For Questions That Don't Merit Their Own Thread

[kerbiloid]

12 minutes ago, magnemoe said:

In an static defense ammo is not an major issue.

When you have an ammo factory in dungeon or a regular supply transport.
But not when your stronghold gets isolated.

Also the more numerous are moving parts - the more headache with their support, the more staff you need.

Beam weapon doesn't need elevators, storehouses, etc.
And you always can use your energy source for civil needs.

Edited August 1, 2018 by kerbiloid

[Terwin]

On 7/31/2018 at 5:57 AM, kerbiloid said:

Well, the main argument against underwater civilizations is no flame, no metals.

 

On 7/31/2018 at 6:02 AM, ARS said:

Agreed, and it's very hard to advance if there's no metals since it's a vital resource for developing civilization

The ability to cook is probably more telling than the lack of metals.

In addition to killing off things that will make you sick, cooking food breaks it down so that it is easier to digest and more energy/nutrition can be extracted with less effort on digestion.

This extra energy is important for sustaining things like a large brain with high-energy demands , and also lets us out-compete other animals without being more specialized than they are because we get more energy per unit of food consumed.

[Diche Bach]

15 hours ago, cubinator said:

Really makes you wonder if there's really some monolith sitting around for five million years or something...either way, if we are the only ones I hope we don't blow it for ourselves just yet. And I hope there are lots of creatures out there with great and wondrous abilities, even if none of them have decided that it's worth it to invent, tinker, and wonder to the point of expanding into the stars.

I like to hope that we ARE the "Precursors" that are so commonly portrayed as being the "original" high-tech "lost civilization" that left traces of their intergalactic web in so many flavors of sci-fi, and (to insure that sentience never dies) left retro-virus like codes in simple biota seeded across the multi-verse so that it would eventually, billions of years later evolve into all the humanoid-esque alien forms . . .

We are just infantile Precursor's still!

When we can at the right moment in our organic life cycle undergo the "The Uploading" and become one with the galactic information network, we will have arrived at early adulthood 

[kerbiloid]

55 minutes ago, Diche Bach said:

We are just infantile Precursor's still!

And the first thing which humanity did once getting to space - killed a (she-dog) with ICBM.

[magnemoe]

17 hours ago, Diche Bach said:

In seriousness though . . . I know a reasonable amount about primate psychology, a smidgen about canine psychology, and a tiny bit about animal psychology in general. From my standpoint, most prevailing notions/models/theories of "intelligence" are quite stupid because, not only do they focus only on a limited range of human mental function, they focus almost exclusively on human mental function.

Bat echolocation; geese migrating; cetaceans doing all their crazy long-distance stuff out in the oceans; fungi networks communicating up and down the entire length of the Appalachians . . . obviously none of these things are quite the same as our unique human ability to grasp symbols and syntax and combine them in creative ways to exchange mental states with other human beings. But how other animals manage the wonders which they do with ease is largely obscure to us, and much of our own thinking is clouded with imperfections.

The fact that we evolved into "sentient" beings might well have never occurred anywhere else in the entire history of the universe, and might never occur again. That is how random and unlikely it looks to me based on my understanding of what is well established about human evolution and evolutionary psychology.

Think sentient is kind of an long shot because it takes plenty of effort before it pays off and you need some pretty strict settings for it to work. 
Humans was not really successfully until 40-50K years ago, although erectus family did pretty well and spread wide, they might have evolved more smart species had we and the Neanderthals died out. 
On the other hand smart social animals are also pretty new, we might getting more going for it in deep time but no way to know. 
Still its an obvious great filter. 
I wonder if civilization is another one. Not for humans as we tend to create them everywhere, we like large scale organizations, a bit more like social insects than typical mammals here.  Would some species who used smell as main way to identify each other do so? Or would an small city break down as too many people in it. 

[ARS]

Is it possible to create a suppressor for tank gun? How practical is it?

[p1t1o]

1 hour ago, ARS said:

Is it possible to create a suppressor for tank gun?

Yup. No reason it cant be scaled up. But as Im sure we all know, a suppressor only reduces the sound, I'd imagine that a tank main gun will still be pretty loud, unless you used subsonic ammunition with a reduced charge, which would totally defeat the object of being a tank.

1 hour ago, ARS said:

How practical is it?

Not very. The report of the gun has never been the main weakness in tank stealth. And to take full advantage of a supressor, other changes would be needed which would drop the effectiveness of the tank as a whole (as above). On top of that, an effective supressor would be large and heavy, lowering mobility and ironically, reducing stealth.

This is how main battle tanks are used, in a very simple nutshell - 

• Armoured charge/manouvre, firing on the move.
• Defensive line, firing from fortified positions.

Neither of those things would be enhanced with a suppressed gun - bottom line? 15kg of tungsten hitting a neighbouring tank at Mach3 is a dead giveaway that there is a tank shooting at you.

[kerbiloid]

5 hours ago, ARS said:

Is it possible to create a suppressor for tank gun? How practical is it?

And use another tank to carry it.

Spoiler

 

Edited August 3, 2018 by kerbiloid

[magnemoe]

27 minutes ago, kerbiloid said:

And use another tank to carry it.

  Hide contents

 

Great balls of fire, and yes has seen it before but only the first image. 
And it would not work anyway, the sound would be very noticeable inside the range of an tank gun, the flash is also very noticeable at night. 
Just the engine on tanks are hard to hide. 
On the other hand as projectile is supersonic you will not hear the first one before it hits

The real purpose of most suppressor is to reduce noise level to save the ears of the shooter or to keep inside noise requirements. 
The above suppressor is to reduce the noise of artillery cannons probably as this is at an factory or service station with people living nearby and you want to test and zero the guns without going to an artillery training range hundreds of km away. 
The Hollywood silencer who give an poop is pretty much an myth. Yes with an  large suppressor and an subsonic bullet and an gun designed to be silent it can work in an noisy place. 

The only weird thing is why is has camouflage paint. On the other hand they probably ordered soldiers to paint it with no other instructions. 
We was ordered to paint all vehicles in summer camouflage paint, pretty much the same as above while painting the cars and trucks we also painted two bicycles and an handcart who was inside the building  

[kerbiloid]

He knows everything about camouflage.

Spoiler

We had a similar one. Between tree roots he was invisible.

Edited August 3, 2018 by kerbiloid

[wumpus]

On 8/3/2018 at 4:32 PM, magnemoe said:

The only weird thing is why is has camouflage paint. On the other hand they probably ordered soldiers to paint it with no other instructions. 
We was ordered to paint all vehicles in summer camouflage paint, pretty much the same as above while painting the cars and trucks we also painted two bicycles and an handcart who was inside the building  

If it moves, salute it.  If it doesn't move, pick it up.  If you can't pick it up, paint it (camouflage).

Already camouflaged, do not paint.

[magnemoe]

4 minutes ago, wumpus said:

If it moves, salute it.  If it doesn't move, pick it up.  If you can't pick it up, paint it (camouflage).

Already camouflaged, do not paint.

That explain the sergeant response to we painting his bicycle. It could obviously be picked up but was still painted  

Animal should not be painted as it animal abuse. Use hair coloring kits instead.  

[kerbiloid]

Spoiler

 

Edited August 5, 2018 by kerbiloid

[Gargamel]

On 8/3/2018 at 11:40 AM, p1t1o said:

15kg of tungsten hitting a neighbouring tank at Mach3 is a dead giveaway that there is a tank shooting at you.

Yeah, by the time a tank is firing, the time to discover it and deal with it 'safely' has passed.  

The resulting explosions, from a hit or a miss, would be a gentle clue that something is shooting at you. 

[magnemoe]

On 8/5/2018 at 8:14 PM, kerbiloid said:

  Hide contents

 

If some fat guy sit on you its your own fault. 
And yes i'm an master in camouflage, in the army i camouflaged my helmet, put it on the ground and needed help to find it
Sergeant was not amused

[tater]

 

 

[p1t1o]

36 minutes ago, tater said:

<snerpt>

That is gold.

Does anyone else now have a collection of high-res photos from this site that they are keeping to one day have printed off and framed?

Cant wait to move into a house where I can choose what to nail to the walls...

[Cunjo Carl]

Hey! Any chance someone here's tried out WASM? I was curious about the performance of its switch statements for ~50 cases? The code would be compact, so instruction cache misses shouldn't be a big deal, and I know YMMV with branch prediction and all, but I was hoping someone could give me a general feel of if it's  'zingy' or 'better if you didn't.' I was really crossing my fingers they would decide to toss in jump tables in the end, but now I'm wondering if I can just kludge it with a switch. The portability of WASM is really enticing, as well as its ability to tie in with JS (and its 10billion libraries) which is why I was hoping for it in specific. Thanks in advance!

[K^2]

Any sensible implementation will JIT it, much like JS code, so performance of a switch on WASM shouldn't be dramatically different from native.

There are two things to keep in mind. First, WASM doesn't really outperform JS by a whole lot, because JS itself already runs very close to native. On a good engine, JS has about 20% performance penalty on average compared to native, and WASM falls in the same ballpark. The reason is that no modern engine will run raw JS, but rather transform it to intermediate, optimized representation. C++ to WASM compiler will take exactly the same steps. And then both WASM or JS intermediate will be JITed to native on run time.

The advantage of WASM is primarily that you don't need to waste time parsing, transforming, and optimizing JS. All of this has been done in advance, and WASM is ready to be executed directly. Also, you can potentially compile to WASM from just about anything. C/C++ being the most exciting, IMO. This allows for greater flexibility and greatly improves time to interactive and memory overhead, but raw execution performance is not considered a benefit.

The second part is that WASM is going to go through a lot of changes in the near future. Consider the current iteration as something of a beta test. Everyone seems to be pretty happy with core instruction set, so I wouldn't expect any great changes, but memory layout and bindings are going to change dramatically. Currently, WASM expects a block of flat memory fed to it, and binding it to JS is awkward. So if you are planning a big project that relies on WASM, it might be wise to wait a bit to see how these things resolve.

Edit: Compiler support might also improve in the future. Currently, the two options I've found is to either compile via LLVM, which is a bit of a chore, or use EMScripten, which still uses LLVM under the hood and introduces some bloat. Hopefully, WASM support in Clang will get improved, and maybe other compilers will catch up.

Edited August 13, 2018 by K^2

[Cunjo Carl]

8 hours ago, K^2 said:

Any sensible implementation will JIT it, much like JS code, so performance of a switch on WASM shouldn't be dramatically different from native. [...]

There are two things to keep in mind. First, WASM doesn't really outperform JS by a whole lot, because JS itself already runs very close to native. On a good engine, JS has about 20% performance penalty on average compared to native, and WASM falls in the same ballpark. The reason is that no modern engine will run raw JS, but rather transform it to intermediate, optimized representation. C++ to WASM compiler will take exactly the same steps. And then both WASM or JS intermediate will be JITed to native on run time.

Thanks for the word!

Awww... That'd be too bad to hear if WASM's switches won't be improving over JS' JIT... I was hoping to atleast get compiled JS speeds out of it. I really think that's the biggest drawback of using JS for a ton of applications, is how idiosyncratic and poorly documented its profilers and compilers are. One wrong move and a whole function gets slapped into the 1% native run speed morass of the JIT interpreter on atleast one of the common browsers. Switches and dispatch functions are the most problematic for sure! Even put within a very simple function with nothing nasty to worry the profiler or compilers (in Firefox for preference for me) and with everything being best case scenarios (like consecutive ints 0-100, no fallthroughs, no cache thrashing, etc) it took ~~80ns per switch. Putting in a nearby eval() to force JIT, didn't make it that much slower, which kinda shows how slow it can get. For everything else though, I'd be a happy camper with WASM running at compiled JS speeds (optimized or baseline).

I was really hoping WASM would bring in assembly abilities, like jumps and prefetching! The WASM community seemed really adamant about having only simple code control though, offering only switches as non-binary control. Come to mention it, do you happen to have any idea how WASM's going to emulate C code if it can't do jumps (for the gotos)? It made sense that they could play some tricks with WASM's switches as a hack when compiling, but if they're going to do JIT with any reasonable efficiency, they'll need jumps. Back when I was poking at it (early 2017), the general consensus seemed to be that only bad programmers need the code control of jumps or jump tables anyways, and that jumps would make the syntax ugly. For my case, they weren't wrong on any account , but I'd like to see how they're going to handle this! Maybe I'll get lucky and they'll breakdown and put 'em in after all. I'm mostly just chatting, I can't code at the moment (or even browse more than a hair), but after seeing the 'lexing' thread and I came to wonder if the old roadblocks from that project had been cleared. Sounds like not yet, unfortunately.

It's nice to hear the WASM folks are really pushing the whole C/C++ JIT emulation thing though! It wasn't in the plans when I was poking at it, but I think that'll do them well. The UI library support of JS meeting the functional library support of C/C++ should make a killer duo!

 

8 hours ago, K^2 said:

The second part is that WASM is going to go through a lot of changes in the near future. Consider the current iteration as something of a beta test. Everyone seems to be pretty happy with core instruction set, so I wouldn't expect any great changes, but memory layout and bindings are going to change dramatically. Currently, WASM expects a block of flat memory fed to it, and binding it to JS is awkward. So if you are planning a big project that relies on WASM, it might be wise to wait a bit to see how these things resolve.

It feels like it's been in beta for ages! I'm happy with feeding it static memory, but admittedly the setup to even get a "Hello World" out of it was pretty non-trivial. I guess I'll wait another year and see what they do. Thanks for the update and advice! Cheers.

 

[K^2]

33 minutes ago, Cunjo Carl said:

Come to mention it, do you happen to have any idea how WASM's going to emulate C code if it can't do jumps (for the gotos)?

Probably a switch wrapped in a loop. Just a guess, though. The only way I can think to test it is to write a program with gotos in C, compile it to WASM with LLVM tools, and then disassemble the output.

P.S. Depending on what the optimizer does, WASM switches might avoid a lot of the worst-case problems that show up in JS, so you might gain some performance in this instance. But I haven't seen any benchmarks for this. Benchmarks I've seen are for fairly generic code.

Edited August 13, 2018 by K^2

[YNM]

Is it normal to have an earthquake swarm this large ?

Search parameters :

Main swarm (>5.5 magnitude)

Spoiler

All swarm (>3 magnitude)

Spoiler

 

Edited August 19, 2018 by YNM
Replaced urls with screenshot. The search parameters isn't carried on.

[Green Baron]

Not sure what you mean. That does not look like a single swarm, as in timely and locally connected earthquakes assumed to originate from the same source. But "Swarm" is not a strictly defined expression. In your links different sources are listed and i overlooked a time scale, i see subduction zones, diving slabs, lateral plate movement, transform faults, intraplate volcanism and hot spots all over the world.

If you can give as a more precise time span and source of what you mean we can could try to look up more info.

But principally, single earthquakes as well as connected swarms, some of them even sketching a path of movement or the orientation and location of the area of origin, happen every time all over the world.

Btw.: Mag 2.5 is hardly noticeable if it is not shallow and directly under your feet.

Edit: or, in more direct words: which one do you mean ? :-)

Edited August 19, 2018 by Green Baron

[YNM]

4 hours ago, Green Baron said:

If you can give as a more precise time span

The past 30 days.

I mean, these news, if you haven't known of them :

https://www.bbc.co.uk/news/world-asia-44996035

https://www.bbc.co.uk/news/world-asia-45124972

https://www.bbc.co.uk/news/world-asia-45238018

I'm just wondering, I know relative-energy-wise it's still not as strong as a single 8 magnitude quake, but is that possibly a normal thing to happen every now and then ?

EDIT : Ya, the USGS links is crap. I'll screenshot them.

EDIT 2 : Edited them in my original post about it.

Edited August 19, 2018 by YNM

[Green Baron]

For that area it is nothing unusual. It becomes special mostly because it is densely inhabited. Further, from the shallow center the waves (both shear and pressure) travel fast and the isn't much dampening in the crust, so there is the potential for disastrous destruction from earthquakes, and tsunamis from submarine landslides triggered by earthquakes or volcanic eruptions as well as eruptions themselves, caused by partial melting of the fluid rich oceanic crust as it dives down through the Asthenosphere.

The area lies over a converging plate boundary. In the south the Indo-Australian plate dives into the mantle under the Eurasian one. It carries a lot of water, leading to partial melting in a depth because fluids lower the rocks' melting point. In the upper, less ductile parts of the crust tensions build up. Further, a transform faults to the north east make parts of the plates break up and slide along each other, as the subduction changes the geometry of the plates' projection on a sphere (hi flat earthers :-)). If i am not mistaken (which may well be) the fault coincides with the back-arc volcanism over the convergence (subduction) zone.

Long story short: yes, the area has a potential for disaster, from eruptions (Krakatoa), Tsunamis (Aceh) and earthquakes. I am leaving the latter to be filled out in the future :-/.

Edit: i just saw that there are several transform faults. The 2004 earthquake and subsequent tsunamis are thought to have been triggered by them. Like a wedge slipping out under pressure ...

Edited August 19, 2018 by Green Baron