Hello ISBN-13

[pne]

Starting tomorrow (2007-01-01), ISBNs (International Standard Book Numbers) will be 13 digits long instead of 10. Basically, they'll add "978" to the beginning and calculate the check digit differently. This way, they'll fit into the EAN (International Article Number) scheme. This also prepares for the event of ISBN shortages for certain prefixes; these can then be issued ISBN-13s starting with 979-.

Also, I hope I manage to sleep despite the fireworks here.

Comments

[sparkofcreation.livejournal.com]

*confused* I'm pretty sure ISBNs have come in both 10- and 13-digit varieties for quite a while now.

[pne.livejournal.com]

Good point.

I think that was partly in preparation for this switch, so perhaps a better wording would have been "ISBNs will only come in 13-digit versions from now on".

Also, I think we'll start seeing 13-digit ISBNs advertised more often; previously, barcodes on books had displayed the 13-digit version in the barcode and in digits beneath the barcode, but I very rarely saw them in "plain text" (e.g. when citing a book, or even in the space above the barcode where it typically said "ISBN: x-xxx-xxxxx-x" with the 10-digit version only.

[spamsink.livejournal.com]

Thank God. The check digit confusion, let alone the ill-thought check digit X in the original scheme will go away.

[pne.livejournal.com]

Which confusion?

Also, I've heard that the EAN check digit method doesn't catch all transposition errors the way the ISBN-10 check digit scheme was supposed to.

[spamsink.livejournal.com]

The ISBN check digit is not (never?) equal to the EAN-13 check digit. Converting one to another manually was non-trivial. Occasional presence of an X in a supposedly all-numeric code is confusing for a person not familiar with the system. They had better encode the check digit with A to K.

Naturally, a modulo 10 check digit will catch fewer errors than a modulo 11 check digit.

[pne.livejournal.com]

The ISBN check digit is not (never?) equal to the EAN-13 check digit.

Not in general. I think I've seen cases where the check digit happened to be the same.

Converting one to another manually was non-trivial.

Yes -- you have to generate the check digit from scratch, since the algorithms are different.

Naturally, a modulo 10 check digit will catch fewer errors than a modulo 11 check digit.

I think it's not just the 10-vs-11, but also the way different digits are weighted when calculating the check digit. Apparently, the fact that 11 is prime is also significant.

But that's beyond my depth; it's just what I read about so I'm not sure how significant each factor is.