Can you explain why you oppose CSV? And by the way CSV is a standardized format with its own RFC; it is definitely not home grown. There are mature parsers being able to handle commas and quotes and linebreaks and other special characters in CSVs.
That same RFC explicitly notes that CSV is entirely ad-hoc and homegrown, and that it (the RFC) is an attempt to clean up the existing mess.
Beyond that, hand-editing a CSV file (because eventually you have to do that) with those special characters in it is a huge pain. It remains better than JSON, etc., because being record-based is automatically better than not for this stuff, but it's not a good option. I'm well aware that CSV is theoretically standardized; I have written standard-compliant parsers and writers. (It is awful.) And then, once I had painstakingly written that writer to spec, the next guy's--no, not Excel--trashed my data because CSV has a spec that nobody cares about.
The other day I had to fix a bug in our CSV importer; it turns out that when you install Excel, it changes the mimetype of CSV files from 'text/csv' to 'application/vnd-ms.excel' system wide. Wow! These sorts of shenanigans are never ending in the futile endeavour to support CSV.