Bay 12 Games Forum

Very cool.

Notes:
Currently, files with the specific merging (init.txt, d_init.txt, interface.txt) are overwritten in the writable branch every time the file is opened in the merged filesystem and any comment is discarded.

Do you check the files' modification dates?

• Other suggestions?

You could possibly support a three-way merge on raw files. My raw file merges almost always proceed without any hiccups. (Though that "almost" gives me pause. I'm not sure what would be appropriate to do in the event of a conflict, for a FUSE plugin.)

What do you guys think of

multiplying aspects (attributes * traits * skills * preferences) %'s / 100 / 100 / 100 / (product of weights)

I've only been following this casually, and it's not clear to me exactly what you're trying to calculate in the end (suitability for a role, I gather, but the precise meaning of the numbers is unclear), but I very much doubt that this is what you want. It doesn't appear that the weights will make any of your factors more important in the result than any other. You get the same result whether the weights are (.5, 1, 1.5, 2) or (2, .5, 1, 1.5). If you're looking for a kind of multiplicative average, you probably want something like the geometric mean, which is the nth root of the product of n terms. A weighted form would be to take the product of each factor raised to the power of its normalized weight:

  A^Wa * B^Wb * C^Wc * D^Wd

where Wa + Wb + Wc + Wd = 1. Though I suspect this won't be what you actually need, either.

Quote from: Trouserman on October 12, 2014, 07:14:48 am

I know DFHack has drainaquifer. Is the reverse possible? Could a layer be turned into an aquifer layer?

Yes.  On the map block level, there are three flags to set: flags.has_aquifer, flags.check_aquifer, and designation[x%16][y%16].water_table for each tile you want to turn into a water source.  In addition, df.global.world.raws.inorganics[].flags.AQUIFER for that tile's rock/soil type needs to be true.  (Or perhaps the tile's layer type; I haven't experimented with veins and inclusions.)

Even after all four flags are set, it might take a few steps for the new aquifer to start leaking.

It also appears that the drainaquifer script might be slightly wrong; it can miss aquifers that happen to miss the two tiles it checks, and it ignores the two block-level flags.  Is that a problem in practice?

Thanks. I had to dig around for how to access some of the relevant data, but this was enough for me to put together a toggleaquifer script that suits my purposes. It seems that the rock/soil type does not in fact need to be tagged as aquifer for this to work, and in particular the veins and inclusions can be made into aquifer. So, it was necessary to test that the tiletype is not MineralWall. I also added checks for surrounding open blocks, so that exposed edges don't begin to leak. (Though it is amusing for a slope to just start leaking all over the landscape.)

Presumably, the drainaquifer script is checking the corner and center of the block as a quick test if there are any aquifer tiles in the block. It should probably be checking the block-level flags, instead. And then, as you note, turning them off. (I suspect the only harm leaving them set does is causing some small amount of lag from extra checks for exposed aquifer cells. I'm not sure what has_aquifer does; check_aquifer seems to be necessary and sufficient for aquifer tiles in the block to function.)

Edit: I spoke too soon; check_aquifer doesn't apply to the aquifer tiles within a block, but to the tiles in the block which are adjacent to aquifer tiles. That makes a rigorous script for altering the aquifer status of individual layers a bit more complicated. Just when I thought I was done.

Would it be feasible to make a way to add space to an existing stockpile? Either drawing new areas onto a pile, or merging two piles into one, would be quite handy.

You might want to make vs local ("local vs =" instead of just "vs =") there, since it's in a function and all. Just good practice in programming.

So, Lua isn't lexically scoped by default? I haven't really had a look at the language yet, but that's... disappointing.

There may be other ways to accomplish the deed as well, but I'm not sure what they are.

Falling asleep on the stairs, maybe?

Even though falling down stairs is a rare enough event that using wood probably won't make a practical difference for safety, this does make me wonder if flooring over splatter pits with lead or silver would make them more effective.

If you have enough soil layers to make a 5z chamber, but trees require additional layers of soil beneath them, you could drop a few layers of soil to a lower level with a cave-in.