Christopher C. Aycock

Thanks for trying it out! You are correct that I've been neglectful of locale. And I don't have any parameters for `load`, like separators or headers. I will have to...

The HIR will likely need `TraitsKind` and `ModeKind` to complement `TypeKind`. This will eliminate the need for codegen's `direct_repr` in most cases since higher-order items are preferable.

The middle item was fixed in #77.

It might be helpful to read the [Brown Benchmark for Table Types](https://github.com/brownplt/B2T2/tree/main/Media/Empirical) to see how some of the items in this Roadmap can be represented. See also the [Dataframe Algebra](https://arxiv.org/pdf/2001.00888.pdf)....

Would the `Table` have the same functionality as a `DataFrame`? I.e., queries, joins, aggregations, IO, etc?

I definitely like the idea of a base `Table` and then a `DataFrame` that adds an index. Other than indices, there doesn't need to be a distinction in terms of...

How will the `NA` appear in an array when displayed to the user? Currently we have ``` python [1.0, nan, 3.0] ``` Is the plan that any numerical array or...

I like this idea. Numpy's multi-dimensional requirements makes it really difficult to make an unboxed array of heterogeneously sized arrays.

Some of the more extreme: - IP addresses (both v4 and v6) - fractions, storing numerator and denominator as integers