Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Dunamisdeos]

I have a large megaproject planned now.

I sort of ran out of useful things to do a while back.

[LordBalkan]

First of all

@Unhacker,

Have you already overcome your problems with military? I confess that everytime Im watching a live game (Nookrium, captn duck, whatever) i tend to use the chat to ask some tips about this "problem", I appreciate the help they try to give me, but everytime sounds like "oh, it is easy, just press "m" for military and [foreign language for 2~3 minutes trough menus] and its done!". Always end up saying "Oh, ty it helped a lot" but I just dont want to bore them with my non-understanding.

Well, that been said to the main topic

Last year I've found this tool that help to have a goal to the fortress. Actually it have a lot of other games.
Anyway there it is: http://www.scenariogenerator.net/generators/dwarffortress
Maybe some of you guys already use it, maybe not.
I couldnt dare to use, dont feel confident enough about my military.

Thats it, see ya!!

[mikekchar]

My current fortress I want to see if I revive a mountain home for an almost dead civ if they will try to reclaim ruins of their old hamlets (which I will ruin again as an adventurer since they are all occupied by goblins now).  For the fortress, I reclaimed the old mountain home that was ruined by a forgotten beast.  Got a bit lucky because some helpful olm people killed the forgotten beast while I was walling up the caverns :-)  I just managed to verify that the civ is alive enough to send a caravan and outpost liason, so I just need to attract migrants/visitors.  Currently trying to refactor the worldgen fortress so that it is efficient (good luck!).  On the plus side, the fortress is sitting on top of a magma tube, so I have 20 or so magma smelters on level 25 :-)

[Unhacker]

Have you already overcome your problems with military? I confess that everytime Im watching a live game (Nookrium, captn duck, whatever) i tend to use the chat to ask some tips about this "problem", I appreciate the help they try to give me, but everytime sounds like "oh, it is easy, just press "m" for military and [foreign language for 2~3 minutes trough menus] and its done!". Always end up saying "Oh, ty it helped a lot" but I just dont want to bore them with my non-understanding.

Well... I'm sort of figuring it out, but I'm not really all that confident in its usage. In my other active fort "Riskytowers" I'm sort of experimenting with the military system, this is mostly due to the fact that I got lucky enough to get metric fucktons of various iron ores meaning that I can actually outfit a military.

Haven't overcome my problems with the military but I think I'm getting close, though that might be because I haven't tried messing around with marksdwarf squads.

In terms of Riskytower's goals, I think make a large tower spanning out of the fort, preferably from the staircase. And to have a functioning book binding industry.

In terms of how the fort is currently doing, I need to be able to fully outfit my military before I am able to start generating as much wealth as armoring (I've got the weapons made) my dwarves takes up the entirety of my metalworking industry (all two of the forges, I have three smelters, one makes coke from coal, one makes iron and the other makes platinum and gold)

[Libash_Thunderhead]

Well, I just play until I'm bored.

Edit*
If play-as-long-as-you-can is a goal.

[Thorfinn]

Hm. I'm curious. Can you show explain setup you had and the late-game issues with it?

Was planning to start a new game so I could just use screen shots, but I'm not going to get to it for a bit.

Spoiler (click to show/hide)

|,- = Wall
. = space, workshops as needed
p = seed stockpile
P = stockpile
f,F = farm



L0
(The whole thing is usually less than 20 tiles long so I can cover with 2 bridges. Double wall to avoid dodging through.)
------------
------------
||...PP...||
||...PP...||

||...PP...||
||...PP...||
------------
------------

L-1
(Moat around outside. Surface crops.)
..............
.------------.
.------------.
.||fffppfff||.
.||FFFppFFF||.

.||fffppfff||.
.||FFFppFFF||.
.------------.
.------------.
..............

L-2
(Underground crops and ag workshops. To left beyond workshop space is mostly turkey/pig pens, etc. Outer walls omitted.)
...PPPppfffffff

Main thing is trying to minimize pathing problems in late game. First iteration all stockpiles were built on up/down stairs. You probably know that worked great at the beginning but extremely poorly by late game. Too many choices for too many dwarves. Not too much improvement having double stairs every 5. Might be double stairs every 8 or 10 or more are better. A single central 2x2 is much better in terms of FPS, though it starts making some of the hauling jobs quite lengthy. A trade-off between FPS and efficiency. And that's probably the answer right there. How much efficiency am I willing to lose for the sake of FPS?

I'm going to test putting a wall between the seed stockpile and general stockpile so they are not tempted to path through the general pile. Unfortunately, if it turns out it makes a difference, I think that means I probably need a wall between the fields and the seed pile, too. Which I don't have the option of doing on L-1 unless I make the surface fort too wide to span with a bridge.

But slamming feeder workshops right next to one another was bad from mid-game on. For example, the tanner right next to the butcher, which seemed like a good idea at the time, turned out to be poor for pathing as well. I was better off to put a wall with a single opening between the two, and by later in the game, sealing that off and making them walk into the main hall.

b = butcher
t = tanner

Great in early game, lousy by about pop 100:
bbbttt
bbbttt
bbbttt


Almost as great early game, not quite as bad at pop 100:
bbb|ttt
bbb.ttt
bbb|ttt


Mediocre up until around 150 or so, when it drops off. Merging into the main hall changes it from mediocre to terrible, btw. Better to have a stockpile they path through.
-.---.-
bbb|ttt
bbb|ttt
bbb|ttt

Not !!Science!!, because there are all kinds of differences that could account for it elsewhere, but putting a door betwixt the two and locking it doesn't seem to improve things much, nor does replacing the door with a wall seem to work as well as never having mined the hole in the first place.

[Fleeting Frames]

Okay, first, I'll note that butcher/tanner should be stacked in y, not x direction - x like you describe they're like this:

bbb ttt
bbb ttt
bbb tt

Good points I haven't considered, though. The later trouble with stockpile/stairs makes sense with central staircase, though it hadn't occurred to me - I've been putting off just watching my dwarves as they go on about things to see faults, and my test embarks for finding things out quickly always have low fps for high pop. If you don't mind some bucketing hassle, you could potentially dig downstairs in your farming area, build floors, wet, build farm, remove floors and then put seed stockpiles right underneath on upstairs so they don't go to main staircase and never have a reason to go there expect for seeds unless your stairs are congested.

I know about butcher's workshop ability to overwhelm haulers and because of this I've sometimes left some distance between, say, cages of livestock and butcher's workshop. Limiting the number of haulers accessing the shop directly with a small intermediate receiver stockpile link could help fps with that.

Though seriously, what is going on around those 3 workshop layouts that 2nd is better than first? If hauling straight butcher→tanner, then you should have the tanner's NW and SW corners already act as walls.

And yeah, never having mined is better.

[Thorfinn]

Okay, first, I'll note that butcher/tanner should be stacked in y, not x direction - x like you describe they're like this:

bbb ttt
bbb ttt
bbb tt

Sure. Looking at the farm setup graphic, the workshops are above one another rather than next to each other. Didn't even think about displaying them that way. But makes me wonder whether maybe some of my testing was done with them horizontally placed. Hmmm. Thought I would have noticed the traffic pattern, but maybe not. I wouldn't have been looking for it, for sure. All I was doing was timing how long it took a complete n00b to process 10 turkeys. And at that time, the seed SP was to the right and the finished food SP to the left, so going around may well have been the hauler route. Early testing had intermediate product SP built on down stairs between the two shops, down to a larger finished good stockpile adjacent to the dining hall, which turned out to be abysmal for FPS loss. I think it was that every time someone wanted a drink, he pathed through the crowded shops in addition to the central stairs. That's all I came up with, anyway. The next iteration with intermediate good SP built on floor rather than stairs was much better. But I wouldn't swear the orientation was the same...

Good points I haven't considered, though. The later trouble with stockpile/stairs makes sense with central staircase, though it hadn't occurred to me - I've been putting off just watching my dwarves as they go on about things to see faults, and my test embarks for finding things out quickly always have low fps for high pop. If you don't mind some bucketing hassle, you could potentially dig downstairs in your farming area, build floors, wet, build farm, remove floors and then put seed stockpiles right underneath on upstairs so they don't go to main staircase and never have a reason to go there expect for seeds unless your stairs are congested.

Done something similar. Had seedpiles on stairs down the right side of the 1x7s so there was easy access to bring seeds up from the dining hall. Works well, though I seem to rarely have enough levels before the aquifer, and isn't really applicable until you can afford the time for a second aquifer pierce.

[Fleeting Frames]

Hm, timing a complete noob is a nice baseline always-doable idea.

Pathing through indeterminate SP-downstairs and workshop shouldn't be an issue for touching workshops, but I take it the touching tanner/butcher had both workshops used for north-south pathing too, not surrounded by walls on 3 sides?

For the iSP->, was that place a dead-end room for a drink or did it have an actual, alternate longer path for getting one?

(Just thinking what are non-workers doing pathing through workshops, though this makes a good argument for traffic restrictions - though they're probably no use when dwarves bounce into each other on the stairs.)

On multiple paths, central staircases and rampways, one thing I've had floating around in my head is rampways surrounding workshops/workshop support in center

I.e.

Code: [Select]

▒▒▲▼▒▲▼▒▲▼▒▒
▒▒▲▼▒▲▼▒▲▼▒▒
▼▼AAABBBCCC▲▲
▲▲AAABBBCCC▼▼
▒▒AAABBBCCC▒▒
▼▼DDDEEEFFF▲▲
▲▲DDDEEEFFF▼▼
▒▒DDDEEEFFF▒▒
▒▒▼▲▒▼▲▒▼▲▒▒
▒▒▼▲▒▼▲▒▼▲▒▒
Since the central rampways have annoying unused space. That said, while I haven't tested it (if combining workshops and ramps I tend towards slanted-stacked industries, which is unsuitable for place to go through) I wouldn't expect it to do well fps-wise with the above diagram having 10 alternate paths. If one culls 4 in above, there's always only one closest rampway to workshop, which would work out fine for workshop jobs, but pathing from bottom SE to top NE corner would have all rampways as possible paths if one uses it as central ramway that connects back together on top and bottom. (Perhaps a four-workshop 4-rampway variant would further isolate the paths.)

Though your main hall and stairs experiences suggests that one shouldn't have workshops and main arteries cross, and if that is done for above diagram there will be no communal area at the other end of those ramps.

That the downstairs into dining hall weren't used as pathway is heartening.

[Thorfinn]

Warning! Digression from OP.

Spoiler (click to show/hide)

Yeah, a more complete layout of the original L-2 is:

P = goods pile
X = up/down stair with stockpile on it
+ = floor with stockpile on it
k = kitchen
b = butcher
t = tannery
f,F = fields

PPPkkkXXfffffff
PPPkkkXXFFFFFFF
PPPkkkXXfffffff
PPPbbbXXFFFFFFF
PPPbbbXXfffffff
PPPbbbXXFFFFFFF
PPPtttXXfffffff
PPPtttXXFFFFFFF
PPPtttXXfffffff


This had good throughput on the n00b turkey test (ntt), but later game lagged something awful. (Something that just occurred to me is that I was not using 0 level cooks or tanners. If the butcher was using bigger creatures, or the cook/tanner couldn't keep up, that would slow things down, no fault of the butcher or the layout.)

Next game I replaced all the X with + and the ntt was a little slower, but much less laggy after 100 or so pop. My hypothesis is with only central stairs, there is no checking for pathing through dead ends.

Next build had 3x3 intermediate stockpiles between the kitchen and butcher and between butcher and tanner. Ntt values best so far. I attribute it to clearing clutter faster. Rather than waiting for the kitchen and tanner to need materials for the next job, the haulers just moved it to SPs.

Then I made those intermediates into stairs down to finished goods piles. Ntt roughly similar. But lots of later lag. My guess, again, is calculating pathing through those intermediate SP/stairs.

Then I tried leaving walls and partial walls between the shops. I've not tried more than 2 walls per shop so far. One between the butcher and tanner, and one between the shops and the +. So I had planned on trying a wall between the k and b first, then walls between all shops, but the more I think on it, the more I'm pretty sure the layout a month back had the shops in an E/W orientation, instead of the N/S orientation I'm using this go around. And that would make a difference. Sigh.

But bear in mind, this isn't comparing apples to apples. These are all live forts with different worlds, default worldgen. They all have had at least one set of caverns pierced, as well as magma brought to surface. All had streams, but some were much larger streams than others. Most have had at least some candy harvested. They are all really close to the same mix of professions. Some had no iron. I'm not sure how to go about testing it properly. Each fort is going to develop differently, if only because of the different caverns and veins of metals. Oh, another think that just occurred to me -- even though the n00b had no skill at butcher, they would all have different attributes. Maybe the only useful information out of all this is how it might affect late game lag...

Pathing through indeterminate SP-downstairs and workshop shouldn't be an issue for touching workshops, but I take it the touching tanner/butcher had both workshops used for north-south pathing too, not surrounded by walls on 3 sides?

I didn't even think about the possibility of guys with no business there pathing THROUGH the workshops. Maybe it doesn't matter if I can't stop the algorithm from checking out the possibility anyway. I don't see the weighting will make any difference in whether it includes that path in the options it calculates, just on whether it is discarded.

For the iSP->, was that place a dead-end room for a drink or did it have an actual, alternate longer path for getting one?

Oh, crap, you are right. If a surface dwarf is stopping on L-3 for food or booze on his way to the dining room, it's the same distance. The same is true for dwarves working below the aquifer with my other variant of putting the dining hall in the sunshine. No wonder adding it made for more late-game lag.


Interesting thought on rampways. Is that "only" to convert 2 moves into 1, or is there another purpose?

That the downstairs into dining hall weren't used as pathway is heartening.

That's a really good point. They were a longer path, and I didn't see them used, but I have no explanation for why they wouldn't have been considered in the pathing algorithm, and thus a source of FPS loss. Maybe because from food/booze to table is a basically straight path of as little as 4 weight 1 tiles on the same level, and the algorithm "knows" not to bother looking for something shorter that spans multiple Z-levels? Pure speculation, knowing as little as I do about the pathing algorithm in the game...

[Thorfinn]

More digression.

Spoiler (click to show/hide)

Here's a thought. Same L-2.

L-3
(or whatever is 2 under the aquifer if L-3 is aquifer)

X = up/down stair with booze SP
P = Finished Food SP
| = Wall
m = meeting/dining hall
/,\ = masterwork or artifact door.

XXP|mmmm
XXP\mmmm
XXP/mmmm
XXP|mmmm

There would be no shorter path from anywhere to those piles than through the main passages. And from there, they are only two tiles or so to the dining room. Whatever the algorithm, I can't imagine that it doesn't exit when it gets a path longer than the current best path, which seems like it should start with as straight a line as possible.

This is probably a 2x4 stair column, at least this far, since that's what came through the aquifer. From here, you probably don't need more than a 2x2, or maybe even a 1x2.

I'm still wondering why the back stairs from the meeting hall up to the seed stockpile didn't seem to affect late-game FPS. If it's a real effect, and not just a fluke of that particular map having less going on, that could be really useful in, for instance, the metal processing area, too. Oh, and definitely the setup for the trade depot.

[Fleeting Frames]

Spoiler: "I see the digression, and deem it interesting" (click to show/hide)

The complete layout of z2 seems to have 4 single-wide paths between goods pile and stockpile/updown stairs, and for half of those an encounter could force dwarf to step on workshop centre tile. I'd think that should only matter if dwarves are currently hauling to P and meeting each other, but if I understand you correctly you had it cause lag even when the workshops and stockpiles sat completely unused for months.

The nice thing about the layout is that you can very easily replace the stairs with floors, just 1 use of liquids plugin to compare seconds taken for exact same tasks, then kill process and do it without floors. If you still have a save with X/stockpile, can you post second numbers of 3 (randomness) runs each?
Killing and reloading should ensure relatively same-ish conditions, though I'm not sure what task to use for finish - perhaps a birth that'll occur shortly in fortress.




You could stop the short-side no-business pathing with walls if not restrictions, worst case, while still having tanner and butcher touch.



Aha, that explains the late-game lag.

Rampways are mostly to convert 2 moves into 1, yeah, though they're also supposed to cause less lag than stairs. (And have aesthetically different appearance that one may find more pleasing.)

As a side note, if not using a spiral, a descending rampway can descend into four downwards directions at once with + design compared to descending stairway's 1, though they'll have to have wall behind them towards the rising direction, so semi-floating workshops over such rampways will reduce min-distance to end in one direction only. Have to mind the impassable squares far more with that, though, so it's kinda bigger pain to design(ate).


For second spoiler:

The pathing algorithm will stop flood-filling for a path once it has found the complete path from A to B, yes.

But note that booze is not taken to the dining room - a goblet (if possible) is taken to wherever the barrel is, dwarf drinks on top of barrel (, and then drops goblet).

So for that diagram drinking, what will happen is that thristy dwarf will stop on the stairs to drink, and other dwarves will have to path around them, potentially causing fps loss.

Food, being brought to a table, seems fine though, albeit I've had dwarves make some weeeiiird picks for which tables they eat at, so if you have any tables elsewhere that's going to be a factor.

As a side note, this first you mention weighting. I've been assuming high weight for the stairs and normal for everything else, this correct?

[Grand Sage]

Bloodbaths usually go a long way in entertaining me (in DF). Therefor, I almost always try to include an arena in my fort. Other things I'd like to do:

- Lots of cage traps on surface and caverns. I need them for the arena and

- Execution chambers. haven't done one of these but there are several to choose from: drowning chamber, execution by axedwarf, execution by marksdwarf (in lack of rifles), execution by ballista (hehe), walk the plank (no need for water beneath...)

- money. lots of money and shiny gems. and bars of gold an platinium. and when I say money I mean coins. preferably made out of gold, platinium or adamantine.

- Throne room. At some point, I always want the monarch to come to my fort, and there is no cost to high when it comes to pleasing him/her. adamantine statues anyone?

- breeding grounds are also often on the list, right now for food and the arena.

- defence mechanisms are pretty awesome to construct to, especially if they include ballista, lots of traps and inginious fortifications. you cant overdo it

I usually dont complete this list before an update breaks my fort, but im sure i can find more projects if needed. like temples. or rooms for everyone. or an aboveground city filled with non-dwarves, while my dwarves dwell below ground.

[Quarque]

With my current fort I'm mainly trying to learn how to get the basic industries functioning. For example, this is the first fortress where I managed to get my dwarves to use the pig tail for clothes instead of booze - that's a big step ahead!
It's also the first fortress where I managed to dig down to the magma sea and I built myself a very dwarven smithy deep within the bowels of the earth. My ultimate hope for this fortress is to set up a functioning automated wheelcart railroad to transport ore down and shiny metal objects back up. If I succeed, I will produce a metric fuckton of Brass and build some kind of monument out of it.. not sure yet what it will be.  Haven't found iron; I love brass for being so lucrative and shiny and for the cool name.

Maybe I will attempt to train my first military dwarf in the next fortress. I still fear the 'm' menu. 

[Urist McVoyager]

That military menu is definitely something to learn. Each page has its purpose and if you don't get them down you're going to have subpar units for sure. Not to mention there's all sorts of other requirements outside that screen, that you have to understand as well. Like barracks and supplies and locations for stations/patrolling. I understand it far enough I could probably teach bit by bit. Anyone wanna try learning again can just PM me for it.