For the text variant, it just parses the year field as an integer, so leading zeroes would generate the same year value as if you'd omitted them.

Nice, that's the parsimonious way to handle it.

I'm really digging this system in general. I hope it gets more traction.

On the other hand, valid data may now be unreasonably long. What if someone gives you a string that starts with 10^12 leading zeros?

This is where security and limits comes in:

https://github.com/kstenerud/concise-encoding/blob/master/ce...