Treksit – Interstellar puzzle based on Graph Theory | Hacker News Reader

Treksit – Interstellar puzzle based on Graph Theory

treksit.com

54 pointsmrkstu13 comments

Simon Tatham’s “untangle” puzzle[1] does this, with randomly generated planar graphs. If you ever find yourself being too productive, download his puzzle collection ;-)

1:https://www.chiark.greenend.org.uk/~sgtatham/puzzles/js/unta...

Yes! Tatham's collection is wonderful. The puzzle that he calls Signpost is one of my absolute favorites. Impressive from a programming point of view, too:

> All of them run natively on Unix (GTK) and on Windows. They can also be played on the web, as Java or Javascript applets.

I use the Android version: https://play.google.com/store/apps/details?id=name.boyle.chr...

As someone who uses nodes a lot in Blender and likes to minimize link ("noodle") crossing - I found that game to be very easy :D The hardest thing about it is the very limited window - points often need to be moved into a small space, and then the points that otherwise wouldn't have to be moved - have to be moved apart to make space. Because those aren't points, they are circles that take a lot of space in a very limited space. Not just visually, I noticed a point really can't touch anything other than the two connected segments in order for your solution to be accepted.

Here's the solution to the last level - notice I make no topological change in the GIF, and yet only after it the solution is accepted. Moreover both before and after this movement no point touches a segment, and yet somehow before the movement the solution is considered wrong... The validator is way too pedantic for the little space there is: https://i.imgur.com/by42EZg.gif

Not affiliated, just came across it randomly and it is a fun distraction that stimulates parts of my brain that often aren't.

20 'puzzles' to work through.

Wish there was a counter, something to tell you "this is puzzle #N". Also, it would be good to know that there are only 20 from the start. "Puzzle #N/20"? Should be easy mods to make. Anyway, apart from that, oddly satisfying to flatten those graphs!

Also - you can still move the vertices even after you achieve the flattened state. I found that usually when flatness was first reached, the graph would be kind of scrunched up. You can keep adjusting to achieve maximal symmetry if so inclined.

> Wish there was a counter, something to tell you "this is puzzle #N". Also, it would be good to know that there are only 20 from the start. "Puzzle #N/20"?

Press the up arrow at the bottom

I also found it an entertaining distraction. For a minute; solved five puzzles, won't do it again.

Two suggestions I'd have for it:

1) Don't automatically mark an answer as correct. Sometimes I stumbled on an answer and didn't really know how I got there. Making the user check once they think they've got it is more engaging.

2) A kazoo cover of Cruel Angel's Thesis for the BGM.

Fun, but easy to crack. Try this instead, that may also make you want

to throw your mouse on occasion (one more 'Knot' solving game): https://knots.netlify.app

> "Rearrange the points on the graph so that no lines cross, and no point is sitting on any of the lines."

I do not understand the second part of the statement.

Isn't a line nothing but a series of continous points?

IE, given any two positions (A, B) on Line Foo, aren't there an infinite number of points between them?

so that no control point (the circled grab handles) is sitting on any line segment between two control points

Relevant theory 101: https://en.m.wikipedia.org/wiki/Planar_graph

Thank you for sharing, I found the puzzles engaging. Moving the nodes and lines in puzzle 20 was a pleasant challenge.

Reply on news.ycombinator.com