It’s 2014 and I’m still using line-oriented text formats and sed to do ‘Big Data’. And I’m stilling having to tell the database the difference between an zero-length string and a NULLed int or double.
In spite of fifty years of computer science and software engineering research, I’m still having to do completely moronic shit to get data from one place to another.
Why can’t we have properly self-describing data in a tightly-compressed binary file with a clear and simple way to turn that back into a simple text format so I can read it? Hell, I’d settle with something like JSON (with an attached schema) gzipped.