contification paper: formatting

[Matthew Fluet]

> Yes, it just still doesn't sound right to me.  Is handling the mutual recursion
> actually hard?  

In theory, no.  It's just not in the implemenation of the CPS IL.
Probably will be eventually, and then it would be interesting to compare.

> The fact that the benchmarks don't need it often (but even they
> do some times) isn't too convincing since all it takes is one hot loop which
> is now forced to do frame construction/destruction.

True.  Only logic (not reported) and kit actually have uncontifiable
functions.  All the nested functions (should) get the benefits of
contification.