This is no supprise
As facebook is all about stalking.
I'm putting joke as my icon as I hope I'm jokeing :(
From time to time we capture word verification silliness for posterity. It's been a while, but we've got another for you, this time from Facebook - and this time it ain't so silly. The text reads: "rape now". It was displayed next to the picture of a female work colleague our tipster (a Reg reader, of course) was adding to his …
As facebook is all about stalking.
I'm putting joke as my icon as I hope I'm jokeing :(
"Optical character recognition can stumble on as much as 30 per cent of the words it encounters."
If they don't know what the word is... how can they filter it?
...and it's a shame that an irrational fear of select vocabulary will probably be enough to can its usage in many instances.
Why is this a big deal?
Okay, it's an unfortunate conjunction of words.
I presume your correspondent was not seized with a sudden desire to rape his friend. I further suggest that anyone, faced with this situation, who *was* seized with such a desire clearly has some serious problems in the first place. It would probably be doing a greater service to humanity to do something about that, rather than complaining about the ridiculously rare slightly unfortunate consequences of a wonderful idea (the recaptcha system) which is providing clear and concrete benefits to the world.
Basically: oh, grow up, view this as the tiny triviality it is, shrug, get on with life, and don't turn it into a news story.
computer verification program really bother you your name isn't really Beavis or Butthead is it "it said nodick" hehehehe.
... the unfortunate captcha combination or the fact that Facebook exhorts people to "acd" friends (second graf, "Also, if you acd ... ") instead of add them? Or is this an OCR error in the screen grab software?
If you have watched Beavis and Butthead do America, you'll realize that this pair can't even read their own names. Thus I doubt that they'd find many jokes in the written form. Maybe if it was one of those visually-impaired audio captchas...
I personally think that every captcha should be offensive - given the nature of most of the internet, I just don't think that this would be out of place at all.
But maybe thats the ale talking. Or could be the mind control drugs that the government puts in my beers... Or not. I'm pretty sure that its one or the other.
This from the ReCaptcha FAQs:
"Are CAPTCHAs secure? I heard spammers are using porn sites to solve them: the CAPTCHAs are sent to a porn site, and the porn site users are asked to solve the CAPTCHA before being able to see a pornographic image."
But if a computer can't read such a CAPTCHA, how does the system know the correct answer to the puzzle?
Here's how: Each new word that cannot be read correctly by OCR is given to a user in conjunction with another word for which the answer is already known. The user is then asked to read both words. If they solve the one for which the answer is known, the system assumes their answer is correct for the new one. The system then gives the new image to a number of other people to determine, with higher confidence, whether the original answer was correct.
Where can I become here friend!
Tom: Well apparently they know what one of the words means, and the show the word that they don't know the meaning of to a number of social-networkers. If they all give the same answer then they now know what that word means. All very clever I think.
More to the point, if they don't know what the word is, how do they know that it's been decoded and typed in correctly by a human rather than just being random gibberish entered by a bot?
The words "chocolate teapot" spring to mind (and are probably out there in a captcha somewhere).
They give you two words, as in the example in the article. The OCR couldnt read "rape", and the computer generated "now". You answer both, and if you got the seeded one right, it passes you, and also gives your solution to the unknown word to the system. Of course, they collect a number of solutions for each unknown to verify the solution.
Not sure you got this one did you.
that there is a rather obvious typo in the image you are displaying...
Under the link to "add a personal message" it reads "Also, if you acd <censored> as a friend".
Surely El Reg can't be reporting on mocked up images?
So in theory, if everyone who see's for example the word "cat" types in "dog" the computer will think it is actually "dog"? i say we pick a random word (say "monkey") and anytime it comes up we type "santa" just to confuse other people.
ok it doesnt have to be those words but u get what i mean.
What is more troubling is that you are a complete cretin. Firstly, for actually thinking it said "acd" (when any fool can see that the image size and quality have obviously been reduced from the original screenshot) because there are some graphical elements missing due to said reduction; and secondly, for using "graf" to mean "paragraph".
I don't have to type in a captcha phrase when I add a friend?
1. On the few occasions I've seen bad captchas they've always been the first of the two words. Admittedly a small sample, but is it possible to poison the Carnegie program by entering one word correctly and the other a p155 take?
2. Does anyone know why it's called a Turing Test when my understanding of the Turing Test was can a human tell a computer apart from a human, not the other way around - a computer telling a computer apart from a human. Small difference, but the kind that keeps me awake all night!
I seem to recall from another article on this that the word they know is the one they use to verify your entry. If that's correct, they reckon there's an 80% chance the other word would be correct. In the above case if they'd typed rare now it may still have let them through. They collate the unknowns - it will probably appear on several peoples captcha forms (from people who entered the known word correctly) and they usually just pick the most common word. It's unlikely from the randomness of appearance that several unrelated people will type the the same wrong answer in. No idea how many times it hits people but it's quite a few I think
I thought that the campaign for more yellow flowering plants in agriculture was a minority pressure group, but their campaign slogan seems to be raising quite a stir.....
You don't have to type in a captcha because you've confirmed your email address on Facebook
"the rest of the world knows them as captchas, short for completely automated public Turing Test to tell computers and humans apart"
What on earth makes you believe the rest of the world knows that?
I see the rise of the machines , blame it on the horny AI inside the server
...how do they avoid showing words from their 1000-entry blacklist, when they don't know what one of the words is?
I know, I could probably look it up in their FAQ list, but I like seeing my name in print.
"It was displayed next to the picture of a female work colleague our tipster"
Damn you lucky barstards she looks lovely.
On a serious note though the automated text is almost as inappropriate as GoogleAds, and my spelling.
"the rest of the world knows them as captchas, short for completely automated public Turing Test to tell computers and humans apart. "
What a load of rubbish, the rest of the (albeit techie) world knows it as "Completely Automated Program to Tell Computers and Humans Apart" - nothing to do with public, nor is it a turing test (complete opposite, as Pie Man mentioned)... sometimes they are referred to as a "reverse Turing test".
Not doctored, just resized, and probably using a linear rather then cubic or bicubic algorithm, so it just dropped some rows and some columns, possibly evenly spaced throughout the image. It just happens that it dropped the column that contained the upright of the second 'd' in 'add', as can be seen looking at the letter 'w' above, or 'h' below. Clearly you need content aware image resizing (http://www.youtube.com/watch?v=qadw0BRKeMk).
"Not doctored, just resized, and probably using a linear rather then cubic or bicubic algorithm"
- either that or some fools also typed 'rapo now' as the answer -- trying to fool the machine maybe?
An interview with Luis Von Ahn and talks about recaptcha. Very good article I thought
I posted a response based on the content of the post, as is conventional around these here parts. (Well, either that, or a response based on Paris Hilton). Do let me know if you were expecting something else.
If you're suggesting the post was 'humorous' (in that special way Reg posts are allowed to be without actually making you laugh...), it clearly wasn't, as the opening paragraph explicitly states:
"It's been a while, but we've got another for you, this time from Facebook - and this time it ain't so silly."
So their "security" system consists of training a computer program to solve the problems submitted by their system to screen out computers?
"probably using a linear rather then cubic or bicubic algorithm, so it just dropped some rows and some columns"
"Just dropping" rows and columns isn't even linear, it's not any kind of interpolation at all, it's just nonaliased subsampling (aka blit-based rescaling). But yes, it's absolutely what's going on there, I reckon that grab was taken from an LCD display where the computer's desktop resolution setting doesn't match the LCD's natural resolution.
The drivelling inanity of conspiracy theories based on the "Oh look, I found a pixel out of place - that hasn't *ever* happened to anyone before, so it must be suspicious" line of reasoning really disappoints me. Can't the kooks be bothered any more?
I loved the story of the pirates who wanted lots of accounts on a web site
which used captchas. They created a Pr0n web site, and asked whoever
wanted to access the site to solve a captcha, actually taken from the other site...
To add to the blacklist
Wouldn't want to upset any of my sensitive clients (They will use any excuse to try and get out of paying)
They could just have random letters and numbers, thus ensuring that no real words could be spelled. Of course, then they would probably forget to omit 1/i, 0/O and other similar ones, resulting in far more people being labelled as bots.
Would this be a bad thing though?