How about a define-type-alias?

[bon]

...or some other moral equivalent of the Haskell type statement for introducing a synonym. It would be very helpful as an organisational tool.
Judging by a couple of comments in the thih example, I'm not the first person to think this.

[stylewarning]

I agree this could be useful. I know other communities are divided on the matter (type aliases make code hard to debug). I've just been using "newtypes" (i.e., ADTs with one constructor), which have no run-time penalty in release-mode, but have a penalty in terms of characters written.

[ornamentist]

Would this work for a type alias to a function type, e.g. Relation as :a -> :a -> Boolean?

[stylewarning]

Would this work for a type alias to a function type, e.g. Relation as :a -> :a -> Boolean?

You can't hide the :a even with a type alias. You'd have to have (Relation :a). The above would still work.

[ornamentist]

Thanks - I should have said Relation :a above.

[ornamentist]

Would a macro work to avoid the match/deconstruct steps, particularly for a "simplifying" type alias to a more complex type definition? Some kind of (nested) define-alias macro?

[stylewarning]

Would a macro work to avoid the match/deconstruct steps, particularly for a "simplifying" type alias to a more complex type definition? Some kind of (nested) define-alias macro?

I think having first-class support for something like define-type-alias would be best. I don't think macros would dig us out if that a hole.