Kerbal Space Program

[Ulminati]

 

[Yaar Podshipnik]

I really hope this is a joke. This is a game about green things getting blown up while trying to get into space, not another "battlefield for the equality in everything war".

 

[Spectacle]

Indeed. The lack of individuality among the kerbals just highlights their expendability in the name of progress.

 

[Hellraiser]

The research system is shaping up nicely. Apparently there will be science data storage parts and antennas will now work for data uplink. Also it looks like it is pretty certain the 0.21.2 patch will be the one with the unity upgrade which should fix some issues.

 

[Scruffy]

i can't seem to be able to reach the damn mun

 

[Burning Bridges]

That's not hard. Just get into any orbit around Kerbin, point your ship to the flight direction and fire the engines again when the Moon rises over the horizon. Do this while you watch trajectory (hit 'M') and accelerate until you are on a course that captures you. Don't try to get faster than necessary. From there experiment a bit.

 

[Hellraiser]

i can't seem to be able to reach the damn mun

First try the solution to 90% of all problems in KSP:
Add moar rocket. Add more struts.

If you can get it to the Mun but you can't get into it's sphere of influence (you miss the damn thing in layman's terms), then get into a circular low orbit first. Burn when the angle between your prograde (the direction in which you craft is heading, the green circle marker on the navball that isn't crossed out) and the Mun's position is 45 degrees. When orbiting eastwards (which you should always be doing apart from Polar orbits) that coincides with the Mun rising above the horizon while in low orbit.

Or you can just set up a maneuver node with a planned prograde burn reaching the Mun's orbit and drag it until you get a suitable projected encounter. Execute the burn by pointing at the blue marker and burning as much as the game tells you at the time it tells you to burn.

 

[Scruffy]

i can't seem to be able to reach the damn mun

First try the solution to 90% of all problems in KSP:
Add moar rocket. Add more struts.

If you can get it to the Mun but you can't get into it's sphere of influence (you miss the damn thing in layman's terms), then get into a circular low orbit first. Burn when the angle between your prograde (the direction in which you craft is heading, the green circle marker on the navball that isn't crossed out) and the Mun's position is 45 degrees. When orbiting eastwards (which you should always be doing apart from Polar orbits) that coincides with the Mun rising above the horizon while in low orbit.

Or you can just set up a maneuver node with a planned prograde burn reaching the Mun's orbit and drag it until you get a suitable projected encounter. Execute the burn by pointing at the blue marker and burning as much as the game tells you at the time it tells you to burn.

i've tried that shit for three days, no idea what i do wrong but by the time i get in earth's orbit i have very little fuel left and i can't do the "burning" thing because i just run out of fuel.

 

[Hellraiser]

i've tried that shit for three days, no idea what i do wrong but by the time i get in earth's orbit i have very little fuel left and i can't do the "burning" thing because i just run out of fuel.

Add moar rocket, there are diminishing returns of course but even with demo parts (not sure if you have the full game or not) it shouldn't require too many to get to the Mun. This thing below gets to the Mun as you can see here, it is somewhat of an overkill though IIRC:

There is one trick to making more efficient rockets which I am using in that rocket, asparagus staging (easy tutorial entry on the wiki here). This one fires all boosters and main engines at launch, ditches solid boosters first, ditches two outer tanks and engines next (liquid fuel boosters), then the other two tanks and engines, finished orbit on the central engine and tank. The carrier rocket uses asparagus staging apart from the SRBs (since those don't use liquid fuel that can be pumped).

Fuel lines pump fuel from one tank to the other keeping the target tank as full as possible (the tank you attach it to first gets drained). You use fuel lines to connect the stages you jettison first to those that will be jettisoned next, the idea is that you get rid off dry mass you don't need (empty tanks+redundant engines) as fast as possible. All or most engines (apart from the payload like a lander or interplanetary cruise stage etc.) fire at the beginning, first stage gets drained and dropped, remaining tanks are still full of fuel. The ones to drop next get drained next. You can drag decoupler/engine icons in the VAB or even after launch to change their staging order. You need to do it since the game assumes regular staging when you assemble rockets.

Either way you add enough stages of tanks+engines to maintain a thrust to weight ratio greater than 1 until you get into space (enough not to start falling down basically). The more mass you have the shorter the first stages will last. I think three is the most you need to get a Mun orbiter with enough fuel into orbit.

It's not used in real life (apart from the planned SpaceX Falcon Heavy) due to issues with pumping fuel between stages, it is complicated, costly and increases risk (simpler rockets are safer, less shit to fail on you). Also having all engines fire at onces means the rocket is wide and thus less aerodynamic, but theoretically it is the most efficient way of staging rockets.

 

[Scruffy]

no yeah i've got the demo
will keep trying though, thanks

 

[Hellraiser]

Use the most powerful demo engine (highest thrust). Double long tanks and one of those engines should have a TWR above 1 IIRC, enough for some payload that isn't more fuel and engines. If not go for long+short tanks as liquid fuel boosters to increase the TWR of each stage. The solution is to just add enough of those 2tank+engine stages and organise staging so that it doesn't lose speed at any point. Experiment with fuel lines, you may want to drain and drop liquid boosters in pairs or in fours depending how much thrust you need. Once you learn how to get to Munar orbit the fun part begins, learning how to land there without killing the eager green men (just remember parachutes don't work with no atmosphere, which the Mun obviously lacks).

 

[Absalom]

no yeah i've got the demo
will keep trying though, thanks

Post screenshots and we'd be able to help moar.

Also, if you have a stable orbit over Kerbin, then faff about with the maneuver planner on the map screen until you get an encounter.

 

[Hellraiser]

They have shown some stuff from 0.22 at PAX during a live stream:

WIP tech tree GUI:

Apparently tech will both unlock parts and have passive effects like reducing total mass of structural parts by X%.

Science bay part (the one on the bottom end) and data storage part (the canister at the center):

It has multiple experiments that you start via the contextual menu (and which can probably be assigned to action groups as well).

Command pods can also be used for experiments, I guess for minor ones so you can somehow get the early SCIENCE! points required to develop the science bay and other advanced science parts:

Also it seems the green dolts will get to report what they see. I wonder if in the future their stupidity stat will come into play here.

Also sub-assembly support seems to already be fully working:

Of course they are shared between the VAB and SPH so no more manual copying of rover, plane and other .craft files will be needed. A much needed improvement of the editor.

Overall this is shaping up to be the best update since 0.18. since science is a pretty big new feature that changes how the game is played.

 

[Hellraiser]

Anyway an official features video for 0.22 has been released, the community/PR managers mexican accent makes it a bit unintelligible at times:



It wasn't mentioned but there is supposed to be data loss for data being sent back to Kerbin, hopefully based on distance and antenna power.

 

[The Vanished One]

Good news all around. I just hope there'll be a bit of optimization on the side.

(Also, I'm more and more concerned that once better aerodynamics are implemented, I'll have to re-learn half of the stuff. It'd be great if they made it one of their next priorities.)

 

[Hellraiser]

They haven't talked much about the new unity version they have been trying out recently. But it is pretty much given that the bugs left over by 0.21.1, especially the scene transition one where going from the space center screen or VAB to launch takes way too long, will be fixed. It has been a major performance hit.

I wouldn't expect them to work on aerodynamics for at least one more update after 0.22. It's pretty much given that the next one will focus on another major career mode feature which will need lots of new code. Probably mission generation, I was actually surprised they went with research now for 0.22 instead of missions but it makes sense. You can't have a mission generator work properly if you don't exactly know what tech the player may or may not have.

Still, after those are added they really don't have that much demanding work ahead for career mode. Fleshing out kerbonaut training, working out a prestige system to rate you space program, adding competing space programs (would be cool if you could do joint missions with NPC spacecraft like the real life ISS or the Apollo-Soyuz mission), the economy. After most of the career mode features are done they'll probably go polishing stuff, including aerodynamics and adding re-entry heat and adding requested mostly minor features like resource mining, hinges/robotic arms, the planetary discovery mechanic (along with outer planets, an asteroid belt etc.) and space telescopes.

 

[Hellraiser]

More R&D ahead, apparently experiments will be region/environment sensitive. In practice this means landing on the polar ice caps of duna will count as different than landing in the equatorial flat lands, the same will probably apply to the Mun's giant crater or maria, oceans of Eve/Laythe. Depends how they implemented it, I'm hoping it's based on the procedural tech responsible for shaping terrain features rather than being manually defined area (since it would create issues with boundaries being inaccurate).

Also we'll be getting a new very powerful radially-attached deployable communications dish along with other new antennas:

Data transmission is supposed to be very power-hungry, which is good since the vanilla game really needs more stuff that uses lots of power to justify using gigantor XL panels for more than aesthetics. I'm hard from all the science.

Also if anyone is into coding stuff and the technical software side of how this works there's quite a few things, particularly related to how the planets and orbit to surface transition works in KSP, in the Unite 2013 Unity Dev Conference:



Really cool about how they solved the floating point number issues with astronomical distances.

 

[Absalom]

I'm hoping it's based on the procedural tech responsible for shaping terrain features rather than being manually defined area (since it would create issues with boundaries being inaccurate).

Can you elaborate on this? bit confused

 

[Hellraiser]

Can you elaborate on this? bit confused

There are two ways to define what area constitutes the desert on Kerbin for research purposes, one is to look at whatever code there is that makes the desert as big and in the shape that it is in, the other is to re-trace its boundaries manually based on final output (the shape shown in game). The first one is a better choice since it will always give you 100% accuracy, you stand on a sandy desert texture this is a desert. The latter may have issues in border regions, due to the scale of the game it is possible that you might be 100 or even 500 meters deep into the desert, seeing the edge on the horizon, but the experiment thinks you are not in it. It also has the disadvantage of having to be done manually with each makeover of a planet's surface, however depending on how the game decides that an area is a desert using the first method may be too CPU power intensive. Then again that may be a non-issue since experiments are single player triggered events and you can disguise the time between starting the experiment and getting a result as an experiment running time.

Of course there are some assumptions on my part and my experience with programming is very limited.

 

[Hellraiser]

So apparently 0.22 is in experimental build testing. They already tested a lot of the features earlier during development as individual branches were done so hopefully it should be released within a week. Possibly longer if there are still major bugs lurking in the code somewhere.

One of the new things that is confirmed is collecting rock samples in EVA and that returning them to Kerbin for more research points. That also applies to some other experiments as well where you expose kerbin bacteria to alien conditions and return them back. Also there are the earlier mentioned biomes but just on Kerbin for now. Still in the future landing a rover near biome edges will allow you to get more science done in one mission.

Anyway the shot of the final tech tree was released:

Some of that endgame tech looks a bit enigmatic. Also power generation and probe parts seem to be quite far into the tech tree, bah even ladders and lander legs are. Guess we'll see how it works in practice, after all different tech will have different research cost. I'm guessing the chemical symbols are new types of experiment modules. The fuel icon seems to be fuel lines. The top most branch seems to be bigger tanks+engines with the last ones being mainsails probably, the LV-N is pretty obviously the radiation trefoil.

The starting node looks pretty good although the lack of decouplers in it is both surprsising and possibly quite hilarious. In before somebody tries to decouple solid boosters the jeb way, that is firing a liquid fuel engine on top of it hoping the booster below either explodes from heat or detaches itself due to the force of the exhaust.

 

[Absalom]

looks good, hopefully they'll be more stuff to research than what we've already been using forever tho

Also the landing rovers on kerbin thing sounds boring honestly; rovers in general i don't fuck around with much since they are so fiddly and i rarely bother faffing about kerbin when theres a whole solar system to explore

 

[Hellraiser]

There was a streaming marathon yesterday/today where some of the streamers got access to 0.22. Apart from bugs getting fixed the only thing that is likely to change from that may be some shuffling around with the tech tree. Anyway here's what has been added:

- lots of old flight code got rewritten and optimized, scene and SoI transition loading time bug got fixed. Overall it should run faster by supposedly up to 30%, if all the optimization made it into the update.
- new SAS code logic got improved.
- science and tech tree. All old science sensor parts now do experiments, manned pods can do experiments and the old avionics package is now some fancy atmospheric sensor. AFAIK all experiments are unique that is apart from pods all parts yield unique results. Kerbals can also gather dirt/rocks in EVA for more research, experiments are biome sensitive. Recovering vessels yields extra research from both taking the samples back to kerbin and not losing data *and* from studying the returned craft.
- new landing leg code and remade/upgraded part models. Now all landing legs have proper suspension and shock absorbers, that means they shouldn't jump around as much during 10+ m/s landings. Also helps a bit with landing on slopes. Pistons can break upon impact though, can be repaired in EVA like flat rover wheels. The flimsy and much hated medium legs have a new model and are no longer ass, the micro landing legs model got some tweaks but I'm unsure what they were.
- basic data transmission using antennas, has a nice log detailing the status of the transmission.
- biomes on Kerbin and the Mun, they will come to other planets in the future. Kerbin's are pretty self explanatory, ice caps, deserts, mountains, tundra etc. On the Mun major craters count as separate biomes (so for some experiment their research yield is independent from other parts of the Mun), as do the maria probably as other geological features were also mentioned and maria are the most obvious of those. Burning Bridges probably approves of this.
- entirely new parts: 2 new deployable antennas (the largest one looks quite nice, the bow shaped one shown earlier is a medium sized one), 3 (IIRC) experiment parts including the goo canister, materials lab and some other part I forgot.
- KSC got a bit rearranged, the R&R building got added, runway has lights at night for easier landing, runway slope is less inclined so it should kill less rovers/planes by accident.
- also the much requested sub-assemblies have been added, saving lifters or payload and combining them is now possible without having to rebuild one or the other from scratch.

Stuff that is scheduled for the future (possibly next update): proper distance based data transmission system and hopefully communications relays. The current system has a distance independent data loss for some of the more complicated experiments like sample analysis, so it's better to try to return a vessel with those but how powerful your antenna is and where you are is irrelevant unfortunately. Fortunately this is one thing the devs do want to flesh out more. Also adding more experiments (so more science parts) and biomes on more planetary bodies. Also new tech nodes as new types of parts get added since the tech tree is not final and will evolve as needed, like when we get heat shield and life support.

 

[Hellraiser]

Also the landing rovers on kerbin thing sounds boring honestly; rovers in general i don't fuck around with much since they are so fiddly and i rarely bother faffing about kerbin when theres a whole solar system to explore

I watched some 0.22 stream yesterday/today and have a better grasp of how it works in practice now. There are biomes just on Kerbin and the Mun for now so you can land a rover near a major Mun crater, do experiments, send back data, drive into the crater and do experiments there for more science without diminishing returns since big craters and/or maria are separate biomes. But you don't need to, there's quite a lot of choice in how you want to get science.

Low orbit is biome-sensitive for some experiments, you can put a command pod into an inclined orbit (or sub-orbital trajectory) and have the crew send back reports from observing mountains, the ocean, deserts, polar ice caps etc. for more science in the very early game. Limiting factor is the lack of batteries on the starting tech node since you only get electricity from liquid fuel engines while they're running, while antennas consume a lot of power for data transmissions. Either way no need to land rovers, and those are pretty far in the tree anyway.

All experiments are context sensitive as in they give independent science yields based on what you are orbiting (including the sun if you escape kerbin) and if you are:
- in low orbit
- in high orbit
- in the atmosphere (if it has any), the thermometer and barometer only do experiments here and on the surface, you can still use them to measure stuff elsewhere but without getting data to send back. Also the old avionics package does atmospheric composition analysis experiments IIRC so it works only here.
- on the surface, the old accelerometer is now also a seismometer so you can use it to do experiments here. Also sample collection in EVA.

Both reports from crew observation while inside a command pod and in EVA count as separate experimental data and have separate yields. Either way when you add up all the experiment types, biomes and locations you get quite a bit of choice without any obvious railroading for research missions. So while there are limitations due to part availability you have a lot of freedom in how you design your missions. Some rebalancing, polish and part shuffling in the tech tree may be needed but overall the system looks pretty damn good.

Most but not all cases have different experiment result flavor text, others have placeholder default result text but still yield science independently based on orbit, location and biome. Also experiment yields differ based on the orbiting parent body. The sun gives you possibly the most science I have seen so far, as you exit kerbin you enter high solar orbit and basic "mystery goo" observation gives something like 200 science which is about as much as a tier V tech tree node requires to unlock. Although Jool, Eeloo or Moho could give more for obvious reasons. Oh and for now experiments have a set data loss value so you gain back between 20-100% science yield if you transmit data back, if you return it back to kerbin you get the difference between it and the transmitted value back. So no worry if you are not sure if you can make it back. Experiment parts can erase the data after transmission or keep it.

Saw a guy min-max a vessel for maximum science with what he had unlocked during a Duna orbiting and return mission. He got Duna atmosphere expriments, low and high Duna orbit experiments, high solar orbit experiments, high Ike orbit experiments in one go. Had multiple science parts of each type (including command pods) to store as many experiments as possible as he did return to Kerbin only transmitting back crew reports. Got him about 1000 science point which made him unlock all of tier IV and two tier V tech nodes. So adding more science parts to store more experiments is a good idea if you plan on visiting multiple moons and returning back.

 

[InnerPartisan]

So if .22 is in experimental build testing, that means it's gonna be out any day now, right?
Man, I'm so hyped about the SCIENCE! feature. Can you even imagine what possibilities this opens up for modders? Good Lord.

 

[Absalom]

Indeed. Hopefully this will convince me to waste more time with KSP