This is a great idea. I remember reading a blog entitled something like "ascii-separated values" as a replacement for CSV files. So I took another look at the ascii table...<p>...and when I did, I was surprised to find not just field, record, group and file separators, but also values for packet-based networking, values for network handshaking, support for synchronous and asynchronous data transmission....even support for sessions, heartbeats, etc.<p>... basically in an embryonic form (or a fat-free, pre-crufty form, depending on your perspective) everything you needed to do anything from creating a format to store relational data on a disk to packet-switch networking. Come to think of it teletypes themselves were networked nodes; in retrospect it shouldn't be surprising that ASCII would have rich support for the kinds of devices it was used on.<p>I really hope this proposal has legs. CSV files are in desperate need of replacing. They are ambiguous, insecure, and non-standardized. No matter how careful you are, you'll just get an endless stream of obscure bug reports, where somebody has escaped something weirdly, or forgot to escape something, or switches delimiting characters midstream, etc etc.<p>And from the beginning, they were a completely unnecessary hack and a self-inflicted wound: Dedicated characters does an end run around all such potential problems, and we've had them since the 60's.<p>And what other things might we fix by taking another look at the legacy we inherited from our ancestors?