I mean, give me a break, what is the purpose of having routing numbers and account numbers on checks that hardly anybody can read clearly?
Surely there must be computer technology that could read “real”, human-being-legible numbers at the bottom of checks?
Those mangled numbers are part of a system called MICR, and were created to be machine readable by a magnetic ink reader back in the day.
http://www.asapchecks.com/micr/micr.htm (somewhat relevant link, first one i found googling)
We do have computer technology to make ‘human-legible’ numbers on checks that can be read by computers. I am not sure why MICR is still used, but I have two theories:
A better link http://www.asapchecks.com/micr/micr-faq.htm
They’re legible for most, aren’t they? I can read them clearly, and always could 
They look like the numbers they’re meant to represent … and the stylizations aren’t that extreme.
MICR is still used instead of OCR because it’s much faster and more accurate. Large banks often have to process as many as 20-30 million checks per day or more. Speed and accuracy are essential for them.
MICR dates back to 1955, so just being able to quickly read checks with a computer was pretty impressive.
The design of the font is actually fascinating. It’s sort of trying to act like a barcode but with legible number forms. That’s why certain parts of the characters are wider than they need to be. I wish I could find a good description of it on the web – I read about it years ago in a book and am fuzzy on the details.
But if I understand DMark correctly, he’s proposing a standard set of characters that just look more like the ones we’re used to. The OCR wouldn’t have to be able to read a wide range of characters, just the specific set used.
And doesn’t the USPS uses OCR to read zip codes? Recognizing numbers isn’t that hard.
Kind of. The shapes of the characters are designed so that each will have a unique amount of magnetic ink comprising it. Thus each character will have a unique magnetic signature which can be quickly and accurately read by a sensor very much like a tape head. Each MICR character generates a singal whose strength is proportional to the amount of ink that makes it up. This is why it’s so much faster and more reliable than OCR; rather than needing to process a complex shape to determine visually what character is being read, MICR uses a simple, unambiguous signal level to do the job.
Couple of other links:
http://www.elfring.com/what.htm (talks about specs for MICR)
http://www.asapchecks.com/micr/a-opportunitiesandperils.htm (ditto)
http://www.sri.com/about/timeline/erma-micr.html (discussion of development of MICR and ERMA)
http://www.micr-fonts.com/MICRfont/micr-fonts.html (shows the font set).
Not exactly. They are attempting to incorporate OCR into the mail sorting stream, but it’s not very accurate, and mail that can’t be proprly read by OCR gets put through the old sorting system. It’s getting better, though. As far as ZIP codes, the USPS has had a bar code schem in place for years; its those funny-looking lines you sometimes see near the bottom edge of envelopes. That is read optically, but it’s much faster and more reliable than OCR would be for the same purpose. I expect that one day, OCR will be as fast an accurate as a human reader, if not more so, but for now, the technology still needs work.
You’re not the only one. I’m not exactly sure what’s so difficult about deciphering check routing numbers, either. They look like numbers to me.
I had a hard time reading them when I started working retail, but you get used to them pretty quickly.
The MICR font is similar to (but not exactly like) Westminster font.
I have never seen Wingdings font used on checks.
My guess would be it is used to make it more difficult for people to determine a routing number or account number just by reading the bottom of the check. I have seen websites that display Wingdings in place of letters and numbers as you type out a password in order to protect your password from being easily revealed. (Why this is done instead of the usual string of asterisks, I don’t know).
As far as “protection” of account numbers, using Wingdings in place of numbers probably became obsolete in 1956. All someone would need is a “Wingdings to numbers” chart. In these days of people cracking computer security, the Wingdings system of “security” is paleolithic.
The Soviets had a workaround for the problem – you were required to write number in certain ways, and no way else, though they did provide a square that could be used to mark the numbers properly.
I found that a fascinating difference: at the time (before the Soviet Union fell), the US was working on ways to try to read handwriting, while the Soviets told people how to write things. The difference between the two systems in a nutshell. 
In Soviet Union, checks identify you.
IIRC, the USPS does have a very high success rate reading zip codes, even when they’re handwritten. They throw a lot of money at the problem. The thing is that they only read it once. Then they print a bar code (if the sender didn’t print one), because that is much easier to read with much cheaper equipment.
Those numbers are for the machines to read, not the check writer nor the payee.
Some friends wrote commercial software that could read anything anyone could write!
They are specifically intended to be BOTH machine- and human-readable. Otherwise, they wouldn’t need to look like numbers at all.
Doubtful. OCR software is pretty good nowadays, but many writing styles such as cursive script and highly stylized hands, particularly if written carelessly, are still beyond the ability of even the best OCR program. CAPTCHA, those graphic characters that many websites make you enter to prove you’re a human and not a bot take advantage of this fact. People can read them just fine, but OCR chokes. Or, it’s supposed to. Some poorly-designed CAPTCHA character sets can be read by modern sophisticated OCR software.
Heck, some cursive writing is hard for people to read…
The IRS for some time included in tax booklets examples of how they wanted the numbers drawn. I crossed my 7’s anyway.