This grew out of a game I was marketing a few years called Alexandria (which I'll probably get around to discussing one of these days). The game involves forming words that snake around a hexagonally tiled board and I decided that the basic concept could put a new spin on traditional word search puzzles.
You form the words by moving from hexagon to adjacent hexagon. You can use the same hexagon in different words but not twice in the same word (an n-lettered word has to use n hexagons). This is easier to show than to tell.
In this example Washington and Lincoln overlap in numerous letters.
To hint or not to hint
There's a bigger discussion here about the amount of helpful information should be presented with a puzzle. My general feeling is less-is-more, but puzzles are supposed to be fun. That means making the difficulty high enough to be challenging but low enough not to frustrate.
For word searches, hexagonal or rectangular, there seem to be three schools of thought.
I. List the words in the puzzle. Maybe it's the pedagogical puritan in me, but this is my least favorite choice. It greatly reduces both the challenge and (in a teaching context) educational value.
II. Give a category. This might be my preferred technique when you're having a group do a puzzle. For example, you might break up a class up into small groups, give them something like the above puzzle, and have each group see how many words in the category they can find.
II. Give a category and clues for each word. Probably my preferred approach. These can be crossword style clues or if you're in a hurry, you can give the first letter of each word followed by blanks for the remaining letters
If you want to try making your own rectangular word searches, I'd recommend using a spreadsheet program. For hexagonal puzzles, try saving the blank grid as a separate file then fill in the letters and save each puzzle as an additional file.
Here are a few more examples