The other thing is the micro involved - every few minutes you need to come back to the farm to check on the GSC and silk unless you figure out a way to fully automate (which I haven't). It's not at all like setting the tasks to repeat once you have a large enough stockpile of each item in the process. While working on a ten-level gate, having to check back at intervals did get tedious.
I used a system which includes a 3x3 (I guess 2x3 would work too) island on which the animal stands with the rest of the room made of bridges over a pit 4+ Z-levels deep. Next to the restrained target is either one or two repeating pressure plates tied to the bridges of the room. This causes any web that's either shot over the empty space (before the bridge closes e.g. appears) or shot onto the bridge (before the bridge opens e.g. disappears) to fall into the pit. You get a little waste because web over a space will fly around randomly and possibly fall into otherwise inaccessible areas (e.g. within LoS of the spider/your hostile target if the spider's tamed), but more than 95% will fall into the pit for your workers to collect without having to see the the hostile or the spider. You can stop the production by, as you mention, forbidding passage through one of the doors.
There are a several minor issues with how I went about it. The biggest is that, in my initial design testing, I let my tamed spider kill a few hostile bait targets. This seemed to have the effect of making me unable to assign it to cages, restraints, or pits; it is still listed as hostile in my unit screen (or was until its name because so long that it covered it status), though it doesn't attack friendlies, nor do dwarves run away from the spider when chaining the bait target. The major problem with this issue is that I can't move my spider closer to my main fort, or alter the design of the room without fear the spider will leave its current 2x1 room. I would have to cook up some convoluted system to knock the spider unconscious on a cage trap to move it, but this additionally has problems in that I'd need to get into its room to build the traps, which might free it again.
A secondary problem is that the silk collection is pretty random. Web collectors don't choose the nearest web to them, frequently traveling through several tiles of webs to reach one in the upper right corner, when the entrance is the lower left (probably due to the main fort being SE of the room, but if they were picking the nearest web when they accepted the job, they should be clearing from south to north either way). In any case, this doesn't cause a problem with web collectors since they can travel through webs without getting stuck, but if they abandon the task and drop the tread, a hauler will come waltzing through and trample through several tiles to pick up the thread, often stomping out dozens or even hundreds of webs. Even so, this problem is minor since if you let the system run, you're literally generating thousands of webs per season.
A tertiary minor problem is that it is possible the dwarf who is chaining the bait animal will not move quickly enough before the bait animal activates a pressure plate and causes the bridges to open. I didn't have any issue with this since only my extremely agile dungeon master has animal training and will move hostile creatures, but if you have slower dwarves they might break their legs trying to tie up the hostile bait target.
I should note that someone else describe the idea of using bridges to form a collection pit in another thread about GCSS farming, it is not my own creation.
Author
Topic: Cloth industry worth it? (Read 2878 times)