Print

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

[Rafal99]

Btw:

Do dwarves who deconstruct floors or bridges or do similar actions also make better decision where to stand to avoid ending on isolated "island"?

[Rose]

My guess is that when they chose where to stand for building it, they pretend the wall is built when pathing, thus ensuring they will be able to go back to wherever they were before building the wall.

At least, that's how I'd do it.

[O11O1]

My guess is that when they chose where to stand for building it, they pretend the wall is built when pathing, thus ensuring they will be able to go back to wherever they were before building the wall.

At least, that's how I'd do it.

Suggesting that if you wall off a north passage that's two tiles wide, and both masons go handle it together, they'll each thing they'll finish before the other, and -both- get walled in!

[greenwatering]

now if only miners would pick the closest mining job instead of the one to the northwest and didn't walk all the way around a wall just to mine something from the "correct" side.

[Toady One]

My guess is that when they chose where to stand for building it, they pretend the wall is built when pathing, thus ensuring they will be able to go back to wherever they were before building the wall.

At least, that's how I'd do it.

Suggesting that if you wall off a north passage that's two tiles wide, and both masons go handle it together, they'll each thing they'll finish before the other, and -both- get walled in!

They keep track of the first valid square they walk through from the beginning of the job, and then use that when they have the material.  That stops it from taking the hit on doing extra pathing checks, but I don't think it would be easy to avoid many simultaneous jobs that have a potentially complicated configuration from walling people in.  Outside factors can still cause trouble.  That's virtually unavoidable.  In the two wall case, if one of the walls is finished while the other dwarf is still fishing for stone off in the dead-end AND the first square they walked in wasn't the good square but the other wall's square, then they will still make the mistake.  Doing the extra pathing calculation is not possible without killing the processor the way things are set up, but they'll also often make the correct decision in multi-wall cases (whereas they'd often be systematically failing them before).

[monk12]

Very interesting! I do like hearing about this kind of stuff, though I'm hardly the first to say so.

[ag]

Hm, I'm wondering if the whole problem of dwarves selecting items by linear distance, without taking obstacles into account, could be slightly improved without additional fps cost simply by applying a fixed weight multiplier (or increment; possibly configurable) to z axis movement. The rationale is that moving between z levels often involves additional distance due to stairs or ramps never being placed on every single tile. I.e. count every z level as equivalent 2-3 horizontal tiles, or something like that. This should at least prevent dwarves running all the way to the layer immediately below, when there is an item 2 tiles away.

Edit: the most flexible distance tweak formula is probably base cost + smaller multiplier * z delta; it sort of accounts for the fact that once you get to the staircase it gets easier to navigate, so a lot of the cost is in the initial decision to move vertically.

[bombzero]

Hm, I'm wondering if the whole problem of dwarves selecting items by linear distance, without taking obstacles into account, could be slightly improved without additional fps cost simply by applying a fixed weight multiplier (or increment; possibly configurable) to z axis movement. The rationale is that moving between z levels often involves additional distance due to stairs or ramps never being placed on every single tile. I.e. count every z level as equivalent 2-3 horizontal tiles, or something like that. This should at least prevent dwarves running all the way to the layer immediately below, when there is an item 2 tiles away.

hmm... this seems like a decent idea, might cause a little extra preformance overhead, but it could work for those short trips turning into 3 mile trips.

[orius]

wall/floodgate/etc. builders often make more reasonable decisions about where to stand to avoid walling themselves in

builders make reasonable decisions to avoid walling themselves in

builders avoid walling themselves in

...

<builds temple of the Great Toad>

This alone makes whatever Toady's working on worth it.

The current version of DF is giving us some damn good fixes and nice new stuff to play with.

[MasterMorality]

wall/floodgate/etc. builders often make more reasonable decisions about where to stand to avoid walling themselves in

HOLY CHRIST!

[Moddan]

An era ends.

Spoiler (click to show/hide)

[vidboi]

Will the new 'where to stand' mechanisms apply to mining, smoothing and engraving designations as well as constructed buildings?

Something I'd like to see is miners mining the square next to them from where they are rather than running round a whole load of tunnels to get a better position, or engravers actually smoothing and engraving from inside the fortress.

[Greiger]

That would be cool but it could also be an opportunity to rework how which side to smooth/engrave from works in general.  Because only the side engraved from gets the value bonus from engravings, it can be a pain in the ass to get engravers to engrave a noble's room and not the other side of the bloody wall. 

Currently if you want to completely engrave a room you have to make the walls 2 tiles thick on the north and west sides, or block access to the other side of the wall.   And if you screw it up, theirs no backsies, best you can do is tear down the wall, and either use a utility to cheat in some obsidian to try again, or replace the wall with some valuable materiel constructed walls.

I can't think of a good way to do that without complicating the way to designate engravings by 5 though.  Maybe if they preferred to smooth or engrave from another tile designated for smoothing or engraving... but that wouldn't catch all cases...

[SmileyMan]

That would be cool but it could also be an opportunity to rework how which side to smooth/engrave from works in general.  Because only the side engraved from gets the value bonus from engravings, it can be a pain in the ass to get engravers to engrave a noble's room and not the other side of the bloody wall. 

Are you sure? I don't think that's true...

[MrWiggles]

Will the new 'where to stand' mechanisms apply to mining, smoothing and engraving designations as well as constructed buildings?

Something I'd like to see is miners mining the square next to them from where they are rather than running round a whole load of tunnels to get a better position, or engravers actually smoothing and engraving from inside the fortress.

As how Toady has describe it, it doesnt have any application for those jobs.