User talk:Squirrelloid

AI issues I've noticed

Is there some place to report bugs, because the following have been really annoying me.

And many of them cumulatively add up to many more dwarf deaths than would be expected if dwarves actually followed orders or behaved mildly intelligently. (And since they seem to enjoy ignoring orders which would lead to more intelligent behavior, and you can't actually control their actions, a halfway decent AI seems necessary - the goblins have a far better AI than the dwarves, which is silly).

Animals try to path through tightly closed doors

I've long noticed that making a door tightly closed was the best way to get a pet trapped on the wrong side of the door. This is pretty annoying as I can't see any way to kill off pets without making elaborate death traps, and because tightly closing ends up doing the *opposite* of what its supposed to do all those annoying pets end up cluttering my hallways. So I did an experiment. To a location with 2+ modes of entry, I forbade all but one of them, and tightly closed the last door. All the pets attempting to gain access to the other area immediately moved toward the tightly closed door. While they were stuck "?"ing outside the door, I forbade that door and made another door passable but tightly closed, and immediately the pets moved to the new door. If I forbade all the doors the pets wandered around randomly and stopped trying to access the other section.

So apparently pets draw paths like dwarves, and don't understand why they can't get through tightly closed doors, so each time they fail to follow their path they repath through the same door and sit there - leading to pet pile-ups at tightly closed doors and ultimately pets getting through them when dwarves pass through. Shouldn't a tightly closed door be treated as forbidden by pets, and thus they shouldn't try to path through it?

I have found one way to make a much more pet-impassible door. You dig a way around the door that is strictly longer than the door, and then use traffic designations to restrict the short way. Dwarves will take the long way around the door and never use it. Pets ignore traffic restrictions, though, so they will be stuck against the door unable to pass it. Cheepicus 05:00, 17 October 2009 (UTC)

Dwarves Stay Inside option doesn't work properly

So, ordering your dwarves to stay inside doesn't stop them from targetting items/locations outside to work in. Instead, they walk outside, sometimes many squares outside, before realizing they are in a forbidden area and then head back in. This is really aggravating during things like a goblin siege where as soon as the first goblin dies all your civilians will immediately try to pick up its corpse and loot it, walking outside into the line of fire. As there is a 'dwarves ignore refuse outside' option which causes dwarves to not even target refuse outside, shouldn't 'stay inside' work the same such that nothing outside can be targetted by dwarves? This would actually lead to the expected behavior of dwarves actually staying inside when told to stay inside.

Military 'stay close to station' fails sometimes

I've often resorted to drafting idiot dwarves who can't seem to run in a decent direction and ordered them to stay close to their station and stationed them somewhere away from the menace and towards an escape route. However, it often feels like, despite their station being *screens* away and with a perfectly good route to trace to it, if they can see the chasing goblin they'll often turn and attack it despite being told to stay close to their distant station.

Really, the military interface needs a 'retreat to station' option. But as it is the 'stay close to station' option definitely seems to fail spectacularly on occasion.

Dwarves only run till they can't see goblins, goblins know exactly where the dwarves are

Does it make sense that dwarves only run until the goblin menace is no longer visible (sometimes as few as 2-3 squares in a heavily forested map) and then often turn back around towards the pursuing goblin, whereas the goblin has perfect knowledge of their location even if they are screens away and continues to move directly towards them? This leads to unbelievably stupid AI actions and dead dwarves who should have been able to escape. Shouldn't dwarves who spot an enemy retreat to a 'safe' location? (Meeting zone/hall/etc... report there are goblins out there? Run a marathon in the original sense? something?)

Most of the dwarves I've lost to goblins have died to their own stupidity (and often not immediately heading inside despite 'dwarves stay inside' turned on, a direction that was *also* away from the goblins).

Most of the dwarfs I've lost have been due to their own stupidity. "No, don't drink from the moat while it's filling up. Why are you running towards the goblins? How the #*%^ did you get there?"

Dwarf Fortress, it's like Sim City meets Lemmings. Dangerous Beans 21:02, 4 December 2008 (EST)

Goblins have omniscience, breaks attack AI

Similarly, if you close off all entrances to your fortress (with all dwarves inside), ambushers and siege forces simply mill around outside even if they can't see the locked doors. How can they possibly know that around 3 blind corners you locked all the doors without walking down those corridors? Shouldn't they bring battering rams/something to break down doors if they're a siege force - locking the doors should be an 'expected' response, not something which breaks the siege AI. (And of course, if you unlock the doors, your dwarves gleefully try to take tasks outside despite being ordered to stay inside...).

hi. i dont spose there's any way you can post the seed/world of your underwater fortress? i would like to play around with that :) also, sorry for messing up your pristine talk page :p Twiggie 11:57, 8 September 2008 (EDT)

Quote Page

I understand your concerns, but please try to find other methods of resolving conflict than appealing to authority. I know it's easier, and I'm sure you're frustrated about the current state of the quote page, but I'd rather not have to execute summary judgement on the quote page, which is now harder for me to avoid since I'm now involved. --Briess 16:56, 5 November 2009 (UTC)