Nuke

if you go with generation old intel (ivy bridge) you might save some cash and get decent performance out of it. you might also wait till after intel releases broadwell (which should happen this quarter or the next) and get a haswell machine on sale (broadwell is just a die shrink anyway). just sit on your money and wait for a sale.

switch off your gimbals, ksp doesn't know to reverse the direction of gimballing on engines placed in front of the cg.

yea thats why we have the standard atmosphere: http://en.wikipedia.org/wiki/International_Standard_Atmosphere this is what i used in my simulator (and models exist for mars, venus, etc). however since the part that matters for drag/lift calculations (pressure) its a pretty simple curve that can be approximated by a simple function, it can be used instead. temperature vs altitude is a less simple curve, but unless you want to simulate heating and cooling of objects at different altitudes with any precision you can exclude it from your calculations.

im sure you can find better machine for that price. you could do a decent i3 or i5 machine with a modest current gen graphics card with that.

thats going to be a very interesting game.

cpu: i7 4790k ram: 8gb @ 2400 mobo: asrock z97m-itx/ac (great board so far) gpu: gtx750ti hdd: 256gb 500gb ssds psu: 450w case: coolermaster elite 110 fan: 140mm hydrodynamic bering (one is all that is needed) pretty good build and i can shove it in a backpack if i ever have to go on the road. more info here and heres the parts list (sans the 500gb ssd which was a carry over from my last build)

lots of them use table lookup and interpolation between the two nearest samples (which is how math libraries handle things like trig and square roots). what i did (or rather started to do) was use simplified physics models for the vital parts of the aircraft, such as wings and fusalage segments. i then determine the four forces (thrust, drag, lift, gravity) operating on each one. linear forces just use the sum of all the forces on all physics models, i think angular forces just take the cross product between the position of the object and the force, and sum the results for all objects. it works for the most part, but its too simple to handle things like supersonic/hypersonic flight.

for unmanned stages id probibly accept a lower rate of landing success and dispense with the chutes all together. i would however use a small drogue chute to aid initial deceleration. if everything goes right, it would be jettisoned when the thrusters take over (after ignition and idling, but before throttle up). for manned stages you might have additional backup chutes as well if you have engine trouble. you would attempt to light the engines at a safe altitude and let them idle till you need them. if for some reason the engines dont throttle up, dump their fuel and release the chutes. you might not even need the drogue if your ship is large and low in density. in ksp i like to use chutes to help decelerate during a duna landing, i usually cut them and use engines when i get to below 500m. its one of those fuel saving things i do because im cheap.

ksp runs smoother on my new rig, but i think that is mostly because of the extra cpu power. the gpu is better than my gtx560 i was running, but not by much, and i can run at the next highest setting ok. one reason for getting the 750ti was because its relatively tiny. 3 fan monstrosities were a no go, and that meant either getting something older, something lower end, or throwing out my plans for a tiny pc. hopefully by next year they will have a power/heat efficient model (which seems to be the direction nvidia is going) of reasonable performance that can fit into the limited space of my case, and i can afford a new monitor to go with it since im not building a pc next year. 1080p is getting old.

my 2.0 came with 3 motors (built in rotory encoders for feedback), 2 touch sensors (they are just buttons really), ultrasonic sensor, and color sensor iirc. i think the original had a more basic light sensor, and it also has a sound sensor that is not in the 2.0 set. the software is a different version too, but i think its upgradable. if you look the set up on brick link, there is a parts inventory. 1.0: http://www.bricklink.com/catalogItemInv.asp?S=8527-1 2.0: http://www.bricklink.com/catalogItemInv.asp?S=8547-1 and they do have used 2.0 sets for sale in various conditions. but the price isnt much lower. lego sets are expensive.

i decided i didnt need the smart brick anymore when i was coding in nxc and wiring in custom sensors on the i2c bus. i still use the lego motors and sensors, i just have them connected to an arduino or some other board i made in the kitchen.

yes, the problem is you only get to do it once. neural impulses dont really travel that fast. its not a bunch of wires where the electrons move at very close to c. there are chemical reactions that slow the impulses down to about 200mph (actually there is a wide range of speeds, i just recall the one mentioned in my psychology 101 textbook which i no longer own). thanks to extreme latency, if you splat into the ground at a speed greater than that (lets say for dramatic purposes that this happens at mach 2) you are essentially atomizing your brain faster than your impulses can tell the rest of your brain that its being atomized. most people dont die by running into the ground at mach 2. capital punishment folks might consider using rocket sleds for the sake of being humane. Camacha has it right though, at the moment of death your whole nervous system is in extreme shock. oxygen isnt getting to the brain. the brain is being bombarded with signals that it doesn't have the resources to process. neurotransmitters are all out of whack. there are a lot of things going on which can explain near death experiences.

yea they sill sell the 2.0 kit as of the last catalog i got from lego. its not gonna be cheap. if you want a used one try bricklink, its like the ebay of lego.

you know i dont think anyone's aim is that good. think about it you are going to track a moving target at several miles range. nobody's hand is that stable. you might be able to do it with a scope and tripod, but tracking is going to be very difficult. i think best results would be with an infrared camera on a scope, and have some image processing software do the tracking, this would be used to control a precision turret. i dont think its the kind of thing that "some kid" could pull off. now if you had several thousand bucks, some engineering skills, and a sick desire to make a plane crash, you could pull it off. not that im advocating anyone does that.

i think aeroponic growing would work well in zero g.

what if you make the grow domes co2 rich, and then make the habitable sections low pressure oxygen? obviously the co2 rich atmosphere would be somewhat toxic to humans, but i suppose they can work with a co2 scrubber mask or something. it would definitely help if you could find a source of inert gas on mars. on mars you have a whole atmosphere of co2 at your disposal, surplus oxygen can be used to make rocket fuel. just to feed the colonists would take much more plant matter than is needed for o2 production. then you have foliage which can be either composted (not sure what thats doing to your air) or fed to livestock. the latter helps to consume the surplus o2 if rocket fuel production is not required, and produces manure. otherwise you would either need to store it or vent it.

this is why i dont use portable devices. you pretty much get locked in to what software you end up using. on a pc if i dont like the new version of windows, there are a lot of other operating systems i can replace it with. portables on the other hand are locked down with software licenses and easily voided warranties. then on top of that you are locked into app stores. i didn't like ios any more than i liked android. the only os i have actually liked on portable devices is linux actual (not droid), and i really dont like linux.

i spent the last two days putting my new rig together. it took a day just to sync up all my drives. heres the slide show in completely random order because lazy. the clearances on this thing were tighter than i anticipated. there was a good 2 inches above the cpu cooler. the ram is so close that im pretty much stuck with the stock cooler. the mini pcie wifi board is also kind in the way, so no aftermarket cooler. but its ok because its running really cool during some of the more graphically intense games i have and nothing gets above 56c. the video card was also much better than anticipated. also i like how xmp actually works now, i was able to get my ram timings up to spec without a lot of hassle. last time i went with performance ram i spent 2 days messing around with timings and running memtest86+. the system is much quieter than my previous build, the one big fan approach seems to work really well. i think the only thing i would have done differently is gone with a modular power supply. there are a lot of unused cables in the case, and most of them block a lot of airflow. in previous builds where i had done modular power supplies i had noticed that i used almost every cable and so got practically no benefit from the modular design. in this build nothing ended up needing power. i think the power supply is only hooked up to the 24 pin connector, the 4 pin connector and the 2 ssds. the video card didnt even have a power connector, which says a lot for its low power design. had i used a modular supply i might have gotten away only connecting one sata connector bus. im half tempted to take my wire cutters and heat shrink and give the psu a hair cut (i have crimping tools and soldering skills if i ever wanted to connect more stuff to it). im pretty happy with my build so far. i havent had time to check out ksp 0.25 yet, but i have nothing to do tomorrow and the download just completed. too tired now and its 3am. i need to do a battery of benchmarks too, but i think i will dedicate tomorrow to playing a few games.

thats because its in their interest to get people to upgrade.

some of these formats are older than the average user of the ksp forums. therefore there is support for it in practically every piece of hardware and software out there. actually bc5/3dc is pretty old by computer standards, its just that the amount of legacy hardware/software in the wild that cant run it is significant enough to reduce the size of your market. the reason for using a dx<10 is because of the handful of folks still running xp are significant enough to support, and its easier to just go with one api for everyone than to do conditional support, and obviously you go with the one that is most compatible (oldest). its a lot like 64 bit, its been around for a really long time, but because there is a legacy crowd out there, you do 32 bit to keep them happy and the 64bit people really dont care because they can run it too. the only people left out in the cold are the technophiles who want to run software that reflects the capabilities of their hardware.

i have a feeling the zombies would carry fleas around everywhere they go. we might be able to kill them off in droves, but the black death they bring everywhere would probibly kill a lot. they were burning bodies as fast as they could, but if they get up and start walking, you are going to have a very hard time controlling plague.

my bad, i think i skipped a page there. probibly because i had to get up and do something in the middle of a post and didnt get back till several hours later.

there is also bc5/3dc, but ive never seen a game that uses it. its a 2 channel map which is the same pixel format as a dxt5 alpha channel. dxt5nm just moves red to alpha, leaves the green where it is, and red and blue become ballast (a dirty but effective hack invented by john carmack). ninjad by a couple pages.

i read a paper on lunar tunel boring machines. it got rid of the heat but heating excavated material and removing it from the tunnel.

opengl can do dds just fine. its kind of a misconception that its an m$ format. dxt1 and 5 were actually s3 formats back in the stone age of 3d acceleration, the formats are really old. m$ had its own specifications for them in d3d (likely to avoid paying s3) and the names they used kind of become the defacto terminology over time. they may have created the dds file format, but once the pixel data is loaded the video card doesnt care. yes and no. yes you can open them and edit them, but since the file is lossy. you should load an uncompressed version if you want to make changes. you dont want cumulative data loss over a series of edits.