raptop changed the topic of #KSPOfficial to: Kerbal Space Program official channel | KSP2 current version: 0.1.5.0 | Rules: http://tinyurl.com/KSP-Rules | Δv maps: https://i.imgur.com/CHVnEeE.png https://i.imgur.com/gBoLsSt.png
_whitelogger has joined #KSPOfficial
darsie has quit [Ping timeout: 186 seconds]
SrGato has joined #KSPOfficial
<SrGato> Hi! is there any way to switch between different ships without entering the tracking station?
<SrGato> I mean in ksp1
<Izzy> you can do it from the map view, and you can do it with the alarm clock
<SrGato> alarm clock?
<packbart> the maneuver alarm tool. you can switch to vessels that have alarms set
<packbart> but yeah, I usually use the map view
<SrGato> oh ok
<SrGato> thanks
SrGato has quit [Quit: webchat.esper.net]
<Pinkbeast> Alarm> wait, does KSP1 have a vanilla maneuver alarm now?
<packbart> it does
<packbart> by the same author, I guess
<packbart> since 1.12, I think
<Pinkbeast> Huh, interesting, thanks. (I see the KAC thread is full of people complaining about the stock alarm :-)
<packbart> I still use KAC. it just looks better to me
<Pinkbeast> I'm still in 1.9.1 so it's an easy choice
<Izzy> hoping kOS gets a stock alarm interface eventually
<Izzy> was nice with KAC but I'm not going to duplicate functionality y'know?
Tank2333 has quit [Read error: Connection reset by peer]
Tank2333 has joined #KSPOfficial
<umaxtu> I never really got into KOS
Tank2333 has quit [Remote host closed the connection]
Wastl2 has joined #KSPOfficial
Wastl4 has quit [Ping timeout: 186 seconds]
Althego has joined #KSPOfficial
<raptop> I've had issues with the time warp in the stock alarm thing. I swear, it messes with orbits
Ezriilc has quit [Ping timeout: 195 seconds]
Wastl4 has joined #KSPOfficial
Ezriilc has joined #KSPOfficial
Wastl2 has quit [Ping timeout: 190 seconds]
<Izzy> cases like stations inexplicably deorbiting themselves while time warping?
<raptop> not as such, but my testing was pretty limited, and more of SOI crossing rendezvous going poof
<Izzy> I've had a few cases where I'll dock my tiny capsule to do a "take tourists to a station" mission, only to time warp 15 minutes and end up dipping into the atmosphere
<Izzy> and I can safely say the capsule could not change the trajectory of the whole structure that much if it wanted to
Ezriilc has quit [Remote host closed the connection]
sasamj has quit [Quit: Connection closed for inactivity]
<taniwha> wanna bet they used floats :/
<taniwha> (when you're moving at 2km/s, the limited resolution of floats rears its ugly head when calculating orbits from position and velocity)
<taniwha> 2km/s and 700km from the center
<taniwha> 0.12mm/s and 42cm precision
<taniwha> if they're repeating calculations base on that, the errors will build up fast
purpletarget|ktns has quit [Quit: Any fool can calculate]
purpletarget|ktns has joined #KSPOfficial
<Izzy> KSP seems like a good case for fixed point numbers
<Izzy> "fixed point" would be a bit of a misnomer I suppose, you'd just be using increasingly small units depending on where you put the decimal
<Izzy> and what your "base" unit was
<Althego> that sounds like floating point
<Izzy> well, no, because in floating point the decimal moves, no? fixed point would specify "the decimal point is here" and that would be a design decision
<Althego> until it isnt. all the horrors of changing units when multiplying numbers
<Althego> shif to this, then shift back for the result, etc
<Izzy> whole new fun classes of bugs
<Izzy> :D
<Althego> i saw a whole simulator written this way in assembly
<Izzy> was curious and did the maths, if you declare 0x1000 is 1m, you get just under 0.00025m - 0.25mm - precision and can fit the orbit of neptune into your game world with 64-bit integers
<Izzy> (and then some)
<Althego> and ksp is cheating by making everything smaller
<taniwha> fixed point is *not* the answer
<taniwha> not breaking the code that others carefully built *is*
<Althego> "violance is not the answer. it is a question, and the answer is yes" :)
<Mat2ch> taniwha: for a game like this it is.
<Althego> anyway people mostly dont use fixed point numbers today
<Mat2ch> most people aren't trying to simulate a solar system ;)
<Althego> they can just use doubles
<Mat2ch> Others might. But you get huge problems if you try to add very tiny numbers with very big numbers by using floats.
<taniwha> Mat2ch, go on, go write a game like this
<Althego> i suggest you dont use fixed point numbers
<Mat2ch> That's not how a discussion around this works.
<taniwha> and you get the same problems trying to add very tiny numbers with very large numbers
<taniwha> you're going to run out of room either way
<Mat2ch> Depends on the precision you wanted to achive.
<taniwha> because with fixed point, you're going to run out of very small or very big
<Mat2ch> very big is easy to fix. Very small is just a definition.
<Mat2ch> If you define nothing moves below a certain limit then nothing will move.
<Mat2ch> If you really want to get fancy you could go down to the plank length. But I guess micrometers will be enough
<taniwha> and doubles get you to Gm with um precision
<Mat2ch> which will eat any small number if added to it.
<taniwha> yes, and?
<Althego> if it is so far away, any small movement doesnt matter, you wouldnt see it anyway
<Mat2ch> those small numbers could be important. Because you're not adding them once, but several hundred times
<taniwha> so add them several hundred times before adding to the big number
<Althego> exactly
<Althego> keep them in their own reference
<Althego> just transform the resutl
<Mat2ch> Which is not possible, because a physics engine can't look into the future.
<taniwha> it is possible, and ksp does it
<Althego> no future needed for it
<taniwha> it's called the floating origin
<taniwha> and krakensbane (which is a floating origin for the velocity)
<Mat2ch> how much computing power does that cost?
<taniwha> very little
<taniwha> or even none
<Althego> krakensbane is still a funny name
<Mat2ch> Did you do tests?
<taniwha> did you?
<Mat2ch> How should I? I never had access to the code.
<taniwha> it costs very little to none because you're not doing the "big add" all the time
<taniwha> and working with floats *is* faster than working with doubles
<Mat2ch> Wasn't talking about doubles ever.
<taniwha> if for no other reason than 2x the data points for the same amount of cache misses
<taniwha> actually, you were, you just didn't realize it
<taniwha> because it's a whole system
<Althego> any chance of a real 128 bit system in the near future?
<taniwha> already have 512 bit systems
<taniwha> but when it comes to ints: probably not. 64 bits is a lot
<Althego> as far as i know there is no 128 bit hardware. there are multiple data instructions which might work with more, but no native 128 bit integers
<taniwha> proper 128 bit floats would be nice, but that's expensive in gates
<taniwha> intel does have 80-bit floats, though
<Althego> internally at least
<taniwha> full access at the asm level
<taniwha> but you have to use the stack-based fpu
<taniwha> (good old fp(0-7)
<taniwha> and apparently the thing has bit-rotted a bit (not as fast as the xmm regs)
<Althego> i gave up waiting for arm to replace x86. maybe risc-v, but even if t hat happens it is going to be in 10 years minimum
<Althego> at least intel is gearing up to cut out old parts of x86, which are just there for compatibility
<taniwha> the only thing that will replace x86 is whatever intel's cooking up
<Althego> but they dont plan anything else
<taniwha> you don't know what the do and don't plan. you know only what they've talked about
<Althego> they are lucky if they can keep up with amd for the next 4 years
<Althego> that is true, i dont know. but based on the leaks, it is x86 for a while at least
<taniwha> no, there's no luck involved
<taniwha> unless by luck you mean amd slacking off
<Althego> which is totally can happen if intel falls behind
<Althego> just as it happened with intel
<Althego> and they still couldnt completely get back
sandbox has joined #KSPOfficial
Cranium has quit [Read error: Connection reset by peer]
Cranium has joined #KSPOfficial
<packbart> hmm. "64bit star system precision something something" is a talking point of the Star Citizen pushers
<packbart> (or rather for Star Engine, a modified Amazon Lumberyard, a modified CryEngine)
<packbart> Elite Dangerous already did it before, of course
<Althego> hehe
<sandbox> !mission
<LunchBot> The How It's Made crew films a 15-minute piece at the KSC facility. For reasons unkno-o-own!
<Althego> i dont see that as an outcome
<Althego> however still fits somehow
<Eddi|zuHause> probably some kind of reference?
<Eddi|zuHause> !mission
<LunchBot> You surrepetitiously replace a cat's blini with a waffle. You help Kerbals achieve Kardashev level II.
darsie has joined #KSPOfficial
Fluburtur has joined #KSPOfficial
minas_tirith has quit [Quit: bye]
minas_tirith has joined #KSPOfficial
<KrazyKrl> ;outcome fixup s/You help Kerbals achieve/Kerbalkind is moved to/
<LunchBot> New text is: Kerbalkind is moved to Kardashev level II.
<KrazyKrl> up or down? you say... "why not both!"
<Eddi|zuHause> !outcome add Two failures make a success. Right??
<LunchBot> Added outcome: Two failures make a success. Right??
<KrazyKrl> ;outcome fixup s/??/‽‽/
<LunchBot> New text is: Two failures make a success. Right‽‽
<KrazyKrl> ftfy
<Eddi|zuHause> could be argued either way
<KrazyKrl> ;outcome add Your usage of interrobangs increases by 500%‽
<LunchBot> Added outcome: Your usage of interrobangs increases by 500%‽
<Eddi|zuHause> typically i spell it "?!?"
<KrazyKrl> yea, that's the uncool way of doing it.
<Eddi|zuHause> but i specifically didn't use that in this case
JohnnyJ has quit [Quit: Ping timeout (120 seconds)]
JohnnyJ has joined #KSPOfficial
JohnnyJ has quit [Client Quit]
<sandbox> !mission
<LunchBot> You open the mystery goo at the bottom of the ocean. There are unforeseen consequences.
<Eddi|zuHause> oh noes.
lewildZesty has quit [Read error: Connection reset by peer]
lewildZesty has joined #KSPOfficial
XXCoder has quit [Ping timeout: 186 seconds]
sasamj has joined #KSPOfficial
<packbart> There is a difference between ?! and !? that a single character cannot convey
<Eddi|zuHause> also, the combined char is chosen from a fallback font that is not monospace.
<Eddi|zuHause> (but that's probably a me problem)
<packbart> yeah, it looks a bit weird on my screen, too
<umaxtu> I've fired up KSP2 for the first time since April, has the center of mass never updated when you change the ammount of propellent in a tank?
<umaxtu> built an ssto but it keeps summersaulting forward when I try to land.
<sandbox> somersault
<umaxtu> *facepalm*
<Althego> hehe
<umaxtu> at least its not darting off to the side and rolling over like most of the planes I build in KSP1
<Pinkbeast> Eh, technically "SSTO" says nothing about what happens to it after that, so call it success :-)
<umaxtu> easy to do since the cockpit survived
Ezriilc has joined #KSPOfficial
<FLHerne> SSTOAB
<Mat2ch> AB?
<Althego> and back?
<Mat2ch> all backwards!
<Mat2ch> assembled backwards?
<Althego> ab: afterburner, mwd: microwarpdrive
XXCoder has joined #KSPOfficial
Tank2333 has joined #KSPOfficial
Tank2333 has quit [Read error: Connection reset by peer]
Tank2333 has joined #KSPOfficial
<Eddi|zuHause> !mission
<LunchBot> Across the gulf of space, minds that are to our minds as ours are to those of the beasts that perish, intellects vast and cool and unsympathetic, regard Kerbin with envious eyes, and slowly and surely draw their plans against us. You receive the prestigious Kerman Prize for Idiocy in Spaceflight, presented by Jebediah Kerman himself.
<Eddi|zuHause> oh, a price
<XXCoder> interesting plan from aliens
<Mat2ch> !mission
<LunchBot> You mistakenly input minutes instead of seconds into the barbeque roll maneuver duration. It is unclear which button to press, so Jeb presses all of them.
<XXCoder> I blame jeb
<Mat2ch> It's always Jeb
<darsie> .outcome add Jeb fails to keep your rocket retrograde during entry.
<LunchBot> Added outcome: Jeb fails to keep your rocket retrograde during entry.
Althego has quit [Quit: HMI Module Alpha Humana on approach to Space Station Mercury]
lewildZesty has quit [Ping timeout: 190 seconds]
lewildZesty has joined #KSPOfficial
<sandbox> !mission
<LunchBot> You decide to circumnavigate all the things. Taniwha deletes your recently added ;mission due to your bad grammar.
King_Arthur has quit [Quit: ( ̄▽ ̄)ノ]
King_Arthur has joined #KSPOfficial
<XXCoder> ;mission add Jeb wonders what would happen if he presses a button way beyond what is recommanded press time.
<LunchBot> Added mission: Jeb wonders what would happen if he presses a button way beyond what is recommanded press time.
<raptop> !mission
<LunchBot> Stock aero comes out. A creepy old wizard screams "BINDING MAGICAL CONTRACT!" and pays you in weird gold coins.
<raptop> coins with drag cubes
sandbox has quit [Quit: Leaving]
Fluburtur has quit [Read error: Connection reset by peer]
<FLHerne> .outcome add It why you know but don't works.
<LunchBot> Added outcome: It why you know but don't works.