Color-blind people are bots.

[icomeanon6]

You know those images of text that you sometimes have to copy in order to prove that you're a person and not some spamming bot? Well, I've noticed that it's become popular to have the text in those images be almost illegible.

Examples:

Am I the only one who's noticed what's so moronic about these? If they're only used to deter bots, the words can be perfectly legible and still do the jobs, because the bots can't analyze the images. Or maybe I'm missing something, maybe they're trying to keep out people with poor eyesight.

 

[JX Valentine]

Argh, yes, I know what you mean. The ones with an incredible amount of noise in the background are the worst, really. As in, the ones with tons of dots that are essentially the same color as the letters, so it's difficult to tell if you've missed one -- let alone figure out what one is.

Agreed on the concept of legible letters being sufficient enough.

 

[Yamikarasu]

I'm pretty sure there are programs that could recognize shapes in images and read them as words, but yeah, sometimes it does get really ridiculous. Especially when letter and numbers overlap each other so you don't know in what order they go or what they even are.

And the top one could read F5yylz or F5yytz or F5yydz, that's annoying.

 

[Gummy]

I've always wondered why that make those verification things so so hard to read. And on top of that, some sites only allow to try a limited number of times before you're automatically deemed a bot. But the one I really hate are the case-sensitive ones. Certain letters look EXACTLY the same in both lower-case and upper-case forms.

 

[Gulpin]

Yes, I hate those things. They are hard to read. Some letters are the exact same looking, and all they do is confuse my poor brain o_O

 

[Ruphire]

I know! I agree that they could just put legible letters. Heck! They could even make them huge!

 

[Jorah]

I remember Rapidshare once had images of cats as letters, and it was absolutly impossible to tell what the letter was. It could've been another language system and I would never have known.

 

[jasonresno]

I remember Rapidshare once had images of cats as letters, and it was absolutly impossible to tell what the letter was. It could've been another language system and I would never have known.

I'm so glad someone brought this up. I never could use rapidshare because of those damn cats.

 

[ErickaVolt]

Yeah those damn cats, lol. Sometimes it is confusing because they stick out letters like F5yyFdz I could incorrectly type that as o and l.

 

[/Circa]

I hate it even more when you have other things in the background that look like letters. Then you don't know which ones are right. :S

 

[AsugetRanzo]

I mistake the squiggly lines for "s" or "z" sometimes.
It doesn't help that I read everything backwards either. XD

 

[h POKE]

I hate them all! I try to avoid websites that do that like the plague, but it's really hard. I don't see how making it hard for even people to read is good publicity for the site...

 

[Legobricks]

It's called a captcha. Some are a simple straight alignment of letters and numbers, but the majority are indeed some annoying mal-aligned and often indecypherable combination. Many also have a button beside them that lets you switch to different ones until you can read it, but that still hardly helps.

It's probably since it is possible that a bot would have soft/hardware capable of converting text in images to text as typohraphy.

 

[Idiot!]

Why can't they just give you a number of images and ask users to pick the appropriate picture.

 

[Astinus]

I hate those things. My eyes are bad when it comes to picking out small details on a computer screen, so when the letters can't be easily read, I fail miserably on them.

There should be a better solution when it comes to them. One that can deter bots and allow those with bad eyes to read them.

I have no idea what the second one is even supposed to say. :s Stiff? Still? Stitt? Sti11?

 

[Eon-Rider]

It seems they're becoming more complicated every year. Some CAPTCHA have me clueless for over 5 minutes sometimes.

 

[Petie]

It's probably since it is possible that a bot would have soft/hardware capable of converting text in images to text as typohraphy.

I'm quoting that part of Cassino's post but really supplying this as an answer to the general question of the thread. This software does indeed exist and is referred to as OCR or Optical Character Recognition software. It scans an image for characters and outputs the text it finds. This is usually used to transfer typed pages back into a digital format but can be used by bots to read the CAPTCHAs. This is why they are becoming more and more difficult to read (and even those often times aren't enough).

If you're actually interested in implementing a system like this though, take a look at reCAPTCHA, which can be found [here]. It tends to work a lot better than the other more generic ones and also generally isn't impossible to read.

 

[Soul Eater]

omg. I know what you mean! I always have a hard time seeing the letters and I have perfect eye sight too. They make them look so hard though and hard to read, basically for anyone that doesnt expect to do them. D:

I think they call it Capchtca or...something. I forget how to spell it. D:

I see where you guys are going with it too, because I've always missed a few letters and I've had to do another one. It always pissed me off to know end. I mean, I know they want to keep bots out but, your right, what about those with disabilities (Most mild to moderate disabilities or learning disabilities like dyslexia, make it hard to see letters like that too, not just blind/color blind people or people with bad eye sight.) They are going to have a hard time too. Now, they have it so that you can here the letters but, even that is really hard, because they repeat other stuff in the background, I've tried that once and I couldnt do it that way because it was like, sensory overload for me. x_x

 

[Mooshykris]

I have no trouble reading the CAPTCHA systems, though that may be due to my 20/10 vision :(

However, it gets tedious. And what I care about, in 40 years, my 20/10 vision won't be the same, but CAPTCHA will stay the same :(

~Mooshykris

 

[Venia Silente]

Oh yeah I hated those things.

The problem with captcha systems is not that they are more difficult to solve every day (even for a "normal human"), but that they have to be because that's a requirement of their intended test. Distinguishing a human from a program is not very easy in the Internet when you are limited to visual medium. Which is how most captchas are being done.

As for what Misayu said, that's why people who either are somehow disabled or temporarily impaired "must" have it more difficult when it comes to captchas: were the tests computable enough for them, they would also be plain enough for machines to solve in a blink. Like, removing color/noise from the letter images reduces the complexity of the problem in two orders of magnitude, if I recall correctly from college. The theory of human-machine recognition relies mostly on the main difference between "human data processing" and "computer data processing" being abstraction v/s speed, so for as long visual is the medium they're using, the only way to favor abstraction over speed is to add complexity. Tons of it.

It's a lose-lose-lose-lose-win situation: disabled people lose by design, normal prople lose by being annoyed to no end, advantaged people lose by being unnecessarily tested, content providers lose by distribution and (ugh) reputation, and connection providers win $ by extra traffic, as a captcha serves as a kind of "pay-per-view" access model and web searchers actually act in complicity by hiding the fact that the content is only conditionally accesible.

That kind of stuff can only evolve for the worse, as it was not too long ago in the past that "text images" were unreadable by bots, now they can deal with them with ease. As Petie said, OCR. Being used by everyone including Google. I've heard about sound captchas, too. But... it is harder to implement and filters out even more disabled people.

What I'd like to see would be semantic captcha (like, presenting a picture of a famous event and asking what year did it took place), but then everyone would complain it filters out "non-smart" people... and the cost to implement them at megalarge scale renders them inefficient.