In Search Of The Perfect CAPTCHA

Advertisement

CAPTCHAs, or Completely Automated Public Turing Tests to Tell Computers and Humans Apart, exist to ensure that user input has not been generated by a computer. These peculiar puzzles are commonly used on the web to protect registration and comment forms from spam. To be honest, I have mixed feelings about CAPTCHAs. They have annoyed me on many occasions, but I’ve also implemented them as quick fixes on websites.

This article follows the search for the perfect solution to the problem of increasing amounts of human-generated spam. We’ll look at how and why CAPTCHAs are used and their effect on usability in order to answer key questions: what is the perfect CAPTCHA, and are they even desirable?

The Incentive To Act Human

To understand the need for CAPTCHAs, we should understand spammers’ incentives for creating and using automated input systems. For the sake of this article, we’ll think of spam as of any unwarranted interaction or input on a website, whether malicious or for the benefit of the spammer (and that differ from the purpose of the website). Incentives to spam include:

  • Advertising on a massive scale;
  • Manipulating online voting systems;
  • Destabilizing a critical human equilibrium (i.e. creating an unfair advantage);
  • Vandalizing or destroying the integrity of a website;
  • Creating unnatural, unethical links to boost search engine rankings;
  • Accessing private information;
  • Spreading malicious code.

All of these incentives lead to profitable or otherwise gainful situations for spammers. Automating the process obviously allows for superhuman speed and efficiency.

Those who run websites know that this is a big business and a big problem. Akismet351, the popular spam killer (commonly seen as a WordPress plug-in)2, catches over 18 million spam comments per day and has caught more than 20 billion in its history. Mollom363, which provides a similar service, catches over half a million spam comments per day and estimates that more than 90% of all messages are spam.

No amount of asking nicely will stop the spammers, but their greed can be used against them; using automated systems to increase profit does have a weakness.

Enter the CAPTCHA

On one side of the coin is the spammer; on the other is the humble website owner, a pleasant sort, who experiences common problems:

  • Blogs and forums that sink under the weight of spam posts,
  • Accounts that are registered under false pretences for unlawful purposes,
  • Bots that ruin the dynamics of a website,
  • A dive in the quality of content and the user experience.

Automated spam plagues website owners to no end, so CAPTCHAs are appealing and compelling… initially. The time needed to moderate and review user-generated content versus the time needed to implement a CAPTCHA is what pushes most developers to do it.

In fact, CAPTCHAs are used a lot. The reCAPTCHA124 project estimates that over 200 million reCAPTCHAs are completed daily, and it takes an average of 10 seconds to complete one. The Drupal CAPTCHA project5 logs close to 100 thousand uses per week, and this is just a fraction of websites (those that choose to report back).

CAPTCHAs tackle a problem head-on: they focus purely on stopping spammers. Genuine users are, for the most part, overlooked. That is to say, an assumption is made that the normal behavior of users is not affected.

It’s not true, though. The issue of genuine usability is not new. The W3C released a report back in 2005 on the inaccessibility of CAPTCHAs6, which suggested that some systems can be defeated with up to 90% accuracy. More recently (in 2009), Casey Henry looked at the effectiveness of CAPTCHAs on conversion rates and suggested a possible conversion loss of around 3%:

“Given the fact that many clients count on conversions to make money, not receiving 3.2% of those conversions could put a dent in sales. Personally, I would rather sort through a few spam conversions instead of losing out on possible income.”

— Casey Henry, CAPTCHAs’ Effect on Conversion Rates7

In 2010, a team from Stanford University released a report entitled “How Good Are Humans at Solving CAPTCHAs? A Large Scale Evaluation588” (PDF), which evaluates CAPTCHAs on the Internet’s biggest websites. Unsurprisingly, the results weren’t favourable, the most astounding being an average of 28.4 seconds to complete audio CAPTCHAs. The study also highlighted worrisome issues for non-native English speakers.

Web developers like Tim Kadlec have called for death to CAPTCHAs9, and he makes a strong argument against their use:

“Spam is not the user’s problem; it is the problem of the business that is providing the website. It is arrogant and lazy to try and push the problem onto a website’s visitors.”

— Tim Kadlec, Death To CAPTCHAs10

Completing a CAPTCHA may seem like a trivial task, but studies (like that of the W3C) have shown that that’s far from the reality. And as Kadlec mentions later in his article, what about users with visual impairments, dyslexia and other special needs? Providing an inaccessible wall doesn’t seem fair. Users are the ones who invest in and give purpose to websites.

The question is, are CAPTCHAs so unusable that they shouldn’t be used at all? Perhaps more importantly, does a usable CAPTCHA that cannot be cracked exist? If the answer is no, what is the real solution to online spam?

The World Of CAPTCHAs

The human brain is an amazing piece of work. Its ability to conceptualize, to find order in chaos and to adapt under extraordinary circumstances makes it highly useful, to say the least. For some tasks, it outshines a computer with great ease. In other tasks — mathematics, for example — it is laughably inferior.

Logic would dictate, therefore, that the most successful CAPTCHA would be:

  • A task that users excel at naturally but that computers can’t begin to comprehend,
  • A task that is incredibly quick for users to perform but arduous for computers,
  • A task that minimizes the need for additional user input,
  • A task that is relatively accessible to all users, even those with special needs (that is, the CAPTCHA should be no more difficult than general web usage and the current task demand).

One of the greatest advantages that humans have over machines is our ability to visually recognize patterns. The most popular CAPTCHA technique derives from this.

Web developers have explored many options: simple recognition tests, interactive tasks, games of Tic Tac Toe and equations11 that even mathematicians would have struggled with. We’ll explore the more sensible ideas being implemented online today.

Text Recognition

The most popular type of CAPTCHA currently used is text recognition (as seen with the reCAPTCHA124 project).

13
The reCAPTCHA project aims to stop spam and help digitize books.

reCAPTCHA was created at Carnegie Mellon University, home to the CAPTCHA pioneers and (in 2000) coiners of the term. Now run by Google, the project uses scanned text that optical character recognition (OCR) technology has failed to interpret. This, in theory, provides unbreakable CAPTCHAs, with the secondary benefit of helping to digitize books.

reCAPTCHA’s example of OCR mistakes14
reCAPTCHA’s example of failed OCR scanning.

Concerns of accessibility and usability are often voiced with regard to this type of CAPTCHA. Completely illegible CAPTCHAs are common on the Web, and asking users to perform impossible tasks can not be good for usability.

The reCAPTCHA project does make efforts to provide audio alternatives for visually impaired users, but many more text-recognition CAPTCHAs are being used without aids. As noted in the Stanford University study, audio CAPTCHAs take a long time to complete. The same study also highlighted an undesirable reliance on recognition of English-language words.

Another take on the basic text CAPTCHA was introduced15 in late 2010 by Solve Media, whose solution was to replace text with an advertisement and a related question, a move that many saw as too invasive.

Solve Media CAPTCHA16

Solve Media claims its CAPTCHAs can be solved more quickly than others. While we should be skeptical of marketing talk, there is clearly some potential, given that many global brands transcend a single language. There is potential here for marginal improvement.

While text-recognition CAPTCHAs have a few downsides (e.g. spammers could use a software that would be able to recognize text embedded in the image and try all possible combinations to “break” the anti-spam mechanism), they are undoubtedly recognizable. This fact alone can go a long way in usability decisions.

Logic Questions

Some have suggested that answering simple logic questions would be better than performing visual tasks, the idea being that the complexity of written language would be enough to confuse computers.

The TextCAPTCHA17 service has over 180 million questions in its database, including:

  • The 6th letter in “unrolled” is?
  • What is fifty-eight thousand, five hundred and seventy-four as digits?
  • Which of 3, twenty-nine, 70, 46 or 65 is the lowest?

These CAPTCHA questions are designed for the intelligence of a seven-year-old child. They are far more accessible than text and image recognition, and while this is a big advantage, it comes with a price. First, the time required to read and comprehend these questions will vary because they are unusual and unknown to users. Secondly, computers can still break these CAPTCHAs. Joel Vanhorn18 points to Wolfram Alpha as an intelligence source strong enough to crack them.

With the likes of IBM’s Watson19 recently showcasing an eerily human-like ability to process language, such technology might become mainstream quicker than we think. Instead of worrying about logic questions becoming solvable by computers, we should use this technology to analyze user-submitted content and then separate natural language from the computer-generated content that is common to spam. Services like SBlam!20 are implementing this idea.

Questions that are website-specific, such as “What is the name of this website?” and “What is the dominant color in the image above?”, might be better than general questions. The downside, of course, is that the pool of pointed questions is very small compared to the 180 million possibilities of TextCAPTCHA.

The biggest problem with logic questions is that they’re specific to a language, usually English. Providing millions of questions in every language in order to avoid alienating potential users would be a huge task. When presented with such a daunting prospect, the same question resurfaces: are CAPTCHAs the right solution?

Image Recognition

Many have experimented with photography instead of text. The benefit? No legibility issues. Services like identiPIC21 ask users to identify the object in an image. Microsoft has also researched this method through its Asirra project22.

Microsoft Asirra Project23
Microsoft’s Asirra project on image recognition.

The fact that we haven’t seen widespread adoption of image recognition CAPTCHAs indicates that it doesn’t improve usability. In fact, it jeopardizes accessibility. Visually impaired users have no chance of passing this type of CAPTCHA, and including a description or alternative text would weaken the tests.

In 2009, Google published research24 (by a team led by Rich Gossweiler, Maryam Kamvar and Shumeet Baluja) that looked at an alternative form of image CAPTCHA. The project asked users to correct the orientation of images by rotating them.

Photo rotation CAPTCHA - Google Research

A novel idea, I’m sure you’ll agree, and the research showed a preference for the ease and simplicity of this technique. Sadly, it fails the accessibility requirement (think again of the visually impaired).

Friend Recognition

One of the more interesting CAPTCHA ideas appeared in January 2011 as a result of an effort by social-networking giant Facebook. The company is currently experimenting with social authentication25 in an effort to verify account authenticity. In the words of the experiment:

“We will show you a few pictures of your friends and ask you to name the person in those photos. Hackers halfway across the world might know your password, but they don’t know who your friends are.”

— Alex Rice, Facebook, A Continued Commitment to Security26

Facebook friend recognition CAPTCHA27
A peek at Facebook’s friend recognition test.

What makes Facebook’s project slightly different than the normal CAPTCHA is that the authentication is supposed to filter out human hackers rather than machines.

There is potential for Facebook to roll this out across the web. With 600 million users and millions of websites that integrate with it, Facebook has the ability to use this social recognition CAPTCHA in a big way — and it could prove to be easier than text recognition (Orwellian privacy concerns aside for the moment).

There is one problem. Do you actually know who your friends are? The reality is that friend requests are exchanged between even the barest of acquaintances; remembering names to go with all those faces could be challenging. As intuitive and intelligent as Facebook’s idea might be, it is ultimately flawed because, as humans, we don’t follow the rules.

User Interaction

One method getting a lot of attention has users perform tasks that are impossible for virtual intelligence. They Make Apps28 features a small slider that must be dragged to the right in order to submit a form. It asks the visitor to “Show your human side; slide the cursor to the end of the line to create your account.”

29
They Make Apps uses a slider CAPTCHA.

Obviously this option is inaccessible to people with special needs. Furthermore, developing a script that is capable of moving the slider automatically to activate the “Submit” button would probably be not that difficult. A multilateral version of the slider option is used in the comments section of the Adafruit blog30. Four different sliders have to be matched to the corresponding colors in order to validate a comment and activate submission.

31
The Adafruit blog’s slider CAPTCHA.

An Over-Engineered Solution?

None of the solutions above meet all of the requirements we highlighted for a perfect CAPTCHA. Each of them impairs usability for a large segment of potential users. Even if we went so far as to assume that users generally welcomed traditional text-recognition CAPTCHAs, they would not likely welcome the other alternatives. The extra few seconds the user takes to decipher what is being asked of them negates the benefits. Too slow means not worth it.

Of the solutions available, text recognition (like reCAPTCHA) still feels like the best choice. But the question remains: why are we asking users to perform these tasks? Surely we can beat spammers at their own game by using automated systems to do the work for us. So far we have assumed that a common problem actually exists for CAPTCHAs to solve.

Despite the advances in intelligent computer systems, most spamming mechanisms are stupid. If a submission fails (because of the CAPTCHA or some other reason), the spam bot will move down its list of thousands of websites. Jeff Atwood showed this in his 2006 article “CAPTCHA Effectiveness32.” Despite all the research that goes into CAPTCHA-breaking, most spammers have no incentive to invest effort in defeating them. The sheer quantity of websites available to attack and the speed at which they can do it means that CAPTCHA-breaking is unlikely to concern many spammers.

The BBC is one of the most highly scrutinized institutions in the UK. Its requirements for accessibility are second to none, and its recent examination33 of CAPTCHAs resulted in an emphatic “No”:

“Visually impaired participants expected full accessibility from the BBC and we felt it would affect our reputation to use them. Elderly users had issues with the distorted text. The logic puzzles were found to be odd and patronising. The audio was struggled with. Overall, extremely negative feelings were expressed towards CAPTCHA technology.”

— Rowun Giles, BBC, CAPTCHA and BBC iD34

Alternative solutions exist that prevent automated submissions without resorting to CAPTCHAs and, more importantly, without user interference.

Alternatives To The CAPTCHA

CAPTCHAs, in their purest form, might realize their potential in another field. As website protectors, though, they’re far from ideal. Doing a disservice to users in an effort to combat spam doesn’t cut it on today’s web. Human-powered spam is on the rise (as is unethical link-building), and we should be implementing unobtrusive, invisible methods.

Automated and Manual Spam Detection

We touched on two detection services at the beginning of this article. Akismet351, Mollom363 and SBlam!37 all analyze user-submitted data and flag spam automatically. Mollom sometimes presents a CAPTCHA, but only when it’s unsure. But why not develop your own system that is tuned to the mechanics of your website?

Taking responsibility and removing the burden from users will improve their interactions with and impressions of your website. Manually moderating content is often a sacrifice worth making.

The Honeypot Method

In 2007, Phil Haack suggested38 a clever method of detecting bots: using a honeypot. The idea behind the honeypot method is simple: website forms would include an additional field that is hidden to users. Spam robots process and interact with raw HTML rather than render the source code and therefore would not detect that the field is hidden. If data is inserted into this “honeypot,” the website administrator could be certain that it was not done by a genuine user.

The honeypot method can be made more sophisticated by using JavaScript and data hashing. These obfuscation methods are not hack-proof, but we can assume that robots are not sophisticated enough to enter the required information.

JavaScript can be used to fill in hidden fields dynamically, which server-side validation can check for. Scratchmedia39 provides an example of this hidden field solution, along with an alternative CAPTCHA if JavaScript is disabled.

Additional timestamp and session data checks can also be used to detect automated submissions. A recent discussion40 on Stack Overflow provides many examples and ideas about this, including the implementation of Hashcash41, which is available as a WordPress plug-in42. A jQuery tutorial43 explains a similar method and includes an interesting thought:

“Thieves know to look for stickers, dogs in the yard, lights on the exterior of a home, and other signs of a well-guarded house. They’re looking for high payoff with minimal work and risk.”

— Jack Born, Safer Contact Forms Without CAPTCHAs44

The analogy suggests that, as with CAPTCHAs, the method used does not stop intruders so much as the presence of any hurdle at all. As mentioned, spammers currently have too many targets to bother searching for a back door.

Centralizing the User Base

With the rise of the social web, many websites now allow users to register and interact with one another. Publishing to a third-party website was traditionally done either by registering a full-fledged account or by submitting totally anonymously, both of which methods leave the gate open to spam. In 2008, Facebook announced45 Facebook Connect, which provides websites and their users with an integrated platform that addresses this and other concerns. Twitter followed suit in 2009 with a similar service (“Sign in with Twitter”). Both of these services can be implemented on websites relatively easily, and they eliminate the need for registration and comment forms, which are accessible to robots.

So many websites offer social-networking integration that services like Janrain46 have popped up. Janrain provides an abstracted umbrella solution to ensure that websites are accessible through any account platform.

Janrain social login at Mahalo.com47
Mahalo48 provides social log-in functionality via Janrain.

Other services, such as the commenting platform Disqus49, allow user interaction with built-in spam detection and user sign-in.

Less anonymity and more accountability make users think twice about the content they submit. It also enables human spammers to be detected and banned quickly across entire websites; remove one Facebook profile and the whole Facebook Connect network is safe from that account owner’s dastardly deeds.

Such services, of course, provoke heated debates about privacy, data protection and the like… but that’s for another article. As alternatives for preventing spam without CAPTCHAs while maintaining usability and accessibility, they have great potential.

Recording User Time Expenditure

Another rather simple method that can be implemented without annoying users is to distinguish between users and bots by measuring the time they take to fill out a contact form or compose a comment. By estimating the average time spent on a comment, one could define certain rules. For example, if a submission takes less than five seconds, which is virtually impossible for a human but just enough time for a bot to do its job, you could ask the user to try again. Jack Born’s tutorial50 on a slight variation of this concept for jQuery is worth a peek, since most users have JavaScript enabled. The whole endeavor is based on one crucial assumption: spammers prefer going after the easiest targets and will leave a website untouched if their initial attempt fails (although this can never be guaranteed).

The Perfect CAPTCHA

It would seem evident from years of use and research that CAPTCHAs are far from perfect as a solution. Remove spammers from the equation and we remove the need for CAPTCHAs entirely; this is the mentality we should be aiming for. The perfect CAPTCHA is no CAPTCHA at all.

The Rise of Humans

CAPTCHAs, by nature, function more by blocking spam than by detecting humans (which is their purpose). But they can’t do that when the spammer is not a computer. A better solution would be to remove the incentive to spam altogether. If we can reverse the trend and drive spam from being highly lucrative to being a net loss, then both automated and manual spam will become worthless.

One of the many dark arts of search engine optimization (SEO) is to artificially generate links to the website being “optimized.” Search engines consider inbound links a strong indicator of value. This can be abused, obviously, by posting worthless links on many websites (forums and comment forms are perfect for this). The SEO benefits are so worthwhile that automated spamming isn’t even required. The practice of enlisting cheap human labor is emerging. And CAPTCHAs are not designed to stop this.

We should accept the need for moderation and background detection. CAPTCHAs are a stop-gap solution at best, and are lazy and inaccessible at worst. Whether you choose to fight the good fight or simply put the interests of genuine users first, you have options.

Taking a Stand

If website owners work together to eliminate the incentives to spam, then spam will slowly wear away over time and eventually remove the need for CAPTCHAs. Is that too idealistic? Probably. In reality, we are likely to see a combination of technology and law dealing the death blow to spammers.

Google’s latest algorithm change51 has significantly demoted low-quality content farms (the effects of which are explained by Johannes Beus52). Advances such as this will ultimately remove all incentives to game the system. However, if we website owners don’t evangelize and adopt alternative solutions, then we might just wake up to a world where CAPTCHAs are worthless and our websites are unmanageable.

Understanding the alternatives (whereby spam detection is silent to users) and implementing them on our clients’ websites is a good start. It’s a positive step toward usability and conversion rates (and clients will love that!). If users comment on your website, reward them with a simple experience. This can be done in several ways:

  • Moderation wherever possible
    Disallow certain content to be posted directly to your website, or allow it through maintained and verified account management. Better yet, use a service like Facebook Connect or Disqus; you’ll make things easier for both yourself and users.
  • CAPTCHA alternatives
    Try the honeypot method or another that is invisible to users. Some could potentially be bypassed, but their presence is often enough to thwart automated efforts.
  • Client-side detection
    This can work because, while there are simple workarounds, spammers won’t waste time (for now). Keyword and mouse interactions can be used to detect genuine user input. This option shouldn’t be used on its own but can add extra assurance.
  • Server-side spam detection
    Developers should focus on server-side spam detection that monitors users and flags unusual activity. Specialist services like Akismet are affordable and proven, but bespoke systems can be tailored to the nuances of your website.
  • Social moderation
    Move toward more sophisticated features that allow this. The simple act of voting content up and down can help to push spam away or flag it for deletion.

It seems clear, considering all the pros and cons of CAPTCHA, that the future lies in a system that is invisible to normal web use. For now, using a CAPTCHA should be your last resort.

Further Reading

We express sincere gratitude to our Twitter followers62 and Facebook fans63 for their support and feedback in helping to prepare this article.

(al) (vf) (ik) (sp)

Footnotes

  1. 1 http://akismet.com/about/
  2. 2 http://wordpress.org/extend/plugins/akismet/
  3. 3 http://mollom.com/
  4. 4 http://www.google.com/recaptcha
  5. 5 http://drupal.org/project/usage/captcha
  6. 6 http://www.w3.org/TR/turingtest/
  7. 7 http://www.seomoz.org/blog/captchas-affect-on-conversion-rates
  8. 8 http://www.stanford.edu/~jurafsky/burszstein_2010_captcha.pdf
  9. 9 http://timkadlec.com/2011/01/death-to-captchas/
  10. 10 http://timkadlec.com/2011/01/death-to-captchas/
  11. 11 http://www.flickr.com/photos/ceejayoz/2674227920/
  12. 12 http://www.google.com/recaptcha
  13. 13 http://www.google.com/recaptcha
  14. 14 http://www.google.com/recaptcha/learnmore
  15. 15 http://www.solvemedia.com/index_ss2.html
  16. 16 http://www.solvemedia.com/index_ss2.html
  17. 17 http://textcaptcha.com/
  18. 18 http://joelvanhorn.com/2010/11/10/using-wolframalpha-to-hack-text-captcha/
  19. 19 http://www.engadget.com/2011/01/13/ibms-watson-supercomputer-destroys-all-humans-in-jeopardy-pract/
  20. 20 http://sblam.com/en.html
  21. 21 http://identipic.com/
  22. 22 http://research.microsoft.com/en-us/um/redmond/projects/asirra/
  23. 23 http://research.microsoft.com/en-us/um/redmond/projects/asirra/
  24. 24 http://googleresearch.blogspot.com/2009/04/socially-adjusted-captchas.html
  25. 25 http://blog.facebook.com/blog.php?post=486790652130
  26. 26 http://blog.facebook.com/blog.php?post=486790652130
  27. 27 http://blog.facebook.com/blog.php?post=486790652130
  28. 28 http://theymakeapps.com/users/add
  29. 29 http://theymakeapps.com/users/add
  30. 30 http://www.adafruit.com/blog/2011/03/01/some-adafruit-website-updates/
  31. 31 http://www.adafruit.com/blog/2011/03/01/some-adafruit-website-updates/
  32. 32 http://www.codinghorror.com/blog/2006/10/captcha-effectiveness.html
  33. 33 http://www.bbc.co.uk/blogs/bbcinternet/2010/10/captcha_and_bbc_id.html
  34. 34 http://www.bbc.co.uk/blogs/bbcinternet/2010/10/captcha_and_bbc_id.html
  35. 35 http://akismet.com/about/
  36. 36 http://mollom.com/
  37. 37 http://sblam.com/en.html
  38. 38 http://haacked.com/archive/2007/09/11/honeypot-captcha.aspx
  39. 39 http://www.webdesignfromscratch.com/javascript/human-form-validation-check-trick/
  40. 40 http://stackoverflow.com/questions/4683117/alternative-to-annoying-captcha-in-forms-how-to-smell-the-difference-between-a-h
  41. 41 http://en.wikipedia.org/wiki/Hashcash
  42. 42 http://wordpress-plugins.feifei.us/hashcash/
  43. 43 http://docs.jquery.com/Tutorials:Safer_Contact_Forms_Without_CAPTCHAs
  44. 44 http://docs.jquery.com/Tutorials:Safer_Contact_Forms_Without_CAPTCHAs
  45. 45 http://developers.facebook.com/blog/post/108/
  46. 46 http://www.janrain.com/products/engage
  47. 47 http://www.mahalo.com/login
  48. 48 http://www.mahalo.com
  49. 49 http://disqus.com
  50. 50 http://docs.jquery.com/Tutorials:Safer_Contact_Forms_Without_CAPTCHAs
  51. 51 http://googleblog.blogspot.com/2011/02/finding-more-high-quality-websites-in.html
  52. 52 http://www.sistrix.com/blog/985-google-farmer-update-quest-for-quality.html
  53. 53 http://polldaddy.com/poll/4657952/
  54. 54 http://polldaddy.com/features-surveys/
  55. 55 http://www2.parc.com/istl/projects/captcha/history.htm
  56. 56 http://www.w3.org/TR/turingtest/
  57. 57 http://www.seomoz.org/blog/captchas-affect-on-conversion-rates
  58. 58 http://www.stanford.edu/~jurafsky/burszstein_2010_captcha.pdf
  59. 59 http://www.codinghorror.com/blog/2008/03/captcha-is-dead-long-live-captcha.html
  60. 60 http://www.readwriteweb.com/archives/the_state_of_web_spam_human-posted_spam_is_on_the.php
  61. 61 http://stackoverflow.com/questions/450835/how-do-you-stop-scripters-from-slamming-your-website-hundreds-of-times-a-second
  62. 62 http://twitter.com/#!/search?q=%23smcaptcha
  63. 63 http://www.facebook.com/smashmag/posts/10150112377942490

↑ Back to top Tweet itShare on Facebook

David Bushell is a website designer and front-end developer working at Browser Creative, London. He blogs regularly at dbushell.com and xheight, and shares inspiration and web design related interests at Design Heroes. You can also follow him on Twitter.

Advertisement
  1. 1

    Nice article!
    I firmly believe in “the best CAPTCHA is no CAPTCHA”. Instead of building gates and only letting fewer get through, why not let everyone through and then concentrate on building scripts and checkups along the way that discover malicious behavior from a user? This way the website will detect both traditional bots as well as paid human spammers (which are increasing as technology reaches more developing areas). One way could be to build indicators on a site that triggers flags when suspicious behavior occurs after registration, once there are enough flags triggered for a certain type of user, the user is frozen/deleted. This may be more time consuming to build and will require maintenance, but as long as spammers keep changing tactics, site owners need to do so as well. At the end of the day it will the most user friendly solution and the most non-noticeable.

    1
  2. 102
  3. 203

    I always liked motion captcha as a solution. Not very accessible, agreed. But no problems with color (blind), legibility, etc…
    http://www.josscrowcroft.com/demos/motioncaptcha/

    -1
  4. 304

    I am not hearing impaired, disabled or the like but half the time I can neither visually make out nor hear the captcha resulting in me simply leaving the site; not good for business!

    0
  5. 405

    A very interesting article.
    But I have a question: and what about multilple captchas?
    As an example: honeypot + 5 second count + choose 1 out of X very simple user interaction (this grants no accessibility problems)?

    Users only “feel” one of them, and very simple, but we made 3 captchas.

    0
  6. 506

    Hi,
    Your article was absolutely fantastic, and browsing a website the other day I found the perfect solution. It is a CAPTCHA that uses a small logic game to verify your result, and has platform specific options. Forgot the name though!!

    -2
  7. 607

    I wrote this captcha implementation myself, and I’m just trying to share it with other developers, because it’s so much simpler than many of the other implementations out there.

    http://meta64.com/axis/fb/?id=13655

    0
  8. 708

    With the extent of automated software, day by day traditional Captcha solutions are becoming less effective. For me, the best way to differ a Robot and Human is asking answer of any Question which will make the user thinking about it like “When this site was established”? Regards, Russel from seo company sydney

    0
  9. 809

    Thank you! I’ve been researching what type of captcha to install on several WordPress sites that I moderate. Akismet works well enough for comments, but there are questionable user registrations that appear almost daily. I was considering SweetCaptcha, but had concerns about the complexity of the instructions and the ability of some users to follow through. I also don’t like the implications of certain captcha text instructions, which ask you to “verify your real existence” or “prove you are human.” After reading this, I am much more inclined to go with a honeypot solution.

    0
  10. 910

    Good alternative to Captcha:
    http://areyouahuman.com/

    0
    • 1011

      Seems like a confusing solution. Captchas should be as unobstrusive as possible, otherwise you lose visitors/clients. If a user has to figure out what she has to do to pass a test she potentially does not even understand why she has to pass it, user experience can get very messy.

      0
  11. 1112

    As long as there is valuable content, there will be spam and people who want to do harm… it’s kind of a vicious circle really ;).

    0
  12. 1213

    Excellent article!

    I’ve noticed that a lot of articles here on SM have seem to present some ideas without a lot of thought behind them or any real substance in the article. This one prompted a discussion at my office and is likely to result in my development team implementing a honeypot solution instead of using reCAPTHCA as we do now.

    Great job David.

    0
  13. 1314

    Alternative for the new age :)
    http://www.lirullu.com

    What did you think?

    0
  14. 1415

    I’m surprised that http://areyouahuman.com got no mention! Those guys are doing great work.

    0
  15. 1516

    Great work! This is the type of info that are supposed to be shared around the web. Shame on Google for now not positioning this publish higher! Come on over and visit my site . Thank you =)

    0
  16. 1617

    An alternative to Captcha I prefer is keypic. It’s free of charge and quite efficient. I’d recommend it to anyone who wants a strong protection from spam, without any action needed from the site users. Try and search the web for keypic.

    0
  17. 1718

    I use a combination of Mollom + my own implementations of honeypot and timestamp methods for my own (Drupal powered) website. The numbers are great, Mollom stops most of the attempts, and the other two methods get most of the remaining. From time to time, about once a week, someone gets into, and I have to moderate it, but the effectiveness of these three methods combined is over 99%.

    Highly recommended.

    My honeypot and Timestamp methods are for Drupal 6, but they can be easily adapted to Drupal 7, in case someone wants to try:
    http://www.isegura.es/blog/stop-spam-your-site-being-invisible-honeytrap-drupal-comments-form
    http://www.isegura.es/blog/stop-spam-your-site-being-slow-flood-control-method-drupal

    0
  18. 1819

    Either if this article is dated 2011 , I’ve found it really interesting.

    I think that one additional field to check on the server side (php in my case) just before to submit a database query or email or php function is really effective and transparent.

    But one discussion on another website, made arise one big deal: hidden fields can be tricky if someone uses some kind of website reader that doesn’t processes CSS.

    If you turn off CSS the hidden field is not hidden anymore.

    So I did my version, a visible field labeled “prohibited field”, after the label is the input text, AFTER the input text is another label “<– don't fill this". The input text form POST name could be "subject2" or "emailme" or some other common label name ( for paranoid bots)

    Who fills that field, was warned, and will exit the php execution with an echo that prints "You were warned! You have filled out the prohibited field" + a link to go back to the contact/registration/comments page.
    … hey man, you were warned, next turn skip the prohibited field … <— this thought is for those users that will fill that input text … I'm pretty sure someone real human will, but who care?

    This method IMO sounds interesting also because you can optionally write the label with an image and (anyway, image or not image) it will avoid also those bots that are able to distinguish/recognize hidden fields :-)

    (edit.. HEY! Great idea and alternative!!! An image that in the HTML simulates the CAPTCHA system, a real image but a CAPTCHA fake ;-) , as a matter of fact we will use that image as a label that says "this is an anti-captcha don’t waste your time and leave this field untouched" … this text is quite long, but gives the idea)

    What do you think about this? "Don't write here" … a real anti-captcha

    NOTE: labels can be interpreted by smart BOTs, so
    – you can set more than one (a couple) of anti-captcha fields
    – on top of the page, where you may say that fields marked with a “*” are mandatory fields, you can write that those marked with a “#” are prohibited fields and must be left empty and warn that filling them will fail the form submit.

    Thank you for your opinion

    Robert – Verona – Italy

    0
    • 1920

      P.S. The solution I’ve written here above, is quite recent.
      I had to deal with an effective but simple and easy to implement anti-spam system to apply into an open source (and unfortunately, abandoned) ticket system named lynxHD.

      In these days it started a spam campaign and they were created a lot of tickets.

      What has been magic (I’m not an expert) for me, has been that I’ve removed the submit button from the page, I’ve commented it out… but the tickets were created the same… ?!?!?!?!?!?!?!?!

      Thank to any of you that may explain how that was possible. It was like the BOT was able to generate the submit using some kind of event.

      P.S. Before to implement the visible field trick said above (and that for the moment it works), I went the long way.
      – I used a javascript code that was measuring the time elapsed from the page load and the onsubmit event. This anti-spam It is still active, it was working (I’ve tested and measured it). If you fill out the form in less than 7 seconds, the page reloads itself, like pressing F5 on the browser, the time counter resets and you must start over… but no way, new spam-tickets were creating too fast.
      I’ve increased the 7 seconds up to 1 minute, but they were creating 4 o 5 tickets per minute ?!!?!?!?
      —> maybe the BOT had the javascript disabled <— ahahah, hey now I got it :-((

      At that point I've removed the submit button to discover that the tickets were anyway creating?!?!?!?

      So the BOT was "using" some leak of "something" and was submitting the form skipping, or better to say, not affected by the javascript elapsed time check and skipping the absence of the submit button.

      That has been the moment when I’ve realized that the field to be left empty was the right solution, at least for this kind of BOT.

      But please, if any of you may explain to me why I was getting crazy with that sort of magic spam, thank you in advance.

      Robert

      P.S. the javascript issue maybe the key… the BOTs are not browsers.. this could be the explanation, so any elapsed time measuring method before the submit event, must be done on the server side.
      Found right now http://forums.phpfreaks.com/topic/265035-calculating-elapsed-time-between-pageload-and-page-submission/ it could help those of you that want to deal with this solution.
      Mixing the empty field with time elapsed measurement, could lead to an effective and transparent antispam and anti-captcha solution.

      0
  19. 2021

    I built https://verscaptcha.com, IT’S FREE, FAST, and EASY TO ENTER!

    0
  20. 2122

    The best captcha you can use, is not to use it!

    try keypic.com

    0
  21. 2223

    I quite like the idea of incorporating a standard text CAPTCHA as a honeypot into forms along with the methods from this article http://nedbatchelder.com/text/stopbots.html

    I’ll definitely be implementing this next year on all our forms.

    0

↑ Back to top