That suggests that the bank made a 'mistake' when it recorded the name in its system as well as it could. I don't think that should count as a mistake. The information in the bank's system is as good as it could be, so there is nothing to rectify.
It feels weird to me when privacy legislation turns out to require supporting UTF-8. I think something in the legal process went wrong here.
In the specific case of this bank, like everyone else, they were expected to update systems unable to comply with the legislation within the grace period and yet it seems that they were unwilling or unable to update or replace a system that is incapable of achieving data integrity in a matter as basic as the name of a customer.
I suspect this will be fixed by storing the correct name in an external system and using an ID number or similar to refer to the customer on EBCDIC systems.
No. It requires you to store data correctly. And in the case of a bank storing data incorrectly could have potential ramifications (think two different people, one with diacritics and without).
The law doesn't care whether you use UTF-8 or a manually written translation table, or a 15th-century printing press
Surely this situation wouldn't cause any issues though right? If they are relying on names as unique identifiers, they've got far bigger problems than a users name being spelled incorrectly.
I find this whole ordeal delightful, and applaud the intention of the GDPR and the ways the courts upheld it in this case.