Feedback request: priority queues in containers

[Jim Apple]

I should also note that these timings are quite temperamental on my
computer; using profiling or removing the performGC calls can make C
faster than D and B faster than A.

> Binomial priority queue timings (in milliseconds, averaged over 20 runs)
> A: Simple Brodal-Okasaki list implementation:
> 199.3697
> B: Wasserman nested type implementation, simple functions:
> 235.0143
> C: Wasserman's 'Quick' list implementation:
> 179.7227
> D: Wasserman's proposed nested type implementation:
> 163.47515
>
> D is still the fastest. However, A is a good bit faster than B. Is it
> possible that C would show a similar advantage over D if it were
> optimized as carefully as D?