Revisit usage of filename to determine validity of the data source

[theoaraujo]

Filename as the driver for determining the source data type and if it is valid may become an issue for other platforms, or for other languages.

[damian0604]

Example: youtube takeout uses language-sensitive filenames (kijkgeschiedenis.json vs viewinghistory.json).
But even if we simply support all conceivable languages, then the problem is that respondents still may rename files (or when they download something twice it may be called ".. (1)" or "Kopie van" in various languages or whatever.
I see two possible approaches:

• try to determine file type based on content rather than filename
• give an informative error message to respondent ("I didn't recognize this file name. If you are trying to upload [Facebook|Youtube|whatever] data, then I'd expect that your file is called "..."] Are you sure you selected the right file?