Sandstorm !
Friday Facts #425 - Behind the legs
Re: Friday Facts #425 - Behind the legs
I like the new animations (both the "buggy" and the "not buggy"). I must admit I haver never played enough endgame to actually use a spidertron, or even make one, so I don't feel how different the new spidertron is when I don't see it side by side.
I also like also how these multi legged things give me the creeps - I really don't like whatever has more than 4 legs, especially spiders.
I also like also how these multi legged things give me the creeps - I really don't like whatever has more than 4 legs, especially spiders.
Koub - Please consider English is not my native language.
Re: Friday Facts #425 - Behind the legs
Try teaching Pentapods to walk with both legs at once.
1-3, 2-4, 5-2, 1-3, 2-4, 5-2
Like "making a wave". Should be hypnotizing.
1-3, 2-4, 5-2, 1-3, 2-4, 5-2
Like "making a wave". Should be hypnotizing.
Last edited by Aquilo on Sat Aug 24, 2024 4:17 am, edited 1 time in total.
-
- Inserter
- Posts: 23
- Joined: Fri Apr 26, 2024 12:57 pm
- Contact:
Re: Friday Facts #425 - Behind the legs
It's an unfortunate limitation, but I don't see a way around it that's performant. A gas diffusion model like Factorio uses for pollution has a lot of moving parts, especially when your pollution cloud spans as many chunks as most late-game factories do. They could improve performance by simplifying the model (for instance, they could blob all contiguous chunks containing polluting machinery into one source, sum up its pollution, and dump it on all chunks in a radius all at once) but that would remove a lot of the charm and emergent gameplay of the current cloud mechanics, and I'm not sure it would be worth it.
My money's on some kind of electroreceptive enemy, so it attacks your lightning collectors and power plants first. Maybe the tesla turret is extra effective against them because it not only attracts them, its shots also "blind" them.
Re: Friday Facts #425 - Behind the legs
How is thunder a pollution type? That's an environmental hazard, not something you produce...
I was thinking maybe cultural desecration pollution, robots getting angry at you ransacking the archaeological sites of their builders/victims
Or maybe anti-pollution? Since a lot of your resource extraction on the planet involves cleaning up oil spills and debris, maybe something would prefer the planet stay a barren death heap.
Electric or electromagnetic pollution was thrown around on some other forums but I think if the planet is already a constant storm nothing would care about your activity in that area. Maybe your EM footprint starts to disrupt whatever balance existed on the planet and the storms increase in strength, which is a similar loop to the pollution->evolution of biters and pentapods. Obviously it would still be disappointing if the storms were the only "enemy" on Fulgora as this might indicate.
Re: Friday Facts #425 - Behind the legs
Looks great, thx for sharing!StrangePan wrote: βFri Aug 23, 2024 2:52 pmSo regarding the twitching/jerking: that's my bad. The video I uploaded to FFF actually contained a bug that was later fixed, but I had forgotten this fact at time of writing. If you're curious what the Spidertron movement looks like now, I uploaded it here: https://cdn.factorio.com/assets/img/blo ... smooth.mp4fishlips13 wrote: βFri Aug 23, 2024 12:43 pmI think I agree that the old Spidertron movement looks better. Works great for the Pentapods, doesn't look nearly as good on a robot. It's kind of twitchy and jerky. It's lost some of what made mechanical and cool.
Pony/Furfag avatar? Opinion discarded.
- BrainlessTeddy
- Long Handed Inserter
- Posts: 96
- Joined: Sun May 19, 2019 7:50 pm
- Contact:
Re: Friday Facts #425 - Behind the legs
That was actually a really interesting FFF. But one thing about the spidertron. I see how you made it bend its legs instead of stretching them but the new animation kinda creeps me out. I think I like it. but we'll see
Please consider english is not my native language.
-
- Fast Inserter
- Posts: 121
- Joined: Fri Mar 02, 2018 2:55 pm
- Contact:
Re: Friday Facts #425 - Behind the legs
Ugh, it's like a rock fell off my soul. Thank you!StrangePan wrote: βFri Aug 23, 2024 2:52 pmSo regarding the twitching/jerking: that's my bad. The video I uploaded to FFF actually contained a bug that was later fixed, but I had forgotten this fact at time of writing. If you're curious what the Spidertron movement looks like now, I uploaded it here: https://cdn.factorio.com/assets/img/blo ... smooth.mp4fishlips13 wrote: βFri Aug 23, 2024 12:43 pmI think I agree that the old Spidertron movement looks better. Works great for the Pentapods, doesn't look nearly as good on a robot. It's kind of twitchy and jerky. It's lost some of what made mechanical and cool.
P.S. How about a PLD that shoots flying enemies?
-
- Manual Inserter
- Posts: 2
- Joined: Mon Mar 04, 2019 7:14 pm
- Contact:
Re: Friday Facts #425 - Behind the legs
Spiders are not evolved insects. Spiders are arachnids, and arachnids did not evolve from insects. In fact, among arthropods, you can't really get further from insects than arachnids. Arthropods are divided initially into two clades: chelicerata and mandibulata. Spiders (plus other arachnids, as well as a couple of other things like horseshoe crabs) are in chelicerata. Insects (as well as other hexapods, the myriapods, and even crustaceans) are in mandibulata. So even crabs and barnacles are more closely related to insects than spiders are.
As for the claim that the front two legs are converted from antennae, do you have a source? I've never heard that and I can't find anything that says so.
Re: Friday Facts #425 - Behind the legs
Those chonky bois are so big! It's exciting that simple walls won't be effective, those sort of combat mechanics help to make each planet feel more unique.
You can really feel the engineer's "F this planet" vibe at the end of the video. Yeah, Gleba is definitely getting paved.
You can really feel the engineer's "F this planet" vibe at the end of the video. Yeah, Gleba is definitely getting paved.
Re: Friday Facts #425 - Behind the legs
How is that better as a pollution type ? x)
I was thinking, the more you produce / use electricity, the more you charge the air, and then the environmental hazard, like thunder or sandstorm occur more frequently requiring more of those tower that capture it.
That was considering pollution for a planet where there may be "no" ennemies. Instead the adversity would be the planet itself/environment reacting to the player directly/ through weather intensifying. Rather than through more ennemies.
To compare with pollen, or smell, that can attract Gleba fauna. In other planet maybe "other harmful things" are attracted/ generated by player actions, and need to be kept in check.
-
- Inserter
- Posts: 31
- Joined: Fri Aug 30, 2019 6:19 pm
- Contact:
Re: Friday Facts #425 - Behind the legs
IMO, the old/original leg motion still looks way better. The mechanical movement more resembles a robot.StrangePan wrote: βFri Aug 23, 2024 2:52 pmSo regarding the twitching/jerking: that's my bad. The video I uploaded to FFF actually contained a bug that was later fixed, but I had forgotten this fact at time of writing. If you're curious what the Spidertron movement looks like now, I uploaded it here: https://cdn.factorio.com/assets/img/blo ... smooth.mp4fishlips13 wrote: βFri Aug 23, 2024 12:43 pmI think I agree that the old Spidertron movement looks better. Works great for the Pentapods, doesn't look nearly as good on a robot. It's kind of twitchy and jerky. It's lost some of what made mechanical and cool.
Re: Friday Facts #425 - Behind the legs
OMG, you're right. I don't know where I have my information from. It seems I picked it up in my youth and never questioned but thought it's common knowledge. Sorry. I just read the Wikipedia articles about both, and you're right, spiders are no insects. It seems I thought spiders are insects, because they appear so similar in many ways, but they aren't - they're different. The walking style I spoke of is probably somehow unrelated to this, it just depends on how many legs the creature has.arcosapphire wrote: βFri Aug 23, 2024 10:49 pmSpiders are not evolved insects. Spiders are arachnids, and arachnids did not evolve from insects. In fact, among arthropods, you can't really get further from insects than arachnids. Arthropods are divided initially into two clades: chelicerata and mandibulata. Spiders (plus other arachnids, as well as a couple of other things like horseshoe crabs) are in chelicerata. Insects (as well as other hexapods, the myriapods, and even crustaceans) are in mandibulata. So even crabs and barnacles are more closely related to insects than spiders are.
As for the claim that the front two legs are converted from antennae, do you have a source? I've never heard that and I can't find anything that says so.
Re: Friday Facts #425 - Behind the legs
How long will it take before the 5 foot walkers are modded into pink spongebob dude?
- GregoriusT
- Filter Inserter
- Posts: 315
- Joined: Wed Apr 10, 2019 6:42 pm
- Contact:
Re: Friday Facts #425 - Behind the legs
you have just sped that one up drastically.
Don't underestimate Landmines!
Biters bite, Spitters spit, Spawners spawn and Worms... worm? - No, they throw their vomit! They even wind up to directly hurl it at you! friggin Hurlers...
Biters bite, Spitters spit, Spawners spawn and Worms... worm? - No, they throw their vomit! They even wind up to directly hurl it at you! friggin Hurlers...
Re: Friday Facts #425 - Behind the legs
I think the FFF walker video should be fixed with the new one, at least edit the post to remove the video. That looks horrible and a lot of people might not read it or will go fixated on that and not anything new.
Also I agree that laser's should be able to target the Bio-Missiles to cause them to spawn their squid early.
Also I agree that laser's should be able to target the Bio-Missiles to cause them to spawn their squid early.
- Skellitor301
- Fast Inserter
- Posts: 152
- Joined: Mon Aug 04, 2014 10:04 pm
- Contact:
Re: Friday Facts #425 - Behind the legs
While reading about the search algorithm for leg placement, something came to mind when looking at the different patterns used and the solution used. Is the heading of the spidertron being taken into account with this search algorithm? Cause I can see an argument be made for a bit of extra efficiency when the direction of travel influences the search algorithm and the area it searches.
One idea that comes to mind when considering this is instead of using a branching search, I figured something like a raycast to certain points would probably make finding a viable location easier with a bit of logic put behind how the raycast search is performed. The idea that came to mind was start with doing a raycast of the heading and seeing if a collision is made. If it is then step the leg group like normal in that heading.
If not then I'd utilize the cardinal and intermediate directions, i.e. N, NE, E, SE, etc. So we'd start by checking 0, 45, 90, 135, 180, 225, 270, and 315. Utilizing the heading of the spidertron, you can determine which of these directions to start based on which one is closer to this heading. Example, if the spidertron is traveling with a heading of 156 the closest direction it would start with is 180. The raycast would check the distance from the center of the step to the furthest a leg could go with the spidertron stretching out. If there's no collision detected, it would go to the next closest heading from the initial heading 135, check that, and keep going until either a collision is made or all 8 directions don't find a result. Since most cases at this point are across shores this would cut back search substantially to generally 4 searched directions, at most 8 (at the current stage) depending on how far away from the heading the collision is. At that point, letting the spidertron step out to the usual distance it would step in that direction would be ideal, if it's a back leg then the grid coordinate closest to the collision location. But we also do have other unchecked coordinates in the 4x4 quadrant we've identified as having a viable location.
If we want to further refine this, we essentially will reduce the grid search size from the 8x8 to ideally 1/8th of that size by cutting the grid into 8 pie sections aligned to the 8 searched directions. We determine the slice to search by having one of its edges be along the direction the collision was successful in, and the other edge closest to the initial heading. We can then use 2 or 3 more raycasts to refine the direction pointed from the search center point and the 3 edge points of our initial grid, aka the crust of our pie slice . We would start with the middle raycast and then the next one from that closest to the initial heading, and if we find less than 2 collisions then we check the final raycast for any collision. We should now have at least 1 successful raycast collision coordinate at most 3. At this point, we basically have the collision closest to the initial heading of the spidertron and we check the furthest point along the raycast that is viable for the spidertron to step and choose the closest grid coordinate.
In the rare instance of the 8 initial raycasts failing to collide with anything, we perform the checks done before that point again only we're shifting the 8 points we are checking to the ones located half way between them. We essentially will continue the checks like normal from that point until a viable location is found, only we can safely ignore the already checked directions that failed and reduce the pie slices step in half. If all fail, then essentially we don't have a viable way to move forward.
I know that seems quite a lot, but if the logic is properly followed then a lot of the steps can be ignored or we find a viable spot to step early and we can skip the rest of the logic at that point. Yes it's a logic system that will build on itself, but overall on paper it should perform far less checks on average.
Steps in psudocode:
One idea that comes to mind when considering this is instead of using a branching search, I figured something like a raycast to certain points would probably make finding a viable location easier with a bit of logic put behind how the raycast search is performed. The idea that came to mind was start with doing a raycast of the heading and seeing if a collision is made. If it is then step the leg group like normal in that heading.
If not then I'd utilize the cardinal and intermediate directions, i.e. N, NE, E, SE, etc. So we'd start by checking 0, 45, 90, 135, 180, 225, 270, and 315. Utilizing the heading of the spidertron, you can determine which of these directions to start based on which one is closer to this heading. Example, if the spidertron is traveling with a heading of 156 the closest direction it would start with is 180. The raycast would check the distance from the center of the step to the furthest a leg could go with the spidertron stretching out. If there's no collision detected, it would go to the next closest heading from the initial heading 135, check that, and keep going until either a collision is made or all 8 directions don't find a result. Since most cases at this point are across shores this would cut back search substantially to generally 4 searched directions, at most 8 (at the current stage) depending on how far away from the heading the collision is. At that point, letting the spidertron step out to the usual distance it would step in that direction would be ideal, if it's a back leg then the grid coordinate closest to the collision location. But we also do have other unchecked coordinates in the 4x4 quadrant we've identified as having a viable location.
If we want to further refine this, we essentially will reduce the grid search size from the 8x8 to ideally 1/8th of that size by cutting the grid into 8 pie sections aligned to the 8 searched directions. We determine the slice to search by having one of its edges be along the direction the collision was successful in, and the other edge closest to the initial heading. We can then use 2 or 3 more raycasts to refine the direction pointed from the search center point and the 3 edge points of our initial grid, aka the crust of our pie slice . We would start with the middle raycast and then the next one from that closest to the initial heading, and if we find less than 2 collisions then we check the final raycast for any collision. We should now have at least 1 successful raycast collision coordinate at most 3. At this point, we basically have the collision closest to the initial heading of the spidertron and we check the furthest point along the raycast that is viable for the spidertron to step and choose the closest grid coordinate.
In the rare instance of the 8 initial raycasts failing to collide with anything, we perform the checks done before that point again only we're shifting the 8 points we are checking to the ones located half way between them. We essentially will continue the checks like normal from that point until a viable location is found, only we can safely ignore the already checked directions that failed and reduce the pie slices step in half. If all fail, then essentially we don't have a viable way to move forward.
I know that seems quite a lot, but if the logic is properly followed then a lot of the steps can be ignored or we find a viable spot to step early and we can skip the rest of the logic at that point. Yes it's a logic system that will build on itself, but overall on paper it should perform far less checks on average.
Steps in psudocode:
Click to view
Likely to have some mistakes in the psudocode but I hope it gets the idea acrossclick to view image
Re: Friday Facts #425 - Behind the legs
It was joke, I put smiley faces on the end.
And yeah I can see them going with the EM pollution increasing storm power thing, but again that would be rather disappointing. Since you can 100% lightning proof a part of your factory with a lightning rod, which are a passively functioning building, the obvious and trivial solution would be to simply spam them everywhere and then pollute to your heart's content.
That would obviously not be very exciting, so I hope that even if that is a thing there are still actual enemies you have have to shoot.
Re: Friday Facts #425 - Behind the legs
I did that too !
Well often times players suggestions are quite disappointing when comparing to what the combined power of the devs can create . I thought last time that biters on Fulgora could be a bunch of robotic spiders that would be responsible for the extinction of the previous life form and attack players based on electricity, like pollen here. But since we have Pentapods here, i don't think so anymore. I think it would be too similar leggy creatures. But maybe the leg logic is used somewhere else in diguisePhauxstus wrote: βSat Aug 24, 2024 5:01 amAnd yeah I can see them going with the EM pollution increasing storm power thing, but again that would be rather disappointing. Since you can 100% lightning proof a part of your factory with a lightning rod, which are a passively functioning building, the obvious and trivial solution would be to simply spam them everywhere and then pollute to your heart's content.
That would obviously not be very exciting, so I hope that even if that is a thing there are still actual enemies you have have to shoot.
I could go with it, but also with maybe a flying or a swimming one ? Not sure about a burrowing one, as it was revealed no discontiguous pathfinding 'yet'. Maybe each planet has ennemies with a different number of legs x).
Re: Friday Facts #425 - Behind the legs
Could they not keep the current simulation, but could they not divide it's frequency, i.e. often it does the calculation, by the number of "pollutants", then just simulate one at a time? Or would that not amount to the same savings, due to memory-stuff? (I figure it doesn't take much space in memory, but it could be enough to matter)jamaicancastle wrote: βFri Aug 23, 2024 9:07 pmIt's an unfortunate limitation, but I don't see a way around it that's performant. A gas diffusion model like Factorio uses for pollution has a lot of moving parts, especially when your pollution cloud spans as many chunks as most late-game factories do. They could improve performance by simplifying the model (for instance, they could blob all contiguous chunks containing polluting machinery into one source, sum up its pollution, and dump it on all chunks in a radius all at once) but that would remove a lot of the charm and emergent gameplay of the current cloud mechanics, and I'm not sure it would be worth it.
My money's on some kind of electroreceptive enemy, so it attacks your lightning collectors and power plants first. Maybe the tesla turret is extra effective against them because it not only attracts them, its shots also "blind" them.