Innsewerants

Hi all, I wanted to remind you to back up your csv files before updating ksp when 0.15 rolls in. Just in case something breaks, the data gathering for example. Some updates on what I\'ve been doing during my long bouts of forum silence. -Working on in game gui to view small maps and planetary information, live data view etc. (sneak development preview, note the interface has had 0 attention until I figure out which of all my idea\'s are doable and which aren\'t, the 2 artefacts are just test values. screenshot1, screenshot2 The windows will likely shrink, the planet information is a separate window you can toggle on/off. -Controlling mapgen from ksp (on windows) (Won\'t allow you to use a window exe (mapgen) with mac, but you will get the ingame map and the texture will be saved, you wont have the options of mapgen though . I\'ve seen smitezero\'s java based viewer and it\'s coming along nicely, I am hoping this will provide a solution to mac users for rendering big maps. -Yet another new code for the beam, it shouldn\'t become an infinite resolution minibeam at the poles, my bad. This is a big feature update, my code has grown to 4 times the length of the current dll so this may be a while before it\'s out and well lets just hope 0.15 doesn\'t break everything. I still need to work my experimentations into a usable gui, currently it doesn\'t keep artefacts on kerbin and mun separate for example and the gui itself needs some layout-ing love. congrats corax if no duplicates you are about 1/3rd to a full set lol

Hi all, I\'ve just updated the mapper with an important fix, crashing on pause, or to be specific, unpause, should now be fixed. I\'m sorry for the lack of updates or communication on the forum. Stuff in \'real\' life... I\'m still working on new features ad improvements and dabbling with 2 more mod idea\'s, it\'s just going very slow atm but I felt that this fix could not wait that long though so a little intermediate update was in order. I will backtrack the message on the forum soon, I\'m still overwhelmed at how it grows, thanks everyone for your interest in the mod. -Innsewerants

Yes I will look into that, it would increase the size of each write to the file, I don\'t dare predict how ksp will like that, it has crashed on me just due to its own log spam when I was testing a lot of read outs. For now I\'m on a short break after this release, have seen too much monodevelop and too little sky and clouds lately... wonder why.

The code runs when ksp determines its your active flight and stage. Scanning from inactive flights could be doable, it was the first bug I removed because it was unintended behaviour before I ever put the first release on here lol. How practical that is, is another question entirely. You would have seperate instances of code trying to access the same files simultaniously, which might not be a problem but it spams your console and ksp log file to death and that has had the effect of making my ksp crash happy to an extent that I\'m pretty certain that is why it was crash happy. Mapping works in map view while its the active flight though (see the faq I added to the opening post) I might experiment with multi-sat simultanious scanning and see what happens. If I get it to a form that doesn\'t spam console to death and makes my ksp crashhappy I will implement it. About pausing causing crashes, I will experiment with that but haven\'t had any problems myself with crashing using my mod and instances where the game pauses. I must add though I only use mechjeb and the mapping module, I don\'t have any other parts or mods installed. about munolith Hmm yes I could add artifact detection I think, I would need to add a column to the data file format and add it to mapgen. I was thinking about something similar recording your landings and adding an argument to mapgen to add these sites to the map, same basic idea should work for artifacts. I will ask the maker of that mod if he doesn\'t mind me infringing on his turf a bit as it were though I could start writing it now but it just seems civil to ask first and for the landings I was thinking of an icon or small circle graphic not just a discoloured pixel? No this is not a problem, the sats append data to the data file of the body they are orbiting. So they just add data, the module itself does no checking for duplicates or any other fancy stuff, trying to keep the module part as light as possible seeing as ksp itself isn\'t always the most stable and some people run systems that can (apparently) barely handle ksp vanilla.

v2.0.0 is out. Kerbin mapping, thicker 'beam', high altitudes, new projection options Thanks to guicam for figuring out yet another way to get elevation data which doesn\'t seem to go all nutty. I\'ve listed the new features in the features etc spoiler tag, there are quite a few new things, updated the arguments section with all the new mapping arguments. The natural colour scheme doesn\'t scale yet though and is more or less optimized for Kerbin. I wanted to get Kerbin mapping out on the forum, will do a small update of mapgen when I have tweaked that option a bit more. Hope you all enjoy, Most questions, previous problems etc are irrelevant now, different code. If any problems pop up, please let me know and I will get to it asap. Now I only hope I didn\'t forget to pack anything and wake up to a zillion 'it doesn\'t work' replies.

I\'m not really keeping it private. I\'m just more on irc as many other coders are and the line of communication is more direct and timely over there. Run into a snag? r4m0n will be on to talk to directly for example, no waiting for the forum. The main reason for a bit of privacy was that I wanted to surprise people with Kerbin after all the talk of impossibility instead of getting hopes up and potentially failing to live up to them and didn\'t expect this problem to be so difficult to solve and take this long, but I have literally run out of things you can check on pqs and haven\'t found anything that can at the least be a check to stop collecting data when it happens or prevent this odd problem from happening to begin with. It\'s okay though, my main concern is people getting their data polluted if they use experimental code and share data collected with it here. My apologies if I seemed like coming on too strong.

-In the game structure parts need to be activated by the stage they are in to have their module code run, I could run the code all the time but I doubt that\'s the intention of Squad letting us play around with making modules, you would also end up with 5 seperate instances of code trying to access the same file at the same time which might lead to weird problems and a load of console spam. -Too much time acceleration throws measurements out of whack. I\'ve have tested what you said from very early on before I made the mod public. more than 50% of my time has been spent verifying it works or not and the data is good or not and not writing code. (Well I usually code something in the mapgen tool while I leave ksp sat gathering test data after changing module code but still.) I am working on a completely different method of scanning which should make most limitations and questions being asked at the moment irrelevant, but I have run into a big problem with it, after a few hours of data collecting showed some really weird values and further testing while staring at the satellite and console with a zillion print outs on various things showed it\'s ksp giving me weird values and nothing in my code. About 40 variables onwards I haven\'t found one or a combination of that lets me check for or prevent this happening. I\'ve reached out to Mu yesterday about my findings to hopefullly at minimum confirm the possibility or impossibility of doing scanning in this new fashion. Because it is taking so long I am thinking of porting over the wider beam to the old scanning method currently in use and release an intermediate update together with the new mapgen features. PLEASE do not use this code, it will give crap data which might be hard to detect at first on the mun but it\'s really bad crap data. This is the feature and code that\'s taking so long, it takes so long because the interval at which it gives bad data is really long and I only discovered it after an overnight test when a render showed it overwriting land with ocean and vice versa. I check my code/data overnight and in depth before making code part of the module and have tested all the variables you can get direct from pqsController and found 6 or 7 ways to get terrain data that way but they all do the same, almost like inverted data, crap in long random intervals. @giucam Scan for a really long time, it screws up. It\'s the feature that\'s taking so long and delayed the update past the planned weekend. After a while you\'ll get continents over ocean, ocean over continents. Thanks for pretty much forcing me to reveal the surprise before it\'s confirmed working lol. Please don\'t mix data got with the new method with csv\'s being shared. Hours of checking revealed it, I went hallelujah just like you at first and then checked the data of an overnight session of scanning while I slept and the data seems to invert randomly, at least it drew a swaft of ocean through where it before found land and after watching console output on everything in the pqscontroller while watching my sat confirmed it. I have tested every possible variable you can read on pqsController and they all do the same random screwing up of values in concert. Next time please pm me, in all likelyhood I already tested it and am not using it in the mod for a reason. If it\'s something I haven\'t tested and I confirm it\'s working I will give credit where it\'s due.

As it is no, Kerbin\'s atmosphere gets in the way of being low enough for the terrain to be there to be scanned. I have been looking into ways around this since the beginning but so far no luck. Progress on this is also slow because every time I think of something that might work I need to leave a sat orbiting for hours just to test if it does or not. Mapping every body is a feature I want to have so hang in there I\'ve not given up on safer more realistic altitudes and perhaps Kerbin.

My current dev build has, expect it sooner rather than later when I get around to a public update. --- The part model and animations are now all done and working. Still working on a feature I really badly want in the next update, but if it takes too long I will do an intermediate update with the new mapgen features and the part model.

MapGen only works if there is a data file in the same directory, in a command window you\'ll get a file not found, windows kind of hides that from you run with doubleclick. Try and dip a bit lower if no csv file is being created. the terrain might not be close enough to you to be of sufficient detail. When I am finished with the new part model it will have a soft light effect when its actively getting data, for now you can check the console (alt-f2) it will say start scanning whenever it starts and stop scanning when it stops, this height it will probably switch a lot. I added that to mapgen, if you run it from console as of next update it will tell you the highest point in the data processed and the location. Good idea

No the detailed terrain collision mesh only gets loaded where you are. you could have multiple ksp installations perhaps each with a sat and separate csv you can combine later perhaps, not sure how well that would work out depending on your system. As promised I will try to figure that out, patience might be required sadly. Data in the datafile can go beyond those bounds because ksp isn\'t switching fast enough on occasion, so instead of switching from -180 to 180 e.g. it\'s late and spits out -181, -182 and then catches up and continues correctly 177, 176 etc. I correct for this in the mapgen, I wanted to keep the code in the actual ksp module as light as possible and didn\'t consider people might try and feed this data to other applications. I have been looking into this from the beginning and planning to change the scanning as soon as I got it figured out. That would be weird because the whole data collection code is only run when an If statement checking if the body orbited is the mun is true. That depends on the quad count ksp spits out for you, I\'ve got intermittend high terrain telemetry from as high up as 4.8 but it misses almost everything not at the poles. at 4.2 my quad count is about 2000-2600 below 3.8 I get as high up as 3000+ so I figured a minimum of 1400 would be a good first bed as a limit. I\'ve had many safe polar orbits with a large percentage of terrain scaned between 3.5 and 3.8 km, it\'s fast becoming my favourite orbit for this, most of the terrain scanned but not impacted upon I run the options high on a phenom II 3.2 ghz, 3 gb ddr2 ram, nvidia gtx 260 216core factory oc graphics card. I only get a short load flight stutter usually. The whole altitude limit might be removed altogether at some point as I learn more about this all. As it is coded atm there is no custom scanning beam to widen lol. But yes this is on the to do/try list --- Edit Here\'s my data from 1 polar mapping satellite, left on overnight twice and well every time I test new things for the mod, I hardly get to map myself now http://dl.dropbox.com/u/49791830/ksp%20addons/isa_RAM_topic/isa_RAM_Mun_Data.csv I\'m going to work on something that will properly merge data when I get to it.

oO I never thought this thread would grow like this, I hadn\'t checked since my last reply, sorry for that. Worked on some features and am a bit under the weather. Before I get to reading everything since my last reply here\'s what\'s coming so far when I upload the next update: -As requested: meridian gridlines -Mercator projection option -A dedicated part model! http://youtu.be/HJ14Fm-Bboo I still need to figure out how to keep it from rotating inward to the ship and add a light to signal when it\'s active instead of the console test spam. The rest is less obvious, the console line arguments for the mapgen can now be in any arbitrary sequence and I\'ve added a few. Some tweaks and stuff More might be added it all depends how long I will struggle with the animations of the part.

My process is really just go at it lol I really wanted something like this but no one seemed to be working on it or they deemed it impossible. I did some javascript in unity recently so I thought I would give c# another go motivated by ksp\'s module system. I always failed at c# attempts before lol. Basically I opened monodevelop and my browser whenever I hit something I don\'t understand I\'d google the c# reference material or a general google search if the reference didn\'t help me understand and use it. That\'s basically it. I don\'t learn well from reading code examples or most tutorials, I just write click compile, it will bug me about things that are wrong and while trying to fix them I learn. Use lots of console output to see whats happening when testing the code and go from there again. I think it\'s called being self taught ? I learned doing 3d and photoshop pretty much the same way. It\'s okay everyone, I kind of agree with protecting certain modules source but access through modules is pretty stretching atm so I understand the problem. I would have included it before if a certain someone hadn\'t deleted the guidelines as Damion already noted himself in a private irc chat. I\'ve packaged the source with the zip. If it needs a separate link please let me know. I just don\'t like to show my newby code lol :-[ I had this happen if I left mechjeb staying in e.g. prograde lock and the thrusters activated

It\'s up, you only need to replace the mapgen

It\'s fixed, expect an updated version any moment. here\'s your data generated while you wait

Question was this csv recorded on a computer that, how to put it politely, doesn\'t run ksp quite as well as you\'d want to? That might be important for me, because I never seen coordinates like these from ksp before. and I have, *checks, 2.47 mb of 'srtm' data of my own to compare to collected with this version of the module.

Weird even if the mapgen crashes the console window shouldn\'t, it should just show prompt with amystical exception message that\'s usefull to me. mapgen doesn\'t do anything with the command window, it wouldn\'t know how lol, I don\'t There is some data you collected I need to check, I think you\'ve found an oversight of mine but this needs some figuring out, expect an update soon when I have it sorted. Important to mention maybe, the data itself is fine lol.. it\'s how mapgen deals with it. ksp can give some weird lon and lat\'s I thought I accounted for all of it but apparently it can be weirder

hmm indeed same for me and its surely not a typo or windows wouldn\'t have asked to overwrite my test csv in my dev folder lol Give me a second to try and figure it out Okay thank you for giving me the csv file really helpful as it lets me run it here and see what happens, add some debug output (it slows the mapgen to a crawl that;\'s why they aren\'t in the public) Are you sure it gave you file not found and not an exception? I\'m getting an exception in command prompt. I will need to research this one. 'good' find lol

csv was the plan but ksp likes . for decimal, c# likes , so I decided to use neither as the seperator. being a noob at coding I didn\'t want to complicate things, wanting something out to the community to play with. For now you could use software like notepad++ to do a quick replace. I suspect you are not getting a high enough quad count with the graphics settings and your distance from terrain, I need to add this to documentation. If you check alt f2 for a while, the game will give a message with a number when quads are updated. I will put it on my list here though, as I\'m learning coding it might be better to change that and learn than just changing to extension to a 'propriety format of mine' lol thanks for the suggestion. About the scanning, awesome. I do not possess the knowledge to do that, I could code a scanning raycast but I mean the math involved to correct for angle etc. not a clue. I\'m reading about different map projections so that might be a likely feature in the time ahead I just can\'t give any estimates because of me just starting with it. About the detail limit, it might prevent some people of using my module. I didn\'t make the quad count requirement super high but, well, the game updates the quad count as you get nearer to the terrain so a low setting would make that point potentially when you are just above it. I need to add this to the documentation. I could activate a lowres feature in the cfg but I don\'t know why anyone would want a map of their orbital altitude lol. This continuous stream is actually once per second actual time If you get framerates like 2 fps I would suggest to turn down the graphic settings, it will prevent you from using my module for mapping from orbit. but I\'m thinking of adding a feature to the mapgen to render a specified area, you could still make maps of small area\'s you explored with a rover for example. Note down the coordinates of the corners of the area you explored with mechjeb e.g. and then make the mapgen spit out a map of that area. If people want this I will see about adding area map generation. But my module is very lightweight and doesn\'t seem to affect my own ksp performance at all, I purposely left out anything dealing with doing math or such out, it\'s raw data collection only, the only \'big\' code is really just ensuring good data. I could decrease the scan and file write operations even lower, a cfg setting perhaps. Might help but I doubt it, it\'s writing max 23 bytes once per second shouldn\'t worry any drive. writing takes mere ms, the time in between it\'s all idle waiting for the next second. The best way to use the mapgen at the moment is using the command line. If you don\'t know how this works, a little guide: -Have a window open with the directory the csv gets written to and the mapgen resides. -Select the full path in the address bar of the window (I use w7) and copy it ctrl+c -Go to start -> run type in 'cmd', press enter. -You\'ll get a little window with a dos like prompt. -type 'cd ' right click the window and select paste, press enter -You\'re now in the same directory just from the command prompt. -type 'isa' and then press tab it should auto-complete to 'isa_RAM_MapGen.exe', press enter It will generate a default small map from your data. To use console arguments make it e.g. 'isa_RAM_MapGen.exe 1000 scale' to generate a 1000px wide image with a topography colour scale at the bottom. I haven\'t added an option to render long/lat with it but long/lat gridlines have been requested I will add it to that. If you use more than 1 argument on the command line, the first needs to be map width, this is because I have 2 potential numeric arguments, I think I will change that all to be more robust. The mapgen is really interesting me, I always like doing map stuff as a kid, now that I\'m figuring out c# and did this I think you can expect that I will add some features to the mapgen in the coming time, some might not be all that useful though haha I think I wrote in the readme and this post that it only maps the mun for now. Mapping Kerbin is not realistic at the moment because there is nothing to map when you are orbiting 70 km away from the surface or higher. The game changes the resolution of terrain depending on how close you are to the terrain and eventually makes it a low poly sphere I suspect and possibly drops the collision mesh completely. Iow there\'s nothing to scan but -1. I might enable a separate kerbin file if there is demand but you\'d need to map with a plane or just do area maps with a rover or something when I get around to implementing generating area maps.

Yes but what does that say about the accuracy Please keep in mind it\'s also a learning project for me. Learning c#, mapping and making a ksp modules in less than a week . My accuracy is over the top for a game planet but I want to learn from doing this. I could make mine fill out the blanks using just the bit of data available I guess but hmm wasn\'t really what I was after. I can render a full and inaccurate (you can\'t consolidate data that\'s not there.) map now with my mapgen at 50px wide and blow it up to get more or less the same result. One of the idea\'s is to be a bit like srtm and provide data not images, I just happen to have written an image tool to accompany it as well And just saying I\'m still regularly looking up how basic stuff works in c# regularly, I\'ve been a c# coder now for all but 4 maybe 5 days starting without knowing a word of it lol extrapolation.. I might add that as an option if I figure it out I just don\'t like it, that will not tell you about any unexpected cliffs, it will not predict small craters, (actually im missing craters I mapped in your image but your image makes it seem like its been mapped..) But then I wrote it all this way so people can do with their csv\'s what they want So keep going with that tool if you like it I like the elevation lines, might add that when I figure out how to Today is breaktime, worked on it since thursday, tomorrow I will start and see if/what requests there are or what I would request of myself and work on some features to be added unless major bugs pop up that need to be fixed

It\'s not optimized to work at any altitude, it checks the collision mesh detail and if it can get radar altitude readings from it, In short it cares about altitude above terrain not the orbital altitude indicated at the top of the screen as that really says nothing about actual altitude above the terrain. if they add a plateau 10km high to the mun and you orbit the mun at 12 km it will scan the plateau fine, just not the rest as it would be out of range, most of this is because of how ksp works not my module, So far I\'ve only been able to prevent bad data collecting due to whatever the game is doing, I haven\'t been able to force the game to do my bidding so I can\'t change these limitations at the moment

You can\'t orbit low enough on kerbin due to the atmosphere, when you\'re high enough the collision max drops detail, a little higher still and it disappears, there\'s just nothing to measure unless you are well below 15km on kerbin. The game doesn\'t load high poly collision terrain until you reach a point where the game thinks it\'s needed because you might impact or try to land. To save resources and keep performance up. So essentially if you are high enough (and on kerbin still well within thick atmosphere) there is no terrain, just an optical planet not one I can scan. I will keep looking for ways around that. That\'s not enough information to do anything with. How about checking console in game alt-f2 and giving me whatever errors it might spam on it, or if its the mapgen run it in console and copy/paste the exception it\'s given. I can\'t do anything based on nothing which is what you just gave me 4 - No map view causes useless data so my code prevents recording when in map view. You can use map fine it will just not record while you are in it. 5 - Unless I packed an old module by mistake the csv is already name for the celestial body 'isa_RAM_Mun_Data.csv'. So if any other planets come along or I find a way to map kerbin you\'ll get isa_RAM_Kerbin_Data.csv and isa_RAM_planetx_Data.csv etc. When that happens I will bring out an update because currently it\'s just disabled for other bodies as there are none that can be scanned atm/ 6 - the image in my signature is fresh data collected overnight by a polar sat orbiting 3500-3800 km, hasn\'t hit terrain while I was sleeping and gathers a good portion of elevations, I will add more sats to append to the data when this one has 'finished' 7 - You\'ll have to explain what you mean to this coding newby -Is the stage the part is in active? -Do you run the terrain options very low? The code has a minimal level-of-detail of the terrain check. -Is the sat low enough? at 4.0 km it will only find high enough areas now and then so you might have to wait for it to find a high spot or drop a little lower. -you can check console (alt-f2) if it is spamming errors. Yes, I could have it active always but if you arent the active flight and close to the mun the game wont load the collision mesh or will load a simple low detail flat-ish collision mesh for the mun so scanning that would just turn out bad data. The last part is not really true, or my sig image wouldnt exist. You can scan from any orbital height but if the terrain directly below you isn\'t high enough it will read -1 or your orbital altitude (this is the game not my mapper code) I scanned the highly elevated polar region with a 4 - 4.2 km satellite. The terrain gets to high enough detail and spawned +/- when you aren\'t more than 2000 m above it so recording 3 km high terrain on the poles from 4.5km alt works fine but scanning 1 km high terrain from 4.5km alt wont work. Like I suggested if you really want to map it all, have a polar higher safer sat to get most of it, my sig data was gathered overnight by a sat currently at 3550-3800 but it scanned half of it yesterday while at 4-4.2km. It\'s the height above terrain that seems to make the game decide if it loads collision mesh and at what level of detail. That\'s why my orbit path on the sig image has gaps, those are spots where the terrain wasn\'t high enough to be measured or the game decided to drop below minimal detail level or lose the collision mesh. same for toastar. the middle of the mun is pretty low terrain compaired to the poles. that;\'s why I suggested a higher safer polar sat for high terrain and other sats to record the lower terrain from the less safe orbits. run it from command line and let me know what it says, it will give a file not found but I don\'t know how to work windows interfaces yet, so if it does file not found from a windows double click it wont show a nice dialog (yet) the ; and . for decimal is currently hardcoded for the mapper and mapgen. I will think about it but I\'d have to change a lot around in the code. You could do a replace all with notepad++ e.g. to make the csv compatible with other programs, takes 1 second. I would be honored to be a member of the Kerbil Geographic Society I\'m thinking about making a community project to collect csv\'s and consolidate them, but a complete data set would be 11gb without any duplicate measurements atm. I kind of want to keep the option to generate 2048 maps at least and thinking of putting in an option to render e.g. a map of a region which would need that 2 decimal detail data. Just to make it clear again you can map from 4+ km but you will only fetch data at times when the terrain below you is in range. For people that use mechjeb you can spy on this somewhat, whenever true alt isn\'t simply your asl reading it\'s in range. The mapper does some more checks than just that though but it\'s a good indicator when it\'s scanning. It will also ouput started scanning and stopped scanning to the console everytime it does. I\'m thinking if I get around to a dedicated part model I might try and make that have a light that\'s on when scanning. If you have problems and want to help me fix them (It\'s my first c# learning project after all) please do the following: if it\'s the module in ksp: Check the console alt-f2 and if it\'s spamming something there. If so send me what errors is spamming so I can find out whats happening. If you don\'t get data remember it scans when the terrain below you is in range not simply below a cerain altitude. A 4.5 km sat will mostly only scan around the high polar region and not while crossing the lowlands around the middle. if it\'s the mapgen Run from command line so you can see if it spits out a file not found or if it crashes what the exception is it gives, these usually show me the way to whats wrong.

Yes I\'ve used MonoDevelop for this too, mine does not include mac development options though, I think perhaps because it\'s the one that came shipped with unity.

[Edit] Woops, My bad[/edit]

I will send you a pm if I get my hands on a compiler for mac and figure out how to do it .