MaxPeck

I don't know what happened there with that post. Anyway, I'll download the latest copy of SXT and install it and see if that matters. I just did a mod update for all my non-AVC mods, but I'll give it a shot anyway just to make sure.

Okay... V1.41 is hanging up on Stock Extension parts - Lack's SXT mod. The loader hangs on a couple of the parts, if I remove DTL, it loads fine, if I rename the part.cfg to part.old, it loads fine, so it's definitely something in these parts that's causing the issue. The game also doesn't become unresponsive, it just sits there... when I go into "Force Quit", it's not showing as broken, and the sayings on the loading screen continue to cycle, it just doesn't do anything. I thought maybe it was just taking a while to load, so I let it go for a good half hour with no progress. For reference, I'm running on a Macbook pro 13" retina, KSP 1.0.5. Here are the relevant log sections: [LOG 19:30:28.650] PartLoader: Compiling Part 'SXT/Parts/Command/Ke111/part/SXTke111' [LOG 19:30:28.781] PartLoader: Part 'SXT/Parts/Command/Ke111/part/SXTke111' has no database record. Creating. [LOG 19:30:28.830] DragCubeSystem: Creating drag cubes for part 'SXTke111' [LOG 19:30:28.836] DynamicTextureLoader: Unloading: SXT/Parts/Command/HECS-Ranger/partProbe/SXTHECSRanger [LOG 19:30:28.836] DynamicTextureLoader: Renderer: Cylinder [LOG 19:30:28.836] DynamicTextureLoader: Renderer: Cylinder_001 [LOG 19:30:28.836] DynamicTextureLoader: Renderer: Cylinder_002 [LOG 19:30:28.836] DynamicTextureLoader: Renderer: Cylinder_003 [LOG 19:30:28.836] DynamicTextureLoader: Renderer: Torus [LOG 19:30:28.918] DynamicTextureLoader: Unloading: SXT/Parts/Command/Ke111/part/SXTke111 [LOG 19:30:28.918] DynamicTextureLoader: Renderer: Cylinder_010

I'm not sure if this has been mentioned before, but I have a bug regarding the gaslight towers. If I link two of them together, or link them to the muncrete bases for the inflatable modules, and then I leave and come back to my base, there are multiple pipes suddenly leading between each tower/base. Sometimes they're plugged into separate ports, but sometimes I'll have two pipes going to one port and the game loses its mind when I unlink them to straighten it out. If I leave them alone, they don't seem to hurt anything, just looks weird.

Okay... thanks for that. I didn't realize KAS was just parts and KIS was doing the heavy lifting. Same, thanks for pointing me to the answer.

Not being all that versed at the ins and outs of writing the guts of mods, but just as a matter of speculation, how difficult would it be to create a small part that you attach to a satellite or other communications relay station that designates it to the network as a repeater station? Not to the exclusion of other antenna, but something that assigns priority within the network to that particular antenna? Maybe give a bonus to that satellite or station if it has multiple antenna, like a real repeater station would have... just a thought.

Noooooooooooooooooooooooooooooooooo!

So there I am... walking around on the Mun, when I suddenly notice my Kerbal is showing as suboptimal. He's outside my base, literally surrounded by antennas, both stock and from mods, mostly of the DTS . There are at least 3 communications relay satellites overhead, and all of them have a clear line of sight to Kerbin. Is there a way to manually designate relay stations, or would you consider adding such a thing? It seems to me that any one of the many antenna within line-of-sight, if not throwing distance from my guy walking around should have been sufficient to the task...

Me too. As soon as I call something a base, State Funding just gives up. It still works, but the status window is blank. If I go into the Tracking Station and make the "base" into anything else, State Funding starts behaving again. I've also noticed that it's a little unforgiving of reverts... if you screw the pooch and have to revert to launch and try again, it counts the kernels killed and ships as destroyed anyway. I've had to edit the state funding conf file a few times now to undo things that "never happened". It also sometimes counts dropped stages as destroyed ships, not sure why.

Simple question that I haven't been able to find an answer to - what is the difference between the wrench and the screwdriver? Is it purely aesthetics or do they have a different role? The reason I'm asking is that the last time I played, I had installed KAS after I already had enough tech to have the screwdriver, so that's all I've ever used. I started over with 1.0.5 and now I've gotten to the point where all I have is the wrench and I can't seem to be able to attach anything to anything else. I'm trying to build the buffalo rover from Pathfinder on the surface of Mun, and I can't get anything to attach to anything else. No surface attachments, no stack node attachment, nothing. And yes, I'm holding X and/or H when I try to connect pieces. Thanks!

You know, the problem I've always had with this is that it's impossible to train Kerbals for a mission. Right now, the only way to level up your guys is to actually just do the missions, or to do a bunch of unrelated tasks. In real life, the more you do something, generally, the better you get at it - as you gain experience, you also gain skill. Granted the law of diminishing returns means that the amount of skill gained by training isn't infinite, but there's still something to be said for proficiency. I mean, imagine if the only way we could train the Apollo astronauts was to make them do a bunch of unrelated tasks, and then actually fly to the moon - instead they practiced flying to the moon. It would be like if they ran the simulator once, and then said "nope, already got all we can out of that, let's go plant a flag out back and see what that teaches us." They ran the simulators hundred of times and learned from it. If your guys to mission-essential things, they should get experience for it. Maybe less each time, down to a minimum value, but still something. I'd rather have my engineers repack chutes on Kerbal a bunch of times before trying it on Duna. Likewise, I wouldn't dream of trying to fly to Mun with a level 0 or 1 pilot - and how many people don't? They pack a probe core with more skills than their pilots have and bootleg having a level 3 pilot aboard. So why not have a mechanic where pilots can do training flights and increase their level? Successful completion of x number of flights yields xp. Ditto for scientists - successfully doing x number of science tasks yields a certain amount of xp. I have no idea how that would work with engineers outside of KAS, but there's probably something comparable they could do. Like someone else said - the current system is great as an achievement system, but I don't like having the Kerbal's skills tied to it. I'd rather see skills increase with use.

They are, I realized afterwards that they were not not the best to use as an example. The main difference between them would be play asthetics - the CT32 is a mast, the Omni is a dome. Everything else in terms of range and transmission speed are identical. Bad choice on my part. Anyway, thanks for considering it. Just out of curiosity, is RoverDude's stuff that he's working on for 1.1 going to render this all moot?

Bandwidth is the game's designation, not mine. I'm just rolling with that they provide. Oh, and I made a mistake, I was using the Communitron 32, not the communitron 16. Sorry about that. I do use amps in place of electriccharge just because it's easier to type. So it's not necessarily about the number of mits/second, it's about the number of packets/sec and the power required to blast them out. They are both capable of broadcasting the same 6.6 mits/sec, but the CT32's packet size is 2 mits to the Omni's 1. That means that in one second, the Omni can broadcast 6.6 packets to the CT32's 3.3. It's not about the amount of data sent, it's about the throughput of packets and the power used to send them. Maybe using two similar antenna was a bad example, let me use antenna with different specs, both from Lack's mod... Name Bandwidth Packet Size Packet xmit time Packets/sec Packet power Signal Strength Communitron 32 6.6 2 .30 3.3 15 49.5 Scanning Dish 20 20 1 1 85 85 So here, the scanning dish would have a greater range: it broadcasts 1 packet per second @ 85amps, and so it's broadcasting 85amps/sec, whereas the CT32 can broadcast up to 3.3 packets per second at 15 amps, but only puts out at most 49.5 amps per second. So the scanning dish is sending more power per second than the CT32, regardless of the amount of data being sent. Another example using stock antenna: Name Bandwidth Packet Size Packets xmit time Packets/sec Packet Power Signal Stregth Communitron 16 (The correct one this time) 3.3 2 .60 1.65 12 19.8 Communitron 88-88 20 2 .1 10 14.1 141 So here, the stock 88-88 uses 14.1 amps to send one packet, and can send up to 10 packets per second, so it's blasting out 141 amps/sec. the CT16, on the other hand, only sends 12 amps per packet and with the much lower throughput, can only send 1.65 packets per second, so it's only broadcasting 19.8 amps/sec. So it makes sense that the range you've assigned to the 88-88 (80,900 Mm) is much higher than what the CT16 can do (67.1 Mm). As for frequencies and such, I wouldn't neccesarily worry about that too much. Just assume everything transmits in the same band, and that band is universal. Call it subspace radio and thank Star Trek. Whatever. In this case, what your mod would have to do is recognize something as an antenna, read the bandwidth, packet size, and packet power, and use those to calculate the signal strength. You could then apply a set formula to translate signal strength into a nominal range. I don't enjoy calculus, but when I graphed it, it came out somewhat curve-ish, so someone better at math should be able to come up with a formula to figure it, or just make a linear formula that you like and re-assign the stock antenna according to it. Either way, nominal range would be a function of signal strength. I don't do c#, so unfortunately I can't do much more than have the idea, but there it is. Does that make more sense?

I have a suggestion/request for you. I've been playing with the Lackluster Mod that has a bunch of antennas included in it, and so I had to go and write a MM cfg to include all the additional antennas. It was kind of painful, and I was left trying to come up with nominal ranges for a bunch of antennas using the stock values that are hard-coded as a baseline. In order to come up with something reasonable that wasn't a bunch of arbitrary values, I developed a method for calculating signal quality based on power and bandwidth. What I did was look at the bandwidth and packet size to figure out how many packets per second could be transmitted, and then looked at the power requirements per packet to figure out the power/second being transmitted, which I decided was the signal strength of the transmitter. I then ranked the antenna according to signal strength, graphed the stock ones and then figured out where on the graph the others fit in. Then I just overlaid that graph on another chart of ranges and used that to determine the nominal range for each antenna. I'm sure someone better than me at math could do that part, but the signal strength calculation is pretty consistent. For example: in Lack's antennas, you have the Communitron 32 and the Omni Radar, both of which have similar values: Name Bandwidth (mits/sec) Packet Size (mits) Packet xmit time (secs) Packets/sec Packet Power (amps) Signal Strength (amps/sec) Communitron 16 6.6 2 .30 3.3 15 49.5 Omni Radar 6.6 1 .15 6.6 7.5 49.5 So while they both have the same signal strength, and thus the same range, the Omni transmits data at twice the rate, so it's a higher speed connection over the same range. So in this model, the bandwidth essentially dictates the gain of the antenna - low bandwidth will require more power/packet (low gain) to achieve the same signal strength as a higher bandwidth, but lower per-packet power requirement (high gain). So the high gain antenna would utilize the same amount of power per second as the low gain, but transmit more data over the same range. Bringing this full circle, I guess my suggestion/request would be to code this method into the antenna range mod, so that instead of needing a cfg or MM config for each antenna, your mod could just read the bandwidth, packet size and packet power values and calculate the signal strength, and then apply a nominal range based on whatever scale you think appropriate. Did that all make sense?

Does your school teach analytical geometry? Physics? A lot of what you're doing in KSP revolves around elipses, parabolas and angular motion. This is a fantastic way to illustrate those concepts by having students simulate them. There's also a lot that can be done in terms of calculating the power output of the various motors, determining the amount of potential energy per liter of fuel, etc. Go to your library and do a literature search on use of simulation in learning, and classroom gamification. It would also help if you maybe convinced your math/science faculty of the value of it, and have them help with the heavy lifting with the administration. Are you a teacher? Student? What's your role in this?

xfrankie - that did the trick, thanks!

So I installed this, and the jump beacons are showing up in the VAB, but not the Jump Drive. I have my tech tree maxed out, I've watched the log and it seems to load okay, and the jump drive shows up in the part database. I'm not sure why I'm only getting the beacon and not the drive?

I've been thinking about this too... something akin to moving the two assemblies close to each other and then "welding" them together. Not welding in the reduce-part-count type of weld mod that exists, but something to the effect that if you can make two things touch, a Kerbal should be able to join them. Maybe make a "welding torch" tool that can do this? I don't know the programmatic gymnastics this would require, but from the standpoint of balance, it would make sense to be able to join things, regardless of how many other things are attached to them (i.e. command pods with parachutes attached to a rocket assembly).

Sorry if this has already been brought up, but I didn't see a thread for it. If there's already a thread for this, feel free to merge this. I've been experiencing an issue intermittently where I get a vessel into orbit, and try to warp to the circularization burn. After a few seconds of warping, the vessel is suddenly sailing out of orbit, but with no change in velocity, it just starts gaining altitude like crazy. Throttling up increases velocity, but does not alter the rate of ascent. The vessel becomes unmaneuverable like it's still on rails, and it won't let me re enter warp because it claims the vessel is under acceleration, even with all engines deactivated and/or the batteries dead. Trying to switch to another vessel or the KSC gives a revert warning. I didn't post logs because there's no crash, the vessel just goes rogue. Running on a MacBook Pro mid-2014. Running several mods, but appears to happen most often during a mechjeb auto-warp. Any thoughts?

Yeah, the OS X version is half-baked. Try this, it's worked for most of us, it seems.... http://forum.kerbalspaceprogram.com/threads/125288 Hopefully Squad will get their act together with the U5 update and actually give us a playable game. For my own part, crashes are fairly infrequent after using the steps described in that post, but when they do happen, that's usually my cue to take an intermission.

Hey, can we sticky this? There seems to be an influx of new mac users hitting the usual bumps in the road. Be a lot easier if this was at the top of the forum, instead of all these folks putting up new threads for an old (and well known) problem. Thanks!

I'm assuming your iMac has a retina display? If so... Two things... for some reason, Retina displays freak out when you use textures above 1/4... you should dial that back. Also, make sure you're running in low res mode. Right click on the KSP application in your KSP_osx folder and choose "show info". Under options, make sure "Open in Low Resolution" is checked. I was having this same issue on my brand-spankin' new macbook, and dialing back the textures to 1/4 or 1/8 solved a lot of problems for me. Might not be cinema quality, but it's playable.

Two thoughts after reading this... 1. It almost feels wrong to call this 1.1, with all the work and overhauls being done. If it were me, I would just call it 2.0 and consider this a new fork. U5 sounds like it will, literally, be a game-changer. 2. Reading the devnotes, I appreciate the amount of work you guys are doing on the game's evolution. I fully plan, once the next iteration comes out, that if it's stable and works as advertised, to make a donation to you guys for your efforts. Work should be rewarded, even if it's only just enough to get a 6-pack.

Okay... first question about killing the KSP process... I've only noticed the phantom KSP process when I have to force quit the game after it hangs. So that's when you'll want to do that. If I quit normally, it looks like everything goes back to normal. As far as running in low res- as I understand it, all that does is disable Retina for that window and allows it to run in normal resolution. As I understand the way that retina works, basically it's a super high resolution screen that scales down to normal resolutions to proceed really high-fidelity graphics. Basically, the pixels on the screen are smaller, so you can use more of them to draw and create finer resolution images. Running in low res basically disables that and allows KSP to run under a normal resolution. A conflict appears to be occurring for retina users because of the way OS X packs graphics in a retina display. As far as ATM - there is some debate if it's worth using right now. I think it is. What it does, in a nutshell, is take standard graphics formats (jpg, bmp, png) and compresses them on the fly into more a more efficient format - DDS - that eats less memory. Think of it this way, if you have a jpg image that's 128x128 - and jpg's store data at 24 bits per pixel (8bits for red, 8 for green, 8 for blue), that's 128 * 128 * 24, that's 393,216 bits per image. At 8 bits/byte, that comes to 49k per texture file. In comparison, DDS only uses 5 bits per color channel, so you get 128 * 128 * 15 = 245,760 bits per image, or 30k per texture file. The real benefit comes when your jpg files are 512x512 or 1024x1024 and you use aggressive ATM to reduce them to 128x128. A 512x512 jpg would be 6,291,456 bits, or 786k per texture file. Compare that to 30k per texture file. There's quite a bit of savings there. This doesn't even begin to touch on DXT compression. For your eyes, you won't see hardly any difference between jpg texture maps and DDS texture maps. In fact, most mod designers are already using DDS, which is why ATM might be a moot point. For those who aren't, however, there is some memory savings to be gained from using it.

Tom, Try the tweaks listed in this thread. They really do help. I'm pretty much able to play without crashing, unless I do something really stupid like have a bunch of windows open and then try to change scenes. I play with about 20 mods, including KIS, KAS and a bunch of parts packs, and while my graphics res isn't ideal, I can at least play for several hours without disruption. ///on another note/// I noticed something new this morning. I have a memory monitor app installed, so I can keep an eye on my system memory. I generally hover around ~6gb free space when I'm not doing anything, but after a KSP hang yesterday that required a force kill, I noticed my free memory never recovered. It was hovering around 1.5gb with nothing open. I force killed everything and didn't see any improvement. So I started nosing around in activity monitor and there was nothing in the memory usage tab, but under the CPU tab were two instances of KSP, one of which was running at about 55% CPU usage. I forced them to die, and immediately recovered the missing memory. So thought for the day - if you have a KSP hangup and force kill the app, go into activity monitor and look for lingering processes and kill them too. Otherwise, it seems like you get into a spiral of diminishing returns that leads to eventually running out of memory.

If I were going to describe it, it's like a big outer space hot dog roll. I think I see what you're getting at, but you really have to tweakscale it up a few notches to be useable.