> the simplicity of the language makes it appealing for teaching courses on Turing machines.
I'm not sure. Untyped lambda calculus can be implemented on a universal turing machine and a UTF can be emulated through lambda calculus as they are equivalent, but they're not the same thing. So teaching untyped lambda calculus[0] would not help much with turing machines.
And then, you're got the fact that unlambda removes lambda and uses combinators instead, making things even more perverse.
I'm not sure. Untyped lambda calculus can be implemented on a universal turing machine and a UTF can be emulated through lambda calculus as they are equivalent, but they're not the same thing. So teaching untyped lambda calculus[0] would not help much with turing machines.
And then, you're got the fact that unlambda removes lambda and uses combinators instead, making things even more perverse.
[0] via e.g. Programming with Nothing http://experthuman.com/programming-with-nothing