Your job, should you choose to accept it, is to comment on this thread explaining the most awesome thing you've done this since July 1st. You may be as blatantly proud of yourself as you feel. You may unabashedly consider yourself the coolest freaking person ever because of that awesome thing you're dying to tell everyone about. This is the place to do just that.
Remember, however, that this isn't any kind of progress thread. Nor is it any kind of proposal thread. This thread is solely for people to talk about the awesome things they have done. Not "will do". Not "are working on". Have already done. This is to cultivate an environment of object level productivity rather than meta-productivity methods.
So, what's the coolest thing you've done this month?
Testing an algorithm I developed a byproduct algorithm which outputs 3D crosswords. Automatically given a list of words. Many thousands of them are already on my PC.
This has been virtually unknown before, at least with a few or no black fields.
Now I wonder, can I sell this to newspapers, would their readers enjoy that kind of a puzzle? Six 6 by 6 square crosswords, interdependent with each other by the third dimension at every field.
Or some even bigger cubes, perhaps 10 by 10 by 10? Or a 5 by 6 by 7 crossword?
The building algorithm is here now. But I don't want to spend too much time with this anymore. A middle-man or a middle-firm to newspapers would be appreciated.
That sounds really cool as a word-packing exercise. Thinking about it as a crossworder though, if you reduce or eliminate the black fields I wonder whether such a 3d puzzle would be too easy to solve as each redundant use of a letter reduces the challenge on the solver.
Could it be adapted somehow into a series of 2d slice puzzles where the full 3d solution only becomes evident, or even noticeable later on? for example suppose "puzzle 1" is the plane X=1, Y=1. There is no need for subsequent planes to be parallel to the first; puzzle 2 might be Y=2, Z=3 but without any indication of which dimensions are in play at the moment so the solver has to think about which alignment would suffice? Leaving some clues incomplete so they can only be solved by approach from other dimensions after figuring out the relative position of each sub-puzzle?