the increasing retard-ification of our society. How can people be unable to choose a few non-obvious passwords (hell, just some random sequences of alphanumeric characters will do) and remember them with a mnemonic device?
I assume you're referring to my secretary, who seems to believe that the little light at the top of the keyboard (the one with the words "CAPS LOCK" next to it) is the power light for the keyboard. The one who didn't understand why I wouldn't give her an Administrator account, since her job includes administering some of our (expense) accounts. (She pouted for two days over that one.) The one who refuses to log out of her machine at night, because she likes coming in to work and having her computer ready for her? (Note, that point applies to many of my co-workers.) The one who made me turn off the 30-day password cycling, because she didn't want to remember "all those passwords."
The real problem here is that these people don't see the need for security. They think of computers as fancy toys, and maybe something to write letters. "Big deal--you don't need security for that. I don't care if somebody reads my letter to my brother, or plays my games." While that may be fine at home, I'd really rather people not get into our financial accounts, or our grade records (I work at a university). "Well, who would want to?" Well, for starters, any student who has a grade on that system. Anybody who'd like a little extra cash, from our pockets.
The real problem isn't that they can't use a decent password, it's that they don't want to, because they don't see the threat. Until this changes, nothing will change.
"Make it ten--I am only a poor corrupt official."
--Captain Louis Renault (Claude Rains), Casablanca
http://www.realuser.com/ has allready implemented this technology. You can use their service to store web-logins etc. on their servers.
I tried it once and it was incredible how fast you can remember three or four faces in the right sequence.
good book, bad movie. can you imagine a beowulf cluster of those dolphins? >;)
At the MSFT presentation, Darko Kirovski - who seems to have carefully thought a lot of this stuff through - of Microsoft Research (MSR)talked about some of the simple ways that pure dictionary searches could weed out roughly 25-33% of all passwords.
He also dealt with the possible objections that people might have to trying to remember where they clicked on an image.
As an example, he hypothesized that a physiologist, or a surgeon using this system night very well choose anatomical images as their source material.
He also pointed out that the keyboard layout itself is an 'image'.
Kirovski was sensitive to the behavior change - going from text passwords to image-based sources -
that consumers would have to engage.
He also admitted that there still isn't a password system that is 100% fullproof.However, by the time he finished his presentation, I was convinced that the image-based password system is a huge improvement of the current text-based model.
This isn't to say better means for password protection won't cone along - they will. But right now this is an interesting alternative.
btw, Kirovski is an impressive individual. At the same conference, he spoke of research done on something called 'content screening' as a way to protect piracy. The rigor with which he approached both topics - from several perspectives - was impressive.
I guess we couldn't use image passwords on /. : people would confuse them with those new adds
there's no place like ~
There are better things we can do as sysadmins to help fix this problem.
I really don't get the difference between this image stuff and character passwords. All I see on the keys on my keyboard are pictures... they just happen to be of letters. What do they put on yours?
Great, we are so stupid that we cannot remember a basic password combination. Now have to treat people like 3-year olds we cannot remember a word, so they have to use a "Picture Book" to log in.
Remember the Cow for goes Moooo!
Just think of all the people that will use goatse.cx or a picture of CowboyNeal as their password. :P
Too bad there is not a +10, funny. I vote this for the Oscar or funniest posts. Great job!
And stupid is the idea of having picture for passwords. Am I the only one to see the sent key has actually fewer bits than text ones?
Boy, everytime MS innovates, script kiddies rejoice everywhere!
They are probably looking at something like Passfaces. See http://www.passfaces.com/
You go through multiple screens of multiple faces, and have to get the right combination.
Used it before, hated the idea at first, got quite used to it soon.
I never thought I'd finally be able to use my ass as a password.
"Derp de derp."
Even they realized that most people would likely have some variatin on one favorite gesture to use with MS software.
And complex gestures would begin to resemble an arcane and ancient magic ritual. (which is an idea for a sf story someplace)
"It is a greater offense to steal men's labor, than their clothes"
--
m iso socially aware artistic geek pen-pal, m or f, in '1337 edu. jazz, poetry a must.
email me (click my user info for addy) if you're interested.
Now then. Let the games begin.
........
First of all, here's a bit of a rant. Let me disagree strongly with Darko Kirovski, the "cryptography [...] researcher at Microsoft" (article) who created the prototype, when he says:
Now, I'm just an average slashdot user. I've never worked with anything that is worth so much as protecting my keyboard from being TEMPEST-ed as I type my password. I'm certainly no cryptography expert.
But even *I* know that you can create easily memorizable passwords 20 characters long, and, in fact, far longer.
First of all, let me introduce you all to diceware. Diceware, slashdot. Slashdot, diceware. (How do you do, how do you do).
Now diceware here is run by a guy who knows about security. He's paranoid. He doesn't just "come up" with passwords while trying to avoid using any obvious components -- oh, no, he generates them completely randomly, and accepts whatever he comes up with as his password. So randomly does he generate his passwords, in fact, that he uses casino dice rather than trusting any kind of hardware.
But wait, it gets better.
How does diceware work? Basically, you use dice to choose a group of short English words that, since they're words (or can be treated as words by a human, such as the "word" ijk), are easy to remember.
More specifically, you roll a die five times, and put the five numbers together and find the corresponding word. (For example, if you roll 2, 6, 3, 1, 5, you search the list for 26315 and find that your word is "Frank").
The only caveat is that before using this list, you should manually (or with a program of your own design) check to make sure 1) that no numerical combination is missing and 2) that no word is associated with more than one combination.
In other words, you shouldn't trust the guy who made diceware, and you don't need to. It's just the principal of the thing -- a list of unique items on a one-to-one ratio with a range of numbers, each of the items of which is easier to remember than a mere number. (But, because there are equally many of them, will be equally "random".)
Now let's do a bit of analysis together of how secure this is.
Now let's rip apart Kirovski's statement that you can't remember 20 characters.
Before we do, let's point out that no one needs 20 characters, since even if you take a "character" to mean just any of the 94 ASCII values that a user can easily type, we'll even exclude the tab and space, this comes to (6.5545888 bits of entropy per one-of-94-characters * 20 characters=) 131.0917 bits of entropy. That's more than 128 bit encryption needs for a secure key! And this includes only the following characters:
! blah " this # lameness $ filter % really & sucks ' don't ( you ) agree * of + course , you - do . / 0 1 2 3 4 5 6 7 8 9 : ; ? @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z [ \ ] ^ _ ` a b c d e f g h i j k l m n o p q r s t u v w x y z { | } ~
Obviously, if you include more in the definition of "character", then the amount of entropy in 20 characters becomes ridiculous.
But for now, let's assume that Kirovski really did mean 20 characters, as I have defined them, or 128 bits of entropy. Is this "easily memorizable"? Sure is, if you use diceware.
For each word, we'll roll a die five times and get 12.92 bits of entropy. This means we need 10 words to get 128 bits.
Here are my results:[4]
65566 35115 24266 14326 54314 63345 41616 12265 44346 56243
I look these up in the word list, and get:
"56 junk elba bleat lard wacky sermon annex one swept"
as my pass-phrase. Is this "easily memorizable"?
Sure is:
Picture: arm stretching borders of alaska over canda.
It took me less than thirty seconds to come up with vivid pictures for this, then another minute to associate these sentences and pictures with the actual words (bleat for bleating, swept for sweep or sweeping) and if I remind myself of it in a few minutes, then in a few days, then in a week or two, I'll have it known forever. Compare that with memorizing:[6]
JLEwx;+?o9bH`"|6r%Bo
And you see why diceware is a good idea.
The fact that someone who is a supposed expert in this doesn't know about it is in my opinion inexcusable. (Of course, he might know that twenty characters' worth of entropy can easily be made to be memorizable, but his statement does not reflect this.)
Incidentally, it takes me between six and seven seconds to type "56 junk elba bleat lard wacky sermon annex one swept" carefully enough that it's accurate without my checking it as it appears on the screen (I just closed my eyes and did this five times in thirty-one seconds.) And more than twice as long to type the random 20-character word, if I look at the characters as they appear, even though I use every one of those non-alphabetic characters frequently enough to be able to "semi-touchtype it" (might not hit it on the first try, but I know where it is and I don't look at the keyboard -- in fact, I couldn't now because I use a weird international one. [shrug] But semi-touchtyping doesn't help you when you see *'s instead of the characters...)
As for how much security the average person needs (we're not talking 128 bits here):
well, if you consider an 8 character random combinations of A-Z, a-z, and 0-9 that's 5.954196 bits of entropy per letter * 8 letters = 47.6335 bits of entropy, or less than four diceware words' worth. For example,
56 junk elba bleat.
You don't even need spaces (although I find it easier to type with them) since no diceware word includes a space.
Can you believe it, a simple thing like "56 junk elba bleat" being more secure than a completely random 8-letter mixed-case, alphanumeric word? Wow.
Okay, I've run out of steam. That ends my diceware rant, and I'll address this whole nifty picture thing now.
First let me offer these final notes, which didn't fit into my discussion above.
Okay. Rant ends here.
........
Back on topic:
From the article: "The key -- images, which tend to make more of an impression on people than strings of text characters."
This is true, but it is equally true that it is more difficult to uniquely identify member of a given set of pictures than it is to identify a member of a given set of words.
Picture the face of the last high school English teacher that taught you. Now, this is a fine part of a password, because you can choose it randomly from a large list of objects (people you know), and you will remember that it's your password. (Or rather, it and a few more like it).
That is, if I told you that of the 2000 people you know, the following eight faces, in that order, are now your password, you will have very little difficulty remembering them and their order.
However, how will you make a selection 8 times from one of 2000 people? Supposing you know their names also, you can alphabetically list four at a time, doing a double-binary search (for example, A-M at the top, M-Z at the bottom, and the right side is the upper half of each of these ranges and the left side is the lower half).
You now need to make 5.482892 selections to select each of your 8 faces. That's 43 mouse clicks, each one followed by scanning four faces.
Of course, this is based on knowing the names associated with each face, and it would be easier just to type those in. In which case we're back to diceware.
If you don't know their names, however, just how will you select from 2000 faces? Well, maybe you can mimic the binary search with a selection from characteristic skin color, eye shape, etc. If you spend a few hours learning "human facial classification", I bet you can select just about any face you recall in eight or nine mouse clicks.
However, I doubt most people would be too keen on learning to input a bunch of characteristic features. (Even if the 2000 people aren't really people, but people from "Guess who?", who have either a large or small nose, either are wearing a hat or aren't, etc.[7])
The more specific method the article mentions, selecting a particular pixel range within a person's face, isn't something that people do on a daily basis (so much as memorizing and recognizing faces is), so I doubt most people could remember whether it's Mary's lower-right lip followed by where a dimple would be on her right cheek, then the middle of her left eyebrow, or the other way around. It's just not doable.
Okay, I need to go now. Enjoy the weekend, all.
~lts.
You can skip step (1) if you make a contract with yourself that if you ever roll a combination that for some reason isn't on the list, you will take the time to make word that is not on the list, and use that instead.
We'll note that hardly anyone uses the full ascii set, including control characters, in their passwords, but I suppose it's possible to use every character besides carriage return (and maybe even that), depending on the implementation.
There are only 96 keyable characters in the ASCII standard before all the international extentions and so forth, which include the tab and the space.
[4] If you want, you can follow along (and see that I didn't artificially select a particularly easy combination):
#include "iostream.h"
#include "stdlib.h"
int main(int argc, char* argv[])
{
cout << "Unseeded demo. NON-SECURE!"
(You can add indentation, I remove it because of the lameness filter.)
[5] Napolean's last battleground, I guess. Famous palindrome: "able was I ere I saw elba".
[6] this example from unseeded:
for(int i = 0; i < 20; i++) cout << char(rand() % 94 + '!');
cout << endl;
[7]On an aside, I figured out binary searching all on my own in playing Guess Who as a child. I figured out that the most efficient way of ending up with the opponent's person is, at each question, to pick a characteristic that only exactly half of my remaining choices had -- sometimes this involved making up questions like: "Okay, does your person EITHER have a hat OR a moustache (or both?). Yes or no?"
(Actually, I soon realized that I could get an answer faster by saying "does your person have any of the following:", for that particular form of the question, but that doesn't apply to all boolean expressions I asked).
--
m iso socially aware artistic geek pen-pal, m or f, in '1337 edu. jazz, poetry a must.
I don't know if any of you remember (or knew), but there were several programs for "hacking" AOL which used graphical passwords. ( One the comes to mind had a picture of a girl, a dog, a bird and an arrow and you had to click them in a specific order to unlock the program. )
Another "innovation" by Microsoft.
If the user, despite ISP encouragement, chooses not to use a proxy, that should be his choice. He is paying for the bandwidth, and is assumed to be aware of the possible performance hit.
This was discussed in the vuln-dev mailing list after Comcast implemented transparent proxying.
This raised quite a stink when Comcast's logging habits were revealed. Oops.
There is obviously a performance degradation involved with re-resolving the address given to the cache server. Furthermore, requests now appear to be coming from the server, not the actual user -- potentially breaking host-based authentication systems.
I've also seen these cache systems horribly implemented. An IRC network that I administer recently starting checking for HTTP proxies on connection. This was performed by connecting to the remote user's host on certain ports (80, 3128, 8000, and 8080) and then issuing a CONNECT request. In more than one case, a blatantly stupid ISP redirected _incoming_ port 80 traffic to their server -- WITHOUT any sort of access restrictions on their proxy. Sort of ironic that they were probably using untold amounts of bandwidth for 1337 bounce kiddiots.
Proxying without consent is an Evil Thing.
This sounds like yet another attempt to make things "easier", with no understanding or attention to the security ramifications.
Paralogix has a similar password scheme. You click on a number of objects to create a password.
Sounds good, but it turns out to be very bad.
It turns out that the number of objects used on the screen made for less combinations than you would have if it represented a letter of the alphabet. (About 28 combinations per "drag".)
It gets worse. Due to the way the interface works, it becomes prohibitive to make large passwords. (A keyboard is much faster.) The interface passlogix used was drop and drag. Icons are not going to be much better. (You only have so much screen area to work with.)
Passlogix did one even better though... They made the order of the password not matter. (So "AAB" and "ABA" and "BAA" were equivelent.) For small passwords, it removes a fair chunk of the combinations. For large passwords, it removes almost all of it. (95% at 5 characters and it gets worse from there.) I expect similar things from Microsoft if they actually do this.
I have suspected that Microsoft considers most of their users to be illiterate. It frightens me when I see evidence that my worst fears are confirmed.
"Trademarks are the heraldry of the new feudalism."
If you use an image map in a form input type=image the reply you get is action.x=a and action.y=b - depending on the complexity of the image you might get a pretty good "password" out of an a & b range. Helping people remember where they clicked will be more time consuming than reminding them of their password. But a sequence of images.. not a bad thing to try :)
Of those to whom much is given, much is required.
It wouldn't matter if the images move to a different position everytime you try to login.
The images would have to be doctored to work with software that could convert pixels to numbers and encrypt them.
That would certainly require a couple hundred programmer to doctor those image. Add another 100 for the pixel to number algo. At least! While you are there, while not add 50-60 assistants to make the coffee?
When the user selects his password, he also selects a cursor style or color. When logging in, nine or ten fake cursors in different styles and colors also appear over the image in random locations. All the fake cursors move with the mouse, but their planes of motion are each reflected or rotated some amount from the origin.
All the user has to do is watch the preselected "real" cursor while clicking and ignore the others. No one else will know which one is valid.
Here's the final paper.
For what it's worth, I tried out RealUser when it was in the news, specifically because I group myself with the other people here who say "I never remember faces in real life, so this isn't for me."
... that this is a brand new idea. It's not. It's been around, and there is more than one proof-of-concept implementation.
... that text-mode console users will love this. They won't. In fact, I'm pretty sure I'd never use it. But no one cares. I'm not in the target userbase anyway, because I'm already good at remembering passwords.
... that this is only for people who remember the names of the people they meet at the bar. I never remember names. Luckily, this is more like remembering that you're supposed to remember the name for a specific face. ("Click the faces that would embarass you most if they remembered your name.")
... that users will create stupid authentication keys no matter what. That seems like a problem with the MS technology as it's described, but it doesn't seem inevitable. RealUser, for instance, is arguably superior in this respect. The RealUser face-sequence is system-generated and (presumably) more random than any of my user-chosen alphanumeric passwords.
A few times since then (and just now), each time after months of not giving a single thought to PassFaces, I've logged in to prove that it wouldn't work for me. I assumed that I was too much of a geek to remember a sequence of faces. The fact is, I've logged in successfully on the first try every time.
The point is not:
Offered as an option alongside alphanumeric passwords in technologies with a large population of less-than-tech-savvy users, image-based authentication seems likely to result in lowered use of "remember password" checkboxes and/or "forgot my password" tech support calls. Most technologies that have such a userbase require a graphical interface in the first place.
The weakness of this authentication method against shoulder-surfing is my biggest complaint. Again, RealUser seems better poised to address this problem. Both RealUser and the MS solution will require images to be displayed. However, since RealUser doesn't care where I click within an image, they could change the selection method to a keyboard-based one to make the shoulder-surfing threat more similar to the same threat against alphanumeric passwords.
In my opinion, the most interesting thing about this article is the fact that "Researchers at Microsoft Corp." are making news by "[working] on new types of passwords". I, and others here, already tried the technology they're working on over a year ago and in a form arguably superior to the one the article describes.
goat sex anyone.
Baduh pun
don't you think such pic-passwd's would be easy to crack? Since the 'memorable' parts of the pic would be the breasts, "private parts", and other key parts(particular to an individual's festishes like belly button, legs, lips, etc).
Not too many permutations there. But hell, would it be *fun* to crack(not to mention _getting_off_).
:)
For those who want to see something like this in action, go rent the movie Johhny Mnemonic.
I could never remember which cheeses to click to get past the nag screen for Monty Python's Complete Waste of Time (or whatever it was called); why on earth would I remember any better which body parts, mammals, and reptiles to click in what order so I could log onto my computer??
:)
:)
Or as a friend once put it, "I don't need pictures. I can read and write."
More seriously, it occurs to me that unless the images came up in a random order each time, password sniffers would merely need record mouse click position. And once the password images were ID'd.. Hmm. ISTM such images should be user-defined to be more secure, because otherwise sooner or later some sniffer is going to know how to ID the OS-supplied images that were clicked, regardless of screen placement.
I just had this vision of people using their fave porn thumbnails as their password images, leading to this:
Invalid password: you must include at least two tits, one ass, and one other body part.
(thanks to whoever made the post that inspired this
~REZ~ #43301. Who'd fake being me anyway?
It seems logical to be researching alternative password methods for future HCI methods. With devices from Microsoft like PocketPC and TabletPC text passwords are not the ideal method. Keyboards won't die for some time, and text passwords won't go away either, but as Microsoft includes handwriting support in OfficeXP and on PocketPC they seem to be on the move toward "Pen Computing". They never pulled it of Windows 3.1 for Pen Computing but it probably will in the future. Biometrics aren't necessarily the only alternative; this is research after all.
I have a three year old. Rather than tell him to type some list of letters, I grab a picture from his favorite cartoon and he picks three favorite animals. You have more than enough data here and great ease of use for normal people are viaual creatures. You have data from the picture to use as the base of your encryption key. You have randomness of the orders of the points picked. Also remember, you don't have to pick a "point" you could pick a region or as things like MPEG 7 (moving sprites) catch on, an actual object in the picture (or representation thereof) rather than just the bits. You could make the behind scenes as elaborate as you want and the end user wouldn't be bothered with it.
what about clicking in a certain part of the microsoft startup sound thingy or deleteing your regular boot files or if your really smart making your own boot files that only you have and always carying a disk with you... or voice recognition or sound playback .. or carrying your hardrive with you or hiding it in a concealed hiding spot under your desk and replacing it with a fake drive that looks real or backing up all your sensitive data and use a cd sequence like some video games and stuff or well thats all I can think of right now
don't store anything sensitive online or connected to a sources of transmition phyiscally ..
really privacy only slows society and causes division but I'm a poor ass bum with nothing to loose so I just don't care. I'm sure there are lots of people that feel they have things to "protect". Really it just casts a dark cloud when people turn society into what they fear.
Just use unkown encryption of files or access. Or mask file types or have dual file type i.e. something that is a text file when in reality a jpeg or viceversa... or have reverse binary rendering or a million other posibilities.. anyone that actually wants to hide data can. Anyone that wants to crack it might be able to but probably wont. Just envent your own system. Lots of really simple rules for security and none of them are mainstream. But it doesn't really matter its all tools for domination and power mongering. What ever happened to blind trust. Ok I trusted people with my ftp and they deleted my boot files so they killed that opertunity for everyone else. Even though anyone that wants to hack my system could in a second making it a elite server.. oh well maybe one day everyone will live in harmony and not have the need for selfish need I just hope it is before I am the last one left. Honestly it doesn't matter anymore atleat not until I get a sense of non blah to life if ever.
The picture thing seems like it might be good for kids who don't want to use a mac and might make it funner to use a computer. Also its something to keep intrest in windows and talk about for all you computer geeks(compliment):) Really just wanted to say in advance I know I'm not smarter than you and that makes me that much happier. It sucks not being really really stupid though I think actually its imposible to judge intellegence since anyone living has passed. anyway I still don't get this password and safety thing. Either I'm god or you are and either way either I win
hmm sooner
The perfect dictionary file for the new "crack" program: images.google.com
well, the login scheme would automatically change the mouse position to a random point on the screen. it would not be snoopable unless a guy was sitting behind the user (or unless the guy says each letter out loud trying to remember. i have seen this more than once...)
I don't know why people say it is so hard to remember passwords... I literally don't even know my password at work! I'm serious, here's why:
Think of a password thats easy to remember: your name then birthday, like JSmith032302. Easy right? but a bad password. So hash it on your keyboard. I type my passwords in using the key above and to the right. If it is a number or the key above and to the right is a number I 'shift key' it. I also use 'capitalization'. SO the EZ to remember password JSmith022302 becomes the password IEk(^u)@@#)#.
They don't get much harder than that. And as for using a different password for eash web site, use the website name as part of your 'unhashed' password instead of your name.
"I'm about to drop the hammer and dispense some indiscriminate justice!"
A coworker told me the exact same story when wondering why his password had expired. He was mildly pissed. I understood his frustration completely. In turn, he understood why I couldn't just make an exception for him.
At that point I thought of something, which I've never implemented but I think would be worthwhile. I think that when you change your password, the password strength checker should assign an expiration date based on entropy. If you want to use a password like "Cindy" - fine, but it will expire in 24 hours. If you use "a79xoibf", it will never expire. I'm assuming cracklib has a reasonable way to estimate password entropy.
Has this ever been implemented? I think, over time, such a system would encourage people to use good passwords. Having to remember a new password every week is a drag compared to keeping the same hard password for a year.
"How can you claim that you are anti-crack, while still writing a window manager?" — Metacity README
Choose pictures with numbered points in them...
FRA: STFU GTFO
Just for fun, I have been woking on an interface where the user
draws a simple ascii art picture to access certin
functions of the program. For example, if the
user wants to access email, he or she draws an
envelope using text chracters to read it. The
same thing can be done with passwords. The only
real draw back is that the user has to draw
the pictures exactly as they are stored in the file
I don't have to remember a lot of passwords, because I don't use a lot of passwords. How is this a solution? Well, for any and every account that doesn't matter (e.g. hotmail spam account, anything I sign up for) I use the same, stupid password. I don't care if someone hacks those accounts, all they'll get is all the fake information I entered when I signed up. Then I remember 5 complex passwords (8 chars or more, mixed caps, multiple non alpha chars) for the 5 things that are important.
And those are easy to remember, because they're usually phrases, shortened: "There's no Sex in the Champagne Room!" gives me: "TnSitCR!" as a password. Easy to remember, hard to crack.
Jake
Dating: while( 1 ){ call_girl(); get_rejected(); drink_40(); } return 0;