Every time I have to do any major work with CSVs, I re-lament this exact thing.
I think the only way this could ever become more widespread is to fix all the open source tooling so that it's eventually just supported everywhere - then keep evangelizing for... ~30 yrs.
Probably you should also register a new mime type and extension and make it a new thing - don't overload .CSV any further - but make the same tooling support it.
If time could be turned back, a good idea would be to make CSV mean CSV. Not semicolon separated values, not any other thing separated values, but only comma separated values. To not overload the name in the first place.
And I would rename the format to SSV and make semicolon the separator. Comma is a terrible choice, because it's used as the decimal separator in many countries around the world.
Every time I have to do any major work with CSVs, I re-lament this exact thing.
I think the only way this could ever become more widespread is to fix all the open source tooling so that it's eventually just supported everywhere - then keep evangelizing for... ~30 yrs.
Probably you should also register a new mime type and extension and make it a new thing - don't overload .CSV any further - but make the same tooling support it.