So, I got distracted from writing. With other writing. On the on hand, it's good practice. On the other hand, this is just doing a mix of not tackling my existing editing workload, and adding more for later, so that's a little obnoxious.
What else have I been thinking about... Well, the more I've thought about Halstead metrics, the more I think to myself "I'm not sure these can produce directly comparable metrics between languages, and maybe each language or language family should have metrics derived from a set of principles" because I'm genuinely not sure how you'd come up with a coherent way of classifying tokens that accounts for all of "pure functional language that doesn't allow you to bind anything to a name", "tape-based language that also doesn't allow you to bind anything to a name", and "stack-based language that allows for quoting; like, either quoting somehow turns an operator into an operand, or not quoting is an operator!"
I wouldn't think about this so much if I didn't watch people singing the praises of Radon for measuring Python code, and if Radon weren't so... questionable.
I'm going to try working through some of these on my own code, just for the heck of it. That's going to take a bit, so I'm going to publish now.