Cracking the Google Bar Code

It's been more than 20 years since I sat down one day with a pile of envelopes and cracked the Zone Improvement Plan bar code. Excited, I then turned my hand to cracking the Universal Product Code, but soon gave up in frustration. The series of thick and thin bars seemed to carry no information corresponding to the digits beneath them.

Well, today on the 57th anniversary of the official recognition of the bar code, I finally did it. The key turned out to be that the thick and thin spaces between the bars also carry information, and, even more importantly, there are extraneous digits at the beginning, middle, and end of the code. Furthermore, there are always an odd number of digits (in order to begin and end the series with a bar rather than a space), so in the last half of the code the numerals begin with a space rather than a bar.

This all came about because today the GOOGLE logo is a bar code. I haven't been able to crack it yet, but I've gone to all the work of teasing out the digits. I don't know yet whether they are base-2 or base-4 binary, but nothing I've tried so far has worked. I welcome suggestions!

Here are the digits:

2112142113131341111341111221142211141122141221142331112

There are 55 digits, which to me probably signifies that 211 and 112 are the beginning and ending markers (which is more ZIP code than UPC code), and that the central digit, 1, marks the middle. That leaves 12 numerals of 4 digits each. They can't be read as UPC code numerals, which always consist of 4-digit sequences that add up to seven. How that is supposed to spell Google, I don't know. Maybe instead it's a date: 10-07-1952/2009.

Or, the end markers could just be the digit 2 and there are 13 numerals. Please help!

Added on October 8th:
Well, I was on the right track. The end marker '2' is only there to allow the code to end in a bar; barcodes for that reason always have to have an extra digit as long as the characters contain an even number of digits. This particular code is rather consistent, in that all other characters contain six digits. So I was a bit short on the end-markers; the symmetry of their outside digits was only coincidental.

But the extraneous 'g' truly is a mystery. In the science of textual criticism, this would be referred to as dittography, but of a rather unusual kind.

The entire barcode translates as follows:
-begin sequence-Googleg-end sequence-.

I don't see that it would have been all that hard to decode only 49 characters instead of 55, especially given that the key was freely available online. And the notion of a checksum doesn't seem to fit, as it is only used with numeric codes like UPC.

I'll check back in if anyone can solve the mystery. And yes, a malicious website in China (forexbids.cn) has hacked into many of the sites discussing this mystery!