Consider using string without spaces for braille code metadata

IMO using a string without spaces for the braille code metadata would make more sense.

- It is more intuitive: people associate unique identifiers with things without spaces: [UUID](https://en.wikipedia.org/wiki/Universally_unique_identifier), ISBN, product serial numbers, ...
- We would preserve compatibility with other standards, such as [URN](https://en.wikipedia.org/wiki/Uniform_Resource_Name), and BCP 47 language tags. For instance, this would keep the possibility open to integrate the braille code in a language tag, using the "T" or "X" extension (or maybe even the variant subtag), as was [discussed here](https://github.com/daisy/ebraille/issues/25).