tater

Presumably the existence of facilities that improve means that modders can make it look as different as they like.

Here's something you might want to add for the next version

Ah, the brown fields of home. (my neck of NM is rather more spiky, both in plants and rocks)

I can see that. Again, AI is needed all around. Programming them is fine, but in RL they have done the math, and it's far more exact than KSP nominally is. Having a probe "programmed" to do a correction burn makes sense (you drag the node around, and when the time comes, the probe simply executes that command exactly as you have set it (a 3 second burn as angle X in 57 days, 2 hours 3 minutes, and 23 seconds). When it comes to landings, etc, that's where it needs AI as the "soonest" node you can program can be set to the distance in light seconds from Kermin (disallow setting a node closer than XXX seconds away where XXX is the distance in LS). I'm changing my mind a little on your idea, I can see it as a way to drive multiple kerbals. The game could simply assume that any kerbal not actively doing something is resting up, and dividing the "on duty" guy status so that whenever you want to do something, there is a guy good to go---except if he's a single crew member, then if you utilize him too much, he gets tired. Most of the time it seems like ti would not be used, and could be abstracted, though. You land your lander, then one guy gets out, plants flag, takes soil, files EVA report. Done (in about the time it took you to read that sequence). With FP maybe it might take longer. It seems like an easier way might be to weight science gain by crew members aboard the ship (or within physics distance when gathered). So a single man lander might get 25 points, but the exact same sample gathered by a 2 man lander might get 50. A simpler abstraction.

I think that barring huge changes to kerbals in KSP, this is a pretty bad idea. There is little to actually do with kerbals, so making it harder to get them done means making it simply more annoying to get it done. Adding life support makes far more sense (which supposedly won't happen). For this to be useful, kerbals would need complete autonomy, IMO. AI kerbals that do stuff. THEN, fatigue makes tons of sense. Without the AI, it's just tedium. I can see the point in forcing larger crews, I suppose (having shifts, basically).

Tourism is really no more than any other "orbit at altitude X" contract. It adds nothing interesting, IMO. What we need is stuff to actually do.

Got it on RSS. But that was the last install I tried, the other 2 were stock, with just FAR/DRE and a few part mods. The 2d time I scrubbed out all the part mods except full PF (again, not RSS) and I tried that as installed, then I tried again with the PF dll you have removed (so I only had the one dll from PF in GameData). Always loaded up, showed all the parts, I even launched a couple times before it locked up on me. Usual drill for any lock ups I ever have, I watch memory usage creep towards 3 GB, then it crashes once it crosses. You'd never know I have 12 GB of RAM, lol (that's a KSP issue). I'll try a total vanilla KSP, but I always use FAR/DRE. OK, that worked. Nothing but stock and ATM, and it sits around 2.78GB. I'll add in FAR and DRE and cross my fingers. It's just a memory thing, as it loaded up fine.

Looks really good, but it goes over 3gb ram and crashes on me. I was trying in a minimal 6.4X install. It throws no error loading, and I can build, even launch, then it locks up on me.

Some are self-evident, like oceans. Where certain geological formations are are seemingly random now (some craters have a biome at the bottom while others do not, some are different than others, etc). In addition, I mean randomize the "biomes." If there are 10 kinds on Kerbin (20, 30, whatever) apply them randomly. That means build the terrain randomly, not slap a random name on it. If an area is sand colored (in your last game it might have been green) it is probably desert, green, could be steppe, might be "grassland," dunno. The bottom line is that until we sent an orbiter to Venus to radar map the place, we had no idea what the surface was like. Until we got better pictures of Mars from space, we knew little about the geology. Some "biomes" (we can ignore the fact the Mun, Minmus, etc don't have biomes at all, as they are lifeless) would require surface collection/analysis to determine. The bottom line is that with google, there is no way to have a proper "fog" about planets unless they are randomized. Having that means that "science" is not just for points, but to gather information that is actually useful.

Refuel a competing program's ship trapped in orbit someplace. Salvage a competing program's ship in orbit someplace (and possibly rescue the crew (some might be empty)). Rescue a crew from a crashed lander (competing program). All the above with tight time limits. Like launch within 2 days, complete within XX days (appropriately longer if they are past Kermin SOI).

No one should know where any of the biomes are, ever, they should be randomized, and you should have to actively do science from orbit to detect them. I'm not sure how small it is as a change, but a huge payoff in replay would be to have the option to generate an alternate system each new career. Same basic system, but with the worlds in different arrangements, perhaps aside from Kermin (which needs to be in a habitual zone). Everything else might be up for grabs. With a "seed" system, people could even find cool systems and share them (default would be the extant system).

As if getting funds was actually difficult, even on hard. Or even on hard with payoffs dropped from 60% to 30%. Contracts need work, not just new contracts, or more hoops to jump through, but something more interactive altogether, IMO.

I think all rescue missions should in fact be non-player craft. Stranded kerbals included along with their ship... Another mechanism needs to be added other than the player taking complete control via ][, or at least remove the rcs fuel (or 99% of it) from the astronaut so that the player needs to actually fly to them. in general, there should be a sense that there is a competing space program.

Not just removing dumb contracts, but removing any repetitive contracts. Any "plant flag" contracts should instead require a new biome, for example. Static tested an SRB? You'll never see that contract again.

Instead of having contract artificially require X kerbals (or any other arbitrary requirement), it would be better if there was some reason for them. The best mechanism, IMO, would be AI kerbals. If they could do science tasks on the surface, and instead of the current, land, plant flag, EVA report, surface sample, DONE---in only slightly more time than it took me to type that---then the collection science might take a while. You EVA, and can control a guy, meanwhile the other climbs down and starts setting up experiments, takes samples, etc. Doesn't have to be fancy, it can be fairly scripted. If there are certain experiments on the side of the lander, the AI goes to them, fiddles, some might be removed, and placed on the ground, etc. New experiments could be added just for something to do for the AI. Set it up such that 2 do work faster, and collect more science. Lower the "instant" science, and extend it over time. If you have a base, and it is manned, they do experiments X hours a day as EVA. It would add "life" things, and might make bases, stations, etc more useful (not just science per day generated, but science per day iff there are X kerbals present, then more science per Y additional guys.

Aside from bug fixes, I think the obvious choice is really optimizing KSP, particularly memory utilization. Right now it touches just 3 GB memory use, and crashes. Not having to worry about that would allow more mods In terms of features it's more tricky because what I might think of as a "small" change could very well be difficult to code (constant acceleration trajectories is an obvious one).

Given the inclusion of various mods already, and assuming that once they are "feature complete" they work on optimization, it would be interesting to see a competing space program added. There are numerous mods with Soviet spacecraft, some of which seem pretty "stock alike." No all parts would need to be redone for each program, it might be as simple as an alternate texture for some parts that are reused, plus the more definitive parts (command pods, a few engines, etc). Add a new space suit for them as well (even if just a color change). You'd add an alternate launch site for the other side (Union of Kerbinov Republiks or something like that). The "rescue" missions might default to being the alternate program. Ideally there might even be competing launches, stations, bases, etc that you might encounter in flight. If their space center was inland, and not equatorial, then playing the other side would be an option for increased difficulty as well. Another difficulty setting or play type might be "Space Race." In this case, you'd start off about equal, and the opposing program would have a scripted set of missions it would do in some scheduled timeframe (perhaps a little randomized for replay). You'd lose reputation if the other side beats you to a task. The AI program might intentionally interact with YOU, as well. If you end up with no fuel in orbit, for example, they might send up a rescue mission (their "rescue a kerbal). You could lose a pilot (he defects to their program) and lose rep if they rescue before you do... Dunno,it would add some interesting play options, and would provide a way to test possible AI solutions.

Their OP posts don't mention far more varied rescue missions, including crafts, not just random kerbals. I also don't see that contracts are being generated based upon good launch windows to different worlds, or is it buried somewhere? FP indeed looks pretty good, but I don;t see it saying that it does a few things contracts really need to do. They need to never offer a completed contract again, for example. If you already tested the sepratron in condition X, that contract needs to go away forever, for example. Ditto planting a flag, etc.

Another idea regarding rescue operations… what if there were novel parts that you could not unlock via the tech tree, but you COULD unlock via salvage? The Union of Democratic People's Kerbalist Republics has a stranded spacecraft around Duna, for example. They have their own Nerva type rocket engine that has different Isp vs thrust, has an electric generator built in, etc. Maybe they have a different capsule, too. The mission is to save the crew, but if you return the spacecraft, you can then research the parts... People like unlocking new parts, seems like a neat way to require task completion to get some.

I'd like to see a few, as I posted here in Suggestions. In short, they are: 1. Have a launch window calendar suggest missions based upon good launch windows (explicitly given a time range for launch dates for min dv), regardless of your point in the campaign. Ie: if there is a good window for Jool coming up, you might get a contract to explore even though you have not even landed on the Mun yet because the dv is low in 2 months. 2. Novel rescue/resupply missions that include competing spacecraft, not just random stranded kerbals.

I agree. The real issue is that I imagine most rovers would be collapsed at some level, and assembled in situ. I made something vaguely similar to your (very cool, I might add) design, and I took up the rover in a fairing, and used an assembly tug I built (monoprop and lives docked at station) to stick it to a lander stuck to a NERVA transfer craft.

There needs to be a better solution for rovers. I landed one, but I consider most options shown to be pretty ugly.

I saw that they are working on the movie 3001, which made me think of the movie 2010. That made me thing of Discovery tumbling (nevermind that she should have been spinning, unless the spun up section was going 90 degrees from how I imagined it). Anyway, are spin states kept track of? It might be cool to have some salvage missions that require docking to unlock the ship (instead of using ][ to control the dead ship, you'd have to board them). If the ship was a decent size, and it was spinning, the boarding via EVA would be very difficult if the "wreck" was designed with no hatches near the center of rotation. The only issue is that clamp-o-trons don't care about orientation. It would be fun to figure out a way to have to dock, but first match spin (like the Pan Am shuttle in 2001). Any other ideas for challenging rescues/salvages? The goal is novel mission types that vary in terms of dv required (polar orbits, etc), as well as perhaps even other piloting skills (matching rotation, if there is a way to make that happen). Another idea would be a kind of "procedural" mission. Say there is a rescue op you need to do. Normally, the game would pop the unit (always a kerbal so far, but assume a ship) into an orbit, and show you that orbit. This would put the target in a region… say orbiting Duna. Your data on the orbit would be "based on last transmission" or whatever. You'd not be able to see the target (no more tab key letting you shift worlds unless you have a powered craft in orbit). When you get to the point where the target becomes visible (you are about to enter the Duna SoI, or somewhere very close), the game would generate the actual orbit… it would do so knowing when your encounter with Duna will be, when the periapsis is, etc. It would make the orbit such that it will decay, forcing the player into a race against time (maybe an eccentric orbit that aerobrakes a little. Obviously in this case it would by definition be inclined enough to avoid Ike (freeing the game to procedurally generate the orbit without worry).

Like the current ion engine it would be magic, not a solar sail. A sail requires constant acceleration, and the only way to do this is to have the ship remain the focus constantly, and you can have at most 4X time warp.

The game doesn't do any constant acceleration engines right now, so they'd be absurd (as are ion engines as they are now).