> This is useful because the language of types is much more concise than the language of terms.
actually holds with sufficiently complex types. For example, would the fully-specified type for quicksort actually be shorter than quicksort?
I don't know that there is any reason to believe so. The best we may be able to hope for is that we'll have less work to do for formal verification, since instead of writing both the proofs and the code, we may be able to only write the proofs and have the code for free. However, given that writing the proofs is, at least currently, much, much harder than writing just the code, I'm not sure this type of development would make more than a dent in the software engineering world.