Hacker News Reader

Comment by hinkley | Hacker News Reader

There was a puzzle game that came out when UML was still a thing people talked about doing and put on their resumes. The puzzle was a graph of nodes and edges and you won when none of the edges crossed.

I had fun with it for about an hour and then got this terrible sense if deja vu. And then it hit my like a ton of bricks. This is flattening a UML diagram but turned into a game. I pushed back from my desk and went to get some air, and that was the last time I played that game. And (coincidentally?) close to the last time I volunteered to do make class diagrams.

View on HN

Was it this one? https://www.chiark.greenend.org.uk/~sgtatham/puzzles/js/unta...

I went and found this game again relatively recently. It reminds me so much of decoupling and refactoring software or vice versa. I spent a few years occasionally trying to remember what distant memory of an experience decoupling reminded me of.

I quite enjoy both. Whilst simultaneously kind of hating the process.

Incredible sense of clarity once you've resolved everything that was stepping in the way of understanding. Moments of despair and frustration. Sense of progress combined with feeling this could be never ending. Iteratively trying to step towards something better. It all seeming worth it in the end. So many parallels.

Most likely, yes. These games might be fun for students as a learning exercise. Not so much for me.

I love that game... something about solving puzzles just makes it feel like flow. I always crank up the complexity on it, and can do it for hours.

The other game I can lose time in is Universal Paperclips[1], I looped through it 100 times before quitting.

1 - https://www.decisionproblem.com/paperclips/

Universal Paperclips is one of those true gems and in my opinion one of the standout examples within the incremental genre of what a game can really be. also good incrementals: a dark room, SPACEPLAN, kittens game, and trimps. with the exception of SPACEPLAN, all of these games have always been greatly enjoyable despite feeling like the worst type of work that one could be doing. i find it odd that I'm willing to do that sort of resource allocation and microoptimization thing almost endlessly for free, but the idea of doing it in meatspace for pay seems almost intolerable.

I was a game developer when this came out and all of the artists made jokes about how this is just like unwrapping UV coordinates for texture mapping.

I once made such a game, but that was more based from how interesting graph theory can be, and how you can visualize graphs and interact with them. The most difficult part was thinking how to create a solvable level, i.e. how to create a planar graph. Oh, and ofcourse actually completing the whole game, which I actually managed to do.

Just greedily generate a planar graph? lay out some points randomly, start adding edges at random and skip adding any new edge that crosses an existing edge? Then randomize the positions of the nodes to give you a game board?

Or is there a simpler way?

My first job, someone used the Rational Rose diagram tool to print out their diagram (on a huge wide format printer) and it was amazing…ly awful.

There was a bar of lines three inches wide along the top of the chart. They connected a bunch of rectangles on the left side of the chart to rectangles on the right side of the chart. How on earth you were supposed to trace those lines I have absolutely no idea. But we got some good gallows humor out of it.

I used to joke I’d hire some game engine designers to apply their path finding algorithms to solving this problem.

Reply on news.ycombinator.com