Ah yes, *that*. I had imagined a re-writing of the current fluid dynamics. ...
Why would water without any pressure move fast?
But ya, if there was some way to get it to teleport at maybe 3 instead of 7 but stop once it was level-ish the slowly widening slopes of water wouldn't stick out as much.
Your hopes are a bit screwy though as hydraulics dudes probably don't have an especially good way of representing water in tiles like these.
That and Toady, being a mathematic professor, probably understand hydraulic sufficiently. The only difficulty facing him is HOW to implement hydraulic into Dwarf Fortress without the FPS being shot to hell.
Remember, one of the most difficult problem for computer to simulate is fluid dynamics (if you try to do it realistically).
Though I think the current issue with Dwarf Fortress' fluid dynamic is that water seems to have a really high viscosity (which will result in water flowing really slow down a channel). Of course, this may be an artifact of the fluid being represented in 1/7 chunks (low viscosity fluid will spread to a thin film rapidly, while tracking it in 1/7 chunk inherently slows it's rate of spread).
If you've watched the numbers dance very much it's clear that the units glide around on top of the next layer below them as a group but in a random sort of direction unless they hit a spot two digits lower at which case they fall down to that level and possibly continue the process. The teleport behavior requires actual pathing so it's slower but this non-wander behavior gives the kind of constant directional progress we want.
...
Exactly.
Anyway, I was operating under the assumption that if Toady were able to do it himself and it were easy, he would have already. I hadn't thought of the performance issue, but there are still some odd properties with water that come about because of innaccurate simulation (again, my presumption).
For example:
1) if you channel into the floor and it has access to a larger body of water above it, nothing happens. So a toilet-like U-bend (S-bend?) isn't possible.
2) if you channel a one tile access to a river, split access into 4 sub-channels, let those channels dump into a tube which has a one tile exit for the water at the bottom, and your tube overflows. I guess because the 4-splits serves to suck more water from the river? Pretty sure that's not right.
3) again, you can have a 7-tile of water back-logged up a 1-tile channel accross the map and have it dump into a huge reservoir. Water will actually seep into the ground and dry up before the reservoir fills, if it ever fills. The water should flow faster in the channel and slow down once it gets into to the reservoir. Pretty sure this is also not the case.
Perhaps you guys are right and its not a matter of expertise. Either way, it doesn't seem to be working properly. If I am wrong and water actually does work that way in large quantities, by all means disabuse me of my misconceptions. But I am pretty sure (from an interview I read) that Shoku's got it right.