data-category VS haskell-lox

Even a fairly simple statement like "F preserves direct limits over N" is basically impossible to express like this. You can step further from Hask and work at the type level until the very end (which I believe is the approach taken by data-category), you can resign yourself to only expressing things that can be "defunctionalized" (an appropriate use of the term, I think, if not a correct one) down to Haskell functions, which gets you (Co)Yoneda, Lan, Ran, etc. in the general case and I think Traversable in this particular instance, or you can take some intermediate approach with constrained functions and/or explicit witnesses in your data types, but you can't make proper category theory "just work" the way it should.

That's a great idea. I think for the next post, I will provide a PR. For now, I just created this PR: https://github.com/gdevanla/haskell-lox/pull/1 , where I just duplicated the `Scanner.hs` file. Perhaps, there are better ways to do this!