Bay 12 Games Forum

It's easy enough to get away with everybody being unskilled.  I used to think a proficient weaponsmith and a proficient armorsmith were necessary, but these days, I'm getting better migrants before those start skilling up anyways.

A proficient miner makes things easier for me, so I take one.

Other than that, it's novice weapon/armorsmith (for the mood potential), some military skill, and proficient teacher.  Teacher's one of the harder things to skill up, and I believe it makes a big difference in military training.  (Student's equally hard and equally important, but I figure teacher is better for my starting dwarves.)

Well, you will need a longer bridge.  Long-term vamps can only make it 6 tiles before the bridge will open, but newer vamps will make it further.  These kind of plate/bridge combos need a bit of jiggling sometimes, as they can get stuck closed when an open signal follows a close signal too closely.  Other than that, looks pretty square.

Quote from: Nil Eyeglazed on March 22, 2012, 02:13:49 pm

Needs an inverter too-- floodgates and doors allow passage on an open signal, disallow on a close.  In fact, really, you want to use all doors, no floodgates, to eliminate floodgate latency.

My bad, I got confused for which direction that the doors work when in a mechanism. In that case, we need to replace the doors with something that is open when deactivated and closed when activated. The only possibilities that I can think of off the top of my head are to either use a retracting bridge (which has latency issues) or to use a NOT gate and a door (which requires more of an investment). Though latency should only be an issue when the inner room is small and the path is therefore short.

Easiest inversion of a door is a hatch over open space.  If latency's acceptable, easiest inversion of a floodgate is a raising bridge.

I haven't tested dfhack's tools yet, so I don't know there.  It should probably be safe, though, given what I've heard from the developer of that tool. 

I've had DF crashes with autodump-- suggests that there are things DF is doing with dumping/destroying items that aren't happening with autodump.  Think I'd stay away from it for framerate testing purposes.

1.  No path -- this is what the undump uses, and what some of my earlier science worked off of, its major downsides are reduced stockpile size if the whole thing gets blocked off, or major lag if its a repeating trap.  Eyeglazed's work on creature timings has pretty much convinced me its impossible to make a pressure plate hallway that will make all dwarves drop their stuff while still allowing dwarves outside to continue pathing towards the stockpile without a repeater, and to me that means heavy lag.

I'm setting up a water blockage system right now.  It's possible that dwarves treat water differently, perhaps ignoring water until they reach the actual tile where their path is blocked, because otherwise you'd have path recalculation every time flow turned a 3/7 into a 4/7.  Will let you know.

2.  Scaring -- So I think the scare dump is a definite move in the right direction.  It seems quite conceivable that every dwarf heading towards a stockpile with a chained creature sitting on it will drop his stuff, assuming proper window placement.  Then its just a matter of setting things up so that the tiles dwarves drop food on are all hatches or bridges, preferably controlled by the exit of all dwarves from the hallway.  Unless youre dropping into magma, you can always have my patented Duck Duck Goose failsafe system, where stupid dwarves who fall in the chute land on fluffy ducks and gooses instead of exploding, and yes, pastures and stockpiles can be placed in the same place.

It's not that scaring doesn't work.  It does.  It's just that the lure tile ends up getting clogged more often than with a plate-hatch undump.  It's unreliable.

I'm using a scare-based autodump to drop clothing into the magma.  A guy with a pet ram bit it.  Otherwise, safe.  Can't say exactly how much has gotten dropped.

3.  Dangerous Terrain: a dwarf who finds himself in an unpathable tile will cancel his job and try to path out of it if he can.  Unfortunately, they dont always remember to drop their stuff immediately.  Dwarves can be put on unpathable tiles by making them fall onto wall grates, vertical bars, or (iirc) fortifications.  You can also flood the tile theyre in/about to walk into.  Since making them fall has many of the same problems as method 1, maybe we should look into repeating mist machines to cancel their jobs?  I dont know if water forces pathing to be recalculated, or if its more of a OH NO IM DROWNING moment, so im not sure of fps woes from this...

I'll have more info soon, hopefully.  A system that drowned dwarves approaching from one direction, but not dwarves approaching from the other, might be a good way to transfer food.

I would also like to do a little bit of building-based dangerous terrain research.  In particular, I want to see what happens when you build a retracting bridge over a raising bridge and trigger them at the same time, or a tick apart.  This is more pure research for me, because I believe any system used for something practical would either need to use a complicated delay system (for hatches) or else it would need to be a 20-tile long bridge, and I am just not interested in solutions involving 20 tile long bridges.

4.  Does combat make them drop things...?  I know Ive had a marksdwarf drop her baby to the carp while reloading, ( the carp was 3 zlevels to low to attack, and carp cant fly, she had only herself to blame that time). If so, we could just put a bunch of repeating training spears in the hallway, then dwarves will dodge fall down a level, creating a one way path and making their original job inaccessible, more so they drop their junk and head towards an accessible meeting hall, except whoops, they trigger more spikes, to drop them back to fortress proper, and then a plate that dumps their junk.  All the falling, hopefully none of the recalculation of pathing that kills fps.

Personally, I find casualties unacceptable, dodges are really random, and you can't control path with spikes.  If dwarves drop items when spiked (I believe they do), then the dropped items are going to be on the spike square, and any dwarves coming to retrieve the goodies are going to get spiked too.  It all sounds rather dangerous to me.

It looks like that should work. It would still have some trouble with children following their parents closely though.

Needs an inverter too-- floodgates and doors allow passage on an open signal, disallow on a close.  In fact, really, you want to use all doors, no floodgates, to eliminate floodgate latency.

Adolescent dwarves are significantly smaller than adult dwarves-- you can build in a minimum weight to get rid of the child problem.  Assuming you're confident that none of your children are vampires.  On preview, what ed boy said.

If dwarves bump into each other in the hallway, they might take corner squares, so that can screw it up when children are involved, but the risk is low.

The real risk is that of your bed being used for something other than sleeping.  A dwarf might start the suffering the effects of a syndrome only after making it to bed-- such a dwarf would be a goner.  Any dwarf dying in bed (perhaps just old age) would tie up burial dwarves in infinite loops until you dismantled the room.

A vampire trap could form a similar design, allowing you to do as you pleased with any vampires you caught-- after a while, vampires surrounded by sleeping, superficially reachable dwarves are going to be useless anyways.  EDIT: Catching children in addition might even be seen as a perk for such a system.

Are those flags... made of stone?