Encodings 2.0

[ljedrz]

Encodings 2.0

[billpmurphy]

I like this design! I think it's possible to keep the conversions pretty ergonomic across encodings; it gets more tricky when different list types are involved. Consider the following cases:

• 1.to_church()
• 1.to_scott()
• Some(1).to_church()
• Some(1).to_scott() //same as above except for the numeral
• vec![1, 2, 3].to_church()
• vec![1, 2, 3].to_scott() //same as above except for the numeral
• Some(false).to_church()
• Some(false).to_scott() //same as above

All of this seems possible with the design you outlined (there will be a little duplication of impl Traits for each of the encoding traits to make this work, but that doesn't seem too bad). However, I'm not sure how to make this design work when there are also different types of e.g. lists that are shared across encodings.

(I'll be away for the next 3 days, so I may not respond to this thread right away!)

[ljedrz]

That was actually much easier than I expected. I currently finished refactoring with a single list; it looks pretty good, so I'll merge it.

[ljedrz]

After some consideration I decided against adding different representations of lists; it could easily become a mess and the current one is the one most widely used and imo practical.