logic programming again

I've been thinking about logic programming again.

At some point a couple of weeks ago, there was a spark and before I knew it I was scribbling away notes on a few different things at once. I think the trigger was a problem at work. Useful context is that my job is figuring out how to build valid furniture out of many disparate products, and then writing code to make that happen.

One of the systems at work actually uses Prolog under the hood to help work out some constraints we have between different products. There are a collection of 'rules' that describe these relationships of constraints between products, and out of the other end, one way or another, comes proto-furniture. But if something goes wrong when running those rules, it's really hard to figure out what is wrong or how to solve it.

Anyway, the next day I had found my copy of The Reasoned Schemer and had reintroduced myself with the install of scheme on my computer. I started thinking about some kind of domain specific language for the problems with furniture that would simplify the problem space, as well as ways of showing input and output from the Prolog rules that might give a clue. I was heading off in all directions at once.

I've slowed down a little since then. I'm still thinking about the problem a lot, and some subset of the possible solutions. It happens that work is pretty busy at the moment, with a deadline rushing at my team with an unstoppable force - so the work on at least the job-specific parts of these ideas is paused and it's just my mind going at it.

To keep the thinking ticking over, I have started watching a series of videos 'From Bronze to GM in miniKanren' by Will Byrd, one of the authors of The Reasoned Schemer. I'm about 5 or 6 videos in, and enjoying it a lot so far. While it hasn't so far started directly talking about miniKanren or logic programming, it is making me think about what I want to spend my time focused on, and making me think about programming language design in general. I even sat down this weekend and read the R5RS scheme report to start reading papers.

Now I have a small queue of papers to read, including one referenced by R5RS that seems to talk about continuations. While I've used 'continuation passing style' programming in that I was a JavaScript developer before Promises existed and so I've used many a callback, I haven't used the continuations found in scheme before and I'd like to get a better grip on them. And so - one idea leads to another, and a little spark ends up taking you somewhere you didn't expect.