Still not in love with humanity

[tahnan]

The apartment is clean. It still has a desk, a sofa, and two chairs in it that really ought to be anywhere else.

And I've just discovered that I have opinions about elegance and aesthetics in word search grids. Lord help me.

Comments

[dr_whom]

What are your opinions about elegance and aesthetics in word search grids?

[tahnan.livejournal.com]

So I put, side by side, two drafts of the same wordsearch, with the letters replaced by dots and the unused letters replaced by asterisks. (I'd show them here, but since it's for a forthcoming puzzle, it's silly for me to put up the answer. Roughly, though, it looked like this (http://www.suberic.net/~tahnan/wordsearch%20answer%20diagram.jpg).)

So looking at the two grids:

In Diagram A, the unused letters are much more scattered through the grid. In Diagram B, there's a cluster of six (three in one row, three in the row below offset by one colum), another cluster of six (in a 3-then-2-then-1 right triangle); there are no unused letters in the bottom four rows or the first two columns (the grid is roughly 17-square).

In Diagram A, there are diagonals in a variety of places, many crossing each other (the crosses in the sample diagram above) in a few different places. In Diagram B, there's a set that runs SE from the NW corner and a set that runs NE from the SW corner, and that's all; they meet up and cross each other in a clump near the center. (Note that the diagram doesn't indicate directionality of the words, so "SE from the NW corner" could mean "NW to the NW corner", and there's certainly an aesthetic difference between a bunch of words running parallel and in the same direction, and a bunch running parallel but in opposite directions.)

Consequently, in Diagram A, the longish answers near the edges are broken up by having diagonals through them; in Diagram B, there's a much "boxier" effect in which there's a bunch of answers across the top several rows, a bunch across the bottom several, a bunch down the columns on each side, and only in the center do things jumble up a little more.

There are other things I'm not fond of (e.g., answers that overlap by more than one letter, like having EXAMPLE and PLEASANT in a single row as EXAMPLEASANT, or EXAMPLE and AXELROD as DORLEXAMPLE), but I knew that already.

Looking at the sample above, I'm actually pretty happy with the way it came out--I wrote it for National Chemistry Week, and by "I wrote it" I mean "Professor Morse said 'here's what I want' and my Python script gave him a grid; I was just the messenger". There's a little clustering of unused letters, but not much, and I feel like the words aren't too organized. Incidentally, if you feel particularly puzzleish and like really hard criss-crosses: the wordlist for that wordsearch consists of 51 elements from the periodic table, which ought to be enough for you to reconstruct it (not counting the unused letters), right? Answer available here (http://www.nesacs.org/TheNucleus/Oct06.pdf).

[rikchik.livejournal.com]

It's awesome that you have opinions about word search grids. Anything worth doing is worth critiquing.