Joshua and I thought our new typechecking algorithm for higher-rank polymorphism was easy to implement, but I have to admit that we weren't 100% sure about this, since we've been immersed in this work for months.
However, Olle Fredriksson wrote us within one day of the draft going online with a link to his Haskell implementation. This was immensely reassuring, since (a) he implemented it so quickly, and (b) he didn't need to ask us how to implement it -- this means we didn't leave anything important out in the description!
In fact, Olle is technically the first person to implement the algorithm in the paper, since both Joshua and I have actually implemented variants of this algorithm for different languages. I'll try to write a small ML implementation to go along with Olle's Haskell code in the next few weeks. If you want to beat me to it, though, I have no objections!