I like to imagine a world, a worse one, where programming languages were localised. This might initially have been a versioning nightmare, with different compiler binaries for each localisation. Later, it could become standard practice to ship a single compiler containing all supported localisations, the correct one being chosen from either system language, a project-wide setting, a preprocessor flag in each file, or some combination of these. Everyone would have to learn a little Polish, and "Source Code Translator" would be a profession.
I think if this were the case, we would have quickly moved on from storing source code as text, and begun storing it as ASTs which could be 'viewed' in any localised version of the programming language. This may have had wider benefits than just reading source code in your preferred language.
You can serialize ASTs very easily as S-expressions, or slightly fancier versions thereof. That's still "text" and quite easily readable by humans, but somewhat less oriented to direct editing.
I would love for the block delimiters to be also a preference, like the language in that world. I'd edit every language with Python style blocks.
You mean excel?
Hah!
Reminds me of the first time I tried Excel in English (might have been Google Spreadsheet) and the searchv function was nowhere to be found, not knowing it was actually called vlookup ("buscarv" in Spanish).