Data collection version 2

After a couple of slight missteps with the first round of data collection, the repo at this point reflects the exact state of everything when the final data collections were carried out.

In particular, the Racket and Typed Racket Makefiles are now the same thing, and it has been modified to ensure that the largest iteration and size counts are not run together since those were proving to be mind-bogglingly slow. As in, hundreds to tens of thousands of times slower than the next-slowest run. I'm pretty sure one of them was timed as taking abut 70 minutes to complete a single run.

Data collection version

This is the repository as it was when the first benchmarking data were collected, on 19 June 2020. This version is tagged so as to ensure that one can always return to the appropriate commit, in the interests of research reproducibility. It is not, however, expected to be the state of the repository when I submit the work to a relevant venue.