Print

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

[Draco18s]

Or we could go "it looks like ore!" and be quiet about it.

[Slumb]

I think there are bigger and more fun things on the to-do list than improving the randomness of DF.

[Greiger]

I have noticed the lines of symbols in unmined rock as well. Usually way way down at th bottommost z-level.  I actually used to think it was some kinda hint at underground features, because the first one I searched thoroughly came up with a chasm.

But I'm not complaining.  This is an alpha as had been said many many times before.  The randomness is good enough for my enjoyment.  Though I suspect the OP was noticing it may show itself more blatantly sometime in the future.

Just as a side note how did that guy in the linked-to tread get the features to show up on the local map in the fortress screen?  Anyone know?
EDIT: Never mind I heard of Dwarf Prospector But I was unaware it was that fancy. ^^ Nice work 0x517A5D.

[ January 04, 2008: Message edited by: Greiger ]

[Armok]

After reading this I have actually been converted.

This it something that needs urgent fixing, and one of the VERY few tings that is actually separate enough from game play to be done by someone except Toady, at least/especially the seeding if I understood things right.

The streaks in unmined space is something obvious and visible, but in how many places does this cause problems to players indirectly whiteout being noticed?
For example I would not be surprised if some of the weird mining behaviors was involved in this.

I really don't know anything about this more than what is said in the first post, but it seems fairly obvious now that this is a problem that needs fixing, albit not the highest priority one.

[WillNZ]

quote: I don't understand most of this thread. I feel so un-nerdy... I wish I was an uberhax0r too.

You say that now...

[Eagle of Fire]

quote:For example I would not be surprised if some of the weird mining behaviors was involved in this.

My miners does not exhibit weird mining behavior(s). They, just like in 23a, starts from the upper left and go down to the lower right map as preference, then always choose a square next to them to continue mining.

[sphr]

I think the run-time simulation behavior is not the main concern.  What I feel is more of a concern is the world and map generation.  Basically, I think everybody agrees that a need a consistent world generation system, which should not be affected much by things such as new features or random number problems.  In that respect, I think low level details about random number handling will be of concern.  Frankly speaking, this can work as long as the generation and use schemes are constant, no matter whether the PRN sequence is "faulty" or not.  the hard choice is just: live with the knowledge of potential problem as such, or make a breaking (in the world generation sense) change.  But when ToadyOne makes another breaking change to the world generation part, this is something "extra" that he might just want to update, not that it will cause any end player to notice the difference (since the change of world generation is the big differene that's catching all the attention).

@ToadyOne
not sure about this as I haven't tested the 33g world generation, but are the new changes like adding of gods causing the world generation to be totally different due to different order of accessing random numbers?  If that is the case and you are interested in a way of making the random number generation consistent so that say even after you add a new layer like gods and religions (and having to generate all the names and spheres and such during WG), stuff like maps and stuff can still stay the same, decoupling the original RNG from dynamic use could help.  

In one para, it is a method which uses the root RN generator to only generate seeds for other random number generators.  If there is something dynamic (e.g. like generating names for a random number of creatures), a single static seed is generated from the root generator, and that seed is used in another one-use generator to generate all the names.  The root generator only generates 1 number (thus static).  Ideally, if say the terrain generator and other generators are made independent, it means that adding new things like gods will not affect map (since they use independent random number generators seeded once by the root generator)

Can discuss/clarify in another thread if you are interested.

regards

[Sean Mirrsen]

I don't know if it's related or not. It's probably part of the simplification used to speed up and/or simplify the world generation process.

When I look around for mountainous sites, I frequently scan around with cliff display on. Thus, I have noticed large amounts of completely identical (height-delta wise) map segments adjacent to eachother. It doesn't necessarily mean that the actual height map is the same (and of course not in the presence of rivers), but the fact itself is disturbing, especially in this light. With a more refined RNG we could have a lot more diverse worlds.

[Fenrir]

quote:Originally posted by WillNZ:
<STRONG>You say that now...</STRONG>

Waddaya mean?

[Toady One]

Anybody who's concerned about the streaks of characters in the hidden areas should reread the OP's second post and keep in mind that this is an isolated case.  It has absolutely nothing to do with the main RNG used by the game.

This second post is also related to decompiled/reverse-engineered closed source code.  I know I haven't really brought this up when people have been modding and so on, and I've helped people with structure definitions and so on in the past, but at the same time, I don't want to have to engage in a defense or make apologies any time I code something.  Sure, the code should be defensible, but that kind of scrutiny is aggravating and makes working on the game a miserable experience, in part because I am forced to deal with the snark and hyperbole that often accompanies these discussions.

Sean:  Nothing to do with the RNG.  It has to do with the corner selection, which isn't random for the world map, and there's a Req for it some place.  I'd have added a random element already, but doing so might break some assumptions of other parts of the seamless map code, so I'm being cautious about it.  It uses random corner selection at the next level down, and I don't remember off the top of my head what I was worried about on the world map when I set it up at first.

sphr:  I've used a similar method for the seamless maps, and I did something like that for the pantheons to try to keep the current seeds working for at least another version, but I haven't set anything like that up for the basic steps of world generation.  It would help a bit, but it's still not a guarantee because of rejections.  If I change something that causes a rejection to go away at any of the steps, you can get an entirely different world than you got in a previous version, because it picks an earlier world.  Similarly, if the world fails at one of the steps in the new version, then it'll have to roll up another one, regardless of how you set up your seeding.

[Sergius]

quote:Originally posted by Eagle of Fire:
<STRONG>
My miners does not exhibit weird mining behavior(s). They, just like in 23a, starts from the upper left and go down to the lower right map as preference, then always choose a square next to them to continue mining.</STRONG>

In the 2D version, dwarves would always prefer to continue mining an adjacent tile even if it was against their "upper left" preference, as long as it was connected to the current one. So digging a tunnel half a mile to the west to connect to another open area was fairly linear (the dwarf just chose whichever end he preferred, unless the player manually forced him to start on the eastern end). Then as he got to the western end, he would just mine the damn tile out. As it is now, they go to the eastern end, dig a tunnel, stop one tile short of connecting the two areas, go all the way to the eastern tunnel exit, go all the way around from the outside to the western end, and dig it from the west side.

[qwertyuiopas]

There are still uncountable numbers of worlds, you junt won't get many of them if the world seed is set to random. An external seed selection utility to make the choice for you would let you get the most worlds with the least effort from you or Toady One The Great.

[Eagle of Fire]

Sergius: As I said, my own miners don't exhibit strange mining behaviors. I suggest that you give them higher wages, they should stop playing mind tricks on you.  

[ZrO4]

quote:...makes working on the game a miserable experience...  

Will there be any '0x517A5D' leather shoes made?

[Toady One]

ZrO4, the thread doesn't need to go that direction.