newtype-ts VS NamedType

Curious what your issue was with duck-typing. Were you effectively looking to create ADTs that are required to go through a specific step-by-step process, not simply 'look like' the thing that was expected?

If so, you might be interested in [newtype-ts].

[newtype-ts]: https://github.com/gcanti/newtype-ts

this one https://github.com/gcanti/newtype-ts ?

You don't need classes for nominal types: https://github.com/gcanti/newtype-ts

Creating newtypes (via newtype-ts) for fractions and percentages is worthwhile if you're very concerned about correctness, though it's unfortunate that absent typeclasses they're not terribly ergonomic.

I haven't used this library in particular or this one, but bless the authors and all the others for trying to make it happen… the project I'm now leading uses strongly typed strings for different component addresses, and it saves me to constantly look up the documentation what the functions need – it's beautiful.

A "nominal" solution with no runtime overhead can be provided by branded types, other words for this are "opaque types" or "newtypes". Some libraries like newtype-ts exist to help with the construction of these. That library README describes it as:

You can also look at the fp-ts ecosystem for inspiration. newtype-ts and io-ts's Branded type with io-ts-type's UUID as a good example

Alternatively we can use other more sophisticated techniques to achieve nominal typing or libraries such as: newtype-ts or io-ts.

One using fp-ts and newtype-ts

We are going to use newtype-ts to build a branded type:

This article seems to be targeted towards beginners. From the title, I was expecting it to cover something like https://github.com/joboccara/NamedType, which is a technique I would recommend.

I haven't used this library in particular or this one, but bless the authors and all the others for trying to make it happen… the project I'm now leading uses strongly typed strings for different component addresses, and it saves me to constantly look up the documentation what the functions need – it's beautiful.

I think you are thinking of something like https://github.com/joboccara/namedtype

The basic jist is that if you use strong types, you can dispatch more easily inside the function. Because of how the template function is generically written, I don't need to do any manual calls to the template parameters, it all gets figured out through automatic template deduction. In this example I'm using really simple types, but in a more production system I would use something more along the lines of Jonathan Boccara's NamedType. Here is a great article that explains it pretty well [link].