recursion-ninja

There is a branch [`haskell-function-memoize`](https://github.com/amnh/PCG/commit/fa555908e20a2d431cf736b6887b6a1bc9937ac2) in which I added an initial `HashTable`, `ST`, and `TVar` implementation. It currently doesn't compile because of a type error involving the `s` state type...

I switched to an `IO` based implementation. It kinda seems to work... in a finicky way. I'll try shoving this in our codebase and measure the performance against the FFI-based...

The experiment was inconclusive. There is a fold over the bit vectors (`DynamicCharacterElement`) to perform the pairwise or threeway Sankoff algorithm for the cost and median that takes up all...

After completing #141, we can use the benchmarks to quantify if this Haskell memoized TCM is better for usage with string-alignment median lookups.

We have moderate confidence that the `memoize` function is working as intended and is thread safe. However, the memoized function is not being retained and reapplied between string alignment functions....

We should also look into using [`concurrent-hashtable`](https://hackage.haskell.org/package/concurrent-hashtable).

I have added the memoized functions to the metadata and excised the usage of compact region from the code. However, this has broken the `SAVE` and `LOAD` commands, due to...

Looks to be around 10% faster. That's an effortless yet significant improvement!

If someone wants to spend a few more days toying with GC options after `ghc-9.0` is released, that would be a valuable time investment. We can take the apparently best...

I'll just note that in the dynamic character rework that should be merged into `master` soon, the type of `DynamicCharacter` will be changed to: ``` data DynamicCharacter = Missing {-#...