In asking for help with an application, I wondered aloud in text whether (or perhaps assumed) it were possible to have:
trait DisplayType1 { ... }
trait DisplayType2 { ... }
// ...
and then somehow register that {:T1} and {:T2} should be handled by those traits.
cf. “The current mapping of types to traits”
In my case, I wanted for a CLI to have a human-readable ‘pretty’ output (OK, Display works fine here) as well as JSON output (that should be different to Debug) and perhaps others in the future (e.g. one might imagine YAML, none, coloured, not).
It was suggested that I bring this up here, so here I am. (Note I’m not here to repeat my ‘what should I do’ question, just to identify a use case for something like this in the language.)