Puzzle 34 - alternate configuration showing that the two gear bits need not make a latch


Strictly speaking, this is not a solution to puzzle 34 as defined, but an aid in understanding it. Helped me at least.

I think this alternate configuration is relevant because it makes it easier to understand the XOR implementation:

  1. Unpacks the 4 lines of the XOR logic table in a more obvious way than puzzle 34 and its solutions;
  2. Does not use the latch configuration of gear bits which I found to be misleading here (unnecessary for this problem).


You can convert the canonical solution into a non-latch version too - just swap the added gear bit with the ramp down/left from it.


I’m working on something that takes images of a TT board and reads the board layout. (Nothing fancy – TT pieces are nice bright colors so this is an image-processing-noob-friendly project.) Would you mind if I steal your jpeg and use it for testing?


No problem.
Please make a topic for that software, it’s a great idea.


I don’t think it’s ready for its own topic yet, but . . . https://github.com/MichaelSpencerJr/TTBoardReader

(Mostly I just made a logger and a DLL. I have . . . feature creep issues . . . when I don’t have a team to bounce things off of and stop me from going crazy. If this were a book, it would contain a super clean professional table of contents, and then all of the pages inside would be blank.)

The plan is to get the happy path working (nice clean well-lit well-aligned image) and then ask the community to take and share some sample board images, ideally from reasonable but odd angles and/or with potato cameras and bad lighting. But I need to do some work before I ask anyone else to do work.


Can you make it cross platform? I run Linux. Thanks :grinning:


Whoa. That is awesome. I can’t wait to try it out! I’ll help in any way I can. So cool!


Drat, I thought this would stay somewhat hidden here. :slight_smile:

No it’s not awesome yet. I deserve no praise yet. Praise scratches that emotional itch, gives you the positive feelings from “I made a thing!” – and if you get that praise from an unfinished work, you lose some of the motivation to keep working on it. Plus I’ve got a new distraction at home: my prime day 960 EVO 1TB just arrived and I’m deciding what to put on it and how.


Although since I have your attention, @paul . . .

Do you have any interest in there being a sort of SpaceChem-like leaderboard, that plots submitted solutions on a histogram for number of pieces, time of execution (however it gets modeled), or other metrics?

If so, I’d be interested in creating a leaderboard service for you to host, if you want to. I can email you with contact info and we can set up a phone call sometime to discuss. I also won’t be offended if you’re not interested in doing that: feel free to defend your boundaries, as I promise that won’t discourage me from contributing in other ways.


Ha ha! Well in that case, I should have said “Yeah right you could build something like that! It would be waaaay too hard. Probably not even possible.”


That would be aweso- er, I mean, nope, don’t waste your time. I’ve talked to the world’s best engineers and they all say it’s impossible.


Right, but seriously we’d need to have a discussion about what your environment can support. If you’re not interested in hosting something official, any such leaderboard will be unofficial and I can make it in whatever I feel like hosting myself. If you’re interested in hosting something, we’d need to talk about what your web hosting can conveniently support and I’d need to target that environment. I have a feeling you won’t be hosting any WCF services on IIS, which would make this super easy for me but is probably a needlessly expensive web hosting bill for you.

Edit: or if we could have some interested community members say they’re willing to run IIS in VMs . . . well, work has been trying to shape us into full-stack engineer types so I have a tiny bit of experience with Chef, Vagrant, Docker, and AWS EC2. I am NOT a proper ops person so I guarantee, left to do this all myself, I would make several critical mistakes that will lead to the whole thing falling over. We could run an haproxy node on EC2 and point it at various volunteer community machines (people who are willing to leave computers on most of the time) – so at least when it falls over, nobody’s risking a machine they care about if it’s a security failure.


Probably oughta move this discussion to another topic, but I do want to interject real quick that I have a little web development experience so I’m mildly interested in helping a bit if I can. (The backends I’ve had experience with were Flask and Django, both based on Python.)