Registering for Yahoo

As far as I can remember I’ve never used Yahoo for anything. I AM the Google generation. Walking across the road, joining the darker side, was a scary thought, but when I saw the short registering procedure I calmed down.

The way they have set up their registering procedure is just about as good as it could get.

As you can see everything is where you expect it to be, divided into chunks like “Tell us about yourself…” and “Select an ID and password”. Even the less web savvy users might think “voilà this is done in 30 seconds”. And they would be right if it wasn’t for that #¤%)/(&=#087AAARGH”#(¤¤//!!…

Freaking spam protection!

I had used only one or two minutes entering my personal information, even choosing a decent account name. Everything went my way just until now. I saw the image representation of something that might look like text. Skewed, distorted, blurred.

Now you say, where have you been? Never seen that kind of spam protection before? Yes I have. Several times. And usually I don’t even think about it. This time I had plenty of time to think about. I just couldn’t get it right.

For example, what letters to the image above represent? If you guessed z7neG6Bz you would have been wrong. I never had a second chance, because now a new image was presented. Impossible to read. The third was even worse. Now at my fourth try I was sure I couldn’t be wrong. I probably wasn’t, but because I had typed in some wrong characters just before Yahoo had thrashed my password. So instead of accepting my, assumingly correct interpretation of the letters, I had re-type my password and then interpret yet another image.

For a moment let’s leave my Yahoo story and have a look at some other examples.

Examples of good and bad image verification

 Example 2, 3. – http://google.com/

These examples are pretty similar to my Yahoo experience. The letters are skewed and the tracking (space between each letter) is set to about -150. It is, at best, very hard to guess the letters.


 Example 1. – http://flatpanels.dk/

In this example from http://flatpanels.dk the letters are not skewed and the tracking is increased to make it easier to separate one letter from another. Instead a hefty grain filter has been applied. This makes it hard for machines to decode, but users without serious visual disabilities should not have any problems. The letters placement is vertically inconsistent, again this should make the bots have a hard time, but users can easily overcome this hurtle.

Conclusion

We see that Google and Yahoo uses the same hard-to-breakdown image-verification method. The image is very hard to interpret for both machines and real users. They have to do something about those nasty abusive bots, but when it happens at the cost of usability the battle against the machines has been lost.

Back to the Yahoo story. Yes, I did succeed after several tries. Not all the letters were hard to guess, but a image verification method is not good enough if only some of the images produced are readable. Back to the drawing board Yang and Page!

Related Posts:

• Google Chrom(ium)
• Testing Password Strength Using Google
• Teamwork Made Easy – GoogleDocs
• Google Chrome – faster and more screen estate
• SEO for Google