Re: Allowing accented characters in the database

Agreed


On Tue, 29 Aug, 2023 at 9:14 AM, Patrik Schindler <poc@xxxxxxxxxx> wrote:


To: midrange systems technical discussion

Hello Vern,

Am 29.08.2023 um 14:39 schrieb VERNON HAMBERG Owner via MIDRANGE-L <midrange-l@xxxxxxxxxxxxxxxxxx<mailto:midrange-l@xxxxxxxxxxxxxxxxxx>>:

You ask about why mess with UTF-8 - I think one reason is its nearly ubiquitous use in json and xml.

Yes, but a database doesn't need to save plain json and xml with a field, yes? As such, conversion from less "problematic" UCS-2 and UTF-16 types to whatever frontend you're using can be done automatically by either the database or the OS. On i, the borderline is somewhat blurry. ;-)

Emojis such as you showed do not convert into CCSID 37 (we are in the US).

Neither do they to 237 in here. It was an example to prove Barbara's point. If you need to have conversion applied, the source data in UTF-8 must only contain those characters which can be actually mapped to the destination CCS.

At any rate, UTF-8 is something we have to mess with here.

Undoubtedly. I'm just trying to point out that the "where to mess with it" might also be something worth to be considered.

:wq! PoC