Hrm. Missed a couple days. I haven’t even been that busy, I just never actually went and took a photo and thought about what to write about. Oops.

Rain all day today. I couldn’t find my hat at first. I have a Tilley hat that I bought nearly 20 years ago because it had the widest brim of anything that I could find: wide enough that it does a pretty good job of keeping the rain off my glasses.

It turns out I had left it in the tractor back on Friday: it had been threatening rain but then I was in the cab of the tractor so I didn’t need it and it was a little hot in there so I took it off and tossed it behind the seat.

Anyway. Very wet today.

looking out at a rain-soaked dirt/gravel footpath through scrubby green plants growing in the driveway. On the right you're looking past the shed door and a round red wooden silo with hostas at its base. Straight ahead is a giant branching silver maple.

I spent a little time today poking through graph representations, wondering what would be appropriate for a Slitherlink implementation. A winged edge structure doesn’t seem like it has quite enough information. And the descriptions of the quad-edge are theoretical enough that I haven’t quite wrapped my head around them yet. Mostly the bits about the vertex/face duality. How do you implement that? Some places say the vertices corresponding to a face are the centroid of the face, but does that still make sense for concave polygons? It seems like it would be simple to construct a polygon where the centroid was outside the polygon. Maybe you don’t need coordinates for the dual vertices if you never use the dual graph? Do quad-edge structures handle graphs with holes? I don’t know if I’d ever need that, but it’d be good to know.

But I think I have a better understanding of the kinds of operations and data that I’d need to manage to represent a graph in a way that I could generate puzzles and analyze the current position of a partial solve for errors and hints and so on. So it was useful. I’ll keep picking away at it.