> Like elf, I'd prefer to prioritize a clean representation of all
> datatypes over absolute compat with existing stuff. It seems a good
> balance can be struck quite easily.

In practice, this means that if you are writing a throwaway script and
know that the data files you have don't contain any weird strings or
symbols, you can use reader in the data using the native reader of a
Lisp implementation and due to the wide agreement on how "normal"
symbols and strings are written, should work out ok. But for programs
that have to read and write unknown data, to ensure correctness, you
should use a dedicated library built for POSE. I find this a reasonable
compromise.