Talk:Validate International Securities Identification Number
Dup?
Is this a duplicate of Luhn_test_of_credit_card_numbers?
If not, what is the task? --Rdm (talk) 07:22, 26 February 2015 (UTC)
Thanks Rdm. I should have spotted the connection with Luhn. The difference is that ISINs can contain alphabetic characters, which must be translated to digits before a Luhn checksum is calculated. CUSIP, the North American stock codes, are shorter but otherwise use the same algorithm as ISINs. I couldn't see a Rosetta page for them.
There is another page on Rosetta for SEDOLs, which use the same letter-to-number technique, but thereafter use a different checksum algorithm, not Luhn.
So I think there is a point in having an ISIN page. What do you think?
--TheWombat (talk) 00:05, 27 February 2015 (UTC)
- Any of these sound like they would work. Another possibility might be to mark pages with something like Category: Checksums (very broad) or Category: Digit Checksums, or something like that.
- That said, note that in my previous reading I did not pick up that letters were being included in the checksum, using a "base 36" sort of mechanism.
- Anyways, I do not have any real strong opinions on this topic. I guess just proceed how you like and see if anyone else weighs in? --Rdm (talk) 09:29, 27 February 2015 (UTC)
Adding an example
I suggest adding an ISIN ending in 0, as it may show a bug in the implementation. For instance, I had the (bad) idea to replace x==(-y)%10 with 10-x==y%10 in the Python implementation (where x and y are expressions in the Python code, and x is the last digit). This transformation is of course wrong, but it fails only when x=0. I have found several valid ISIN ending in 0, here is one: FR0000988040 (from HSBC I think).