Following up on one of my own questions, I experimented a bit with water and pressure plates. The basic idea is to fill an enclosed area with water of depth xxxy, with y being 1/7 different in depth than the x's. One or more pressure plates would then trigger !!Something!! at either y/7 or x/7 depth.
For the purposes of testing I created several areas, filled them to 7/7 water and removed one bucketful of water from each area. A river one z-level above was used to flood each room which hopefully filled everything to a precise level. In an earlier test I was dumping water down stairs into each area but that may have caused some partial evaporation so I wound up redoing everything, multiple times since I'm still noobish.
I started out with 3 miners, 3 masons and one architect with everyone proficient in Mechanics, then embarked with 3 picks, a bucket, enough food and booze to get by and a bunch of stone.
The areas I dug out looked like this--the dark spots are pressure plates:
Later on I dug out 10x1 and 20x1 tunnels that aren't shown in the image.
I didn't bother capturing exact times as that is somewhat dependent on the location of the pressure plates, but I did pause and step through time to see where and how frequently the 6/7 depth changed, and I found that it moved erratically. Sometimes it would move in and out of a tile in one step and other times it would remain stationary for eight or nine steps, but in no case did it move more than one tile from its previous loction. Areas with both x/y dimensions allowed for diagonal and more frequent movement.
Results were a little odd--I'm guessing that it's a function of the largest dimension in a room as opposed to the total area:
2x1 - didn't trigger at all
2x2 - ditto
3x1 - triggered very rarely
3x3 - fairly rare triggers
4x1 and larger - triggered frequently (even the 20x1 tunnel)
4x4 and larger - ditto - the 5x5 room was especially active. I didn't try placing a plate in the corner to see if that reduced the frequency.
In earlier testing I tried setting the pressure plates to a maximum value, eg 6/7 in a 4x1 corridor with depths of 5556. It didn't seem to make a difference.
Water seems to behave like, well, water. (; It's wavy and in perpetual motion, and the peaks and troughs seem to have a hard upper and lower limit, but without the algorithm I can't really say much else about it. Using a 6x1 tunnel with 1-6 pressure plates triggering !!Something!! (or several somethings) every 1-20 seconds could be useful. It brought to mind one of the people here that is building a dungeon for people to solve in Adventure mode. This would be useful with randomizing ways through a particular level that could actually be solved by someone insightful enough to notice and clever enough to figure it out--that seems to be the bulk of people in this forum!