NathanKell changed the topic of #RO to: Welcome to the discussion channel for the Realism Overhaul (meta)mod for KSP! Realism Overhaul Main Thread https://goo.gl/wH7Dzb ! RO Spreadsheet http://goo.gl/Oem3g0 ! Code of Conduct http://goo.gl/wOSv2M ! | [15:01] <soundnfury> Straight Eight Stronk (and) RP-0/1 is basically "Space Agency Spreadsheet Simulator" with a rocket-flying minigame
<Qboid>
[#277] title: Ocean bouyancy not occuring in 1.3.1 RSS/ScaledRSS Textures/Kopernicus -5,-6 or -7 | Did not test with -4, previously -3 worked.... | https://github.com/Kopernicus/Kopernicus/issues/277
Shoe17 has quit [Quit: Connection closed for inactivity]
awang has quit [Ping timeout: 186 seconds]
awang has joined #RO
Hypergolic_Skunk has quit [Quit: Connection closed for inactivity]
Probus has joined #RO
blowfish has joined #RO
ProjectThoth has joined #RO
Mike` has quit [Ping timeout: 207 seconds]
NathanKell|AFK is now known as NathanKell
<NathanKell>
o/
<Qboid>
NathanKell: awang left a message for you in #RO [26.02.2018 13:29:30]: "btw, if an engine only uses one config from ModuleEngineConfigs, should MEC be skipped and all the stats put into ModuleEngines*? Or should the other configs just be deleted?"
<NathanKell>
awang: Should be fine to just use ModuleEnginesRF in that case
Mike` has joined #RO
<ProjectThoth>
I have received another box of pencils.
<ProjectThoth>
I spent the previous hour grafting the ferrule and eraser from one onto the other, because I like that eraser more.
NathanKell is now known as NathanKell|Twitch
<NathanKell|Twitch>
here goes...
<ProjectThoth>
NathanKell|Twitch: Careful!
ProjectThoth has quit [Quit: +++out of cheese error+++]
Qboid was kicked from #RO by *status [You have been disconnected from the IRC server]
Qboid is now known as Qboid_
Qboid_ has joined #RO
Qboid_ is now known as Qboid
NathanKell|Twitch is now known as NathanKell|AFK
<leudaimon>
Mike` I may have reordered the rows to find the engines, that's why diff shows so many changes
blowfish has quit [Quit: Leaving]
Shoe17 has joined #RO
leudaimon has quit [Ping timeout: 186 seconds]
leudaimon has joined #RO
aradapilot has quit [Remote host closed the connection]
aradapilot has joined #RO
<Mike`>
leudaimon: well, that's bad for actually seeing what's changed though
<Mike`>
so for the csv approach we kind of need to kerep the sorting the same way i guess
<leudaimon>
yeah, I realised it when I made the PR... it's very difficult to work on the csv without filtering and reordering though
<Mike`>
we just need to agree on a sorting that makes sense and use that for the PRs, you can always sort them differently during editing. :)
<leudaimon>
maybe we can define an standard ordering
<leudaimon>
yeah, exactly
<soundnfury>
yeah, my kparts.py stuff sorts everything alphabetically (technically, case-blind ASCII) in the generated cfgs precisely so that diffs will be kept small
<leudaimon>
yeah... I was thinking maybe order by year/era then alphabetical, to make reading more straightforward
<soundnfury>
my source files are ordered only within a subcategory, by family (so my Orbital engines subcat goes "A4 A6 Aerobee AJ10..." ;)
<leudaimon>
ah, that is also a good way of doing it
<soundnfury>
btw, since the main remaining work on kparts is the tedious data entry, it's basically on hold until someone tells me "yes we will use this; it is worth your while spending hundreds of hours on it"
<leudaimon>
soundnfury I wonder if there is some way of converting the info already in the cfgs or csv to kparts, at least in a way that saves a part of the work?
<soundnfury>
I don't really see that working, because of the amount of human judgement involved in deciding where to put things, how to factor out commonalities, etc.
<taniwha>
soundnfury: that can always be done after it's massaged
<soundnfury>
e.g. complications like having a part with engine configs A, B and C, and also a part from a different mod for engine B.
<soundnfury>
taniwha: yeah but it won't, will it?
<soundnfury>
doing the conversion by hand means someone is forced to _look_ at every line, and thus spot "hey, this bit doesn't make any sense"
<taniwha>
can be done by loading all the configs, hashing on values, and using that hash in a dictionary
<soundnfury>
aaaaaaaaaaaaaa no.
<taniwha>
doing the conversion by hand is nothing more than sadomasochism
<leudaimon>
well, I really don't know any python so I can just believe you guys know what you are doing
<soundnfury>
doing it *right* means either by hand, or an AI-complete problem.
<soundnfury>
doing it quick-and-dirty means that the conversion won't actually improve things very much.
<taniwha>
soundnfury: actually, no
<soundnfury>
As I see it, the main benefit of the conversion is that it allows us to give _structure_ to the data.
<soundnfury>
but if you want to solve the problem I think is AI-complete, be my guest ;)
<taniwha>
because you can have the computer massage all the data, present it to a person who decides what to do with it, and then the computer does the more massaging based on what the person decided
<taniwha>
doing it by hand opens too many opportunities for typos
<soundnfury>
kparts will accept PRs that seem like a good idea at the time :)
<leudaimon>
the spreadsheet already has lots of structuring features... all the year/era/mod/class, the column listing all the identical parts, etc...
<leudaimon>
I think it could go a long way making a pre-organized thing
<soundnfury>
like I say, send patches :P
<taniwha>
anyway, the right thing to do is to get the computer to do as much of the work as possible
<soundnfury>
taniwha: I don't disagree with that
<soundnfury>
I just disagree on what counts as 'possible' ;)
<taniwha>
it wouldn't be that different to a revision control system's merge tools
<taniwha>
(conceptually. form would be quite different, I'm sure)
<soundnfury>
I think I understand what you're driving at, and I don't think I could write it
<soundnfury>
human-in-the-loop systems are all about human-factors/UI, and I Am Bad At UI.
<taniwha>
so am I :/
<soundnfury>
(Which presumably is why Kathan asked me to write that RP-1 UI xD )
<soundnfury>
anyway, tfw, ttyl
Hypergolic_Skunk has joined #RO
wb99999999 has quit [Ping timeout: 180 seconds]
qwertyy__ has quit [Ping timeout: 186 seconds]
<leudaimon>
anybody knows what cfg determines RCS thrust and Isp per Tech level? The only info I found in the global engine config folder of RO was what I understand as a thrust and proportional value for Isp for each fuel
<leudaimon>
no info for tech levels effect, nor what is Isp = 1
qwertyy has joined #RO
<leudaimon>
nvm, I found it, didn't think it was part of RF
Shoe17 has quit [Quit: Connection closed for inactivity]
Probus has quit [Quit: Leaving]
awang has quit [Ping timeout: 383 seconds]
awang has joined #RO
<awang>
!tell NathanKell* Might that result in some duplication of info between the stuff in Engine_Configs and the engine?
<Qboid>
awang: I'll redirect this as soon as they are around.
<awang>
soundnfury: Might it be possible to export from the spreadsheet to your python format?
<awang>
It's still automated, but at least you can look relatively quickly through the spreadsheet for patterns
<awang>
Especially since the columns are the same for each part, so differences/missing things stand out
<Starwaster>
What the hell is in the concrete in the launch pad that it can explode so violently?
<Maxsimal>
Fuel lines running under the pad, perhaps? Though I'm sure in RL these are built to make sure they they have the appropriate valves and releases to make that unlikely.
<Maxsimal>
Ugh I think Contract Configurator is just trolling me at this point - throwing exceptions on code that looks like stuff I've changed since I last rebooted KSP
<Pap>
leudaimon and Mike` - The default ordering should be by column 1 which is partname. That will be the one that should change the least.
<leudaimon>
Pap I tried it now, but I think it made things worse... maybe the first commits must be removed?
<Pap>
It is committing on top of commits, I would leave it as it is for leudaimon and this will serve as our starting point
<leudaimon>
ok, It should be ordered by part name now
<Pap>
Anyone know JavaScript?
<zilti>
Uhh, I rather not talk about it...
Shoe17 has joined #RO
BadRocketsCo has joined #RO
BadRocketsCo has quit [Ping timeout: 186 seconds]
awang has quit [Ping timeout: 383 seconds]
Technicalfool has quit [Ping timeout: 198 seconds]
xShadowx has quit [Ping timeout: 186 seconds]
xShadowx has joined #RO
aradapilot has quit [Read error: Connection reset by peer]
aradapilot has joined #RO
aradapilot has quit [Remote host closed the connection]
aradapilot has joined #RO
awang has joined #RO
<awang>
zilti: Why not?
<zilti>
awang: Because it's JavaScript
Senshi has joined #RO
<awang>
zilti: It's something we have to live with, unfortunately :(
<awang>
Might as well help make it less painful if possible
<zilti>
Well the few times I was necessarily doing something in-browser I used ClojureScript, so I'm of little help
leudaimon has quit [Ping timeout: 207 seconds]
leudaimon has joined #RO
<awang>
I really should learn clojure someday
<awang>
Know a bit of scheme, but haven't used it for anything major
<awang>
At least clojure is relatively widely used compared to some other lisp dialects
schnobs has joined #RO
<schnobs>
o/
<zilti>
Clojure is amazing imo. I also like Chicken Scheme, but good luck building something that needs a GUI with it. JavaFX is a breeze for GUIs with Clojure though
Daz has quit [Ping timeout: 383 seconds]
awang has quit [Killed (NickServ (GHOST command used by awang_))]
awang_ has joined #RO
Daz has joined #RO
<awang_>
JavaFX is pretty good indeed
<awang_>
Like the generics much better than the system Qt has
<awang_>
Have you ever tried Racket?
<zilti>
No, haven't tried that one yet. But I'm quite invested in both Clojure and Chicken right now, and I'm trying to next learn something non-lisp. Maybe some logic programming like Prolog.
Shoe17 has quit [Quit: Connection closed for inactivity]
BadRocketsCo has joined #RO
<awang_>
Looks like you're exploring quite a few things most people don't touch
<awang_>
You should try a dependently-typed language someday
<schnobs>
Design rocket, tool it all, then find out that you forgot to update the engine type: priceless.
<schnobs>
I'd really prefer if I tooling was a bit more class-y.
<schnobs>
Like, something I set up / pay for ahead of time, then I'm at liberty to make parts within the tooling constraints.
<schnobs>
But something on the order of "anything up to dia X, length Y", would be altogether too simple, wouldn't it?
<schnobs>
Though on the other hand, it's cool that I'm told which parts I'm lacking. I'd wish I could have the same service for engine configs and RCS layout (rather than silenty reverting to the best I have).
BadRocketsCo2 has joined #RO
<zilti>
awang_: I guess so! :) It kinda started when I decided to take a look at functional programming, after I mainly did Java for years and got bored by the routine. So I tried Scala. Didn't really like it though - Scala felt clumsy and somehow... meh. A syntax mess. Heck, the type system is Touring complete!
<zilti>
awang_: Then I got to Clojure, and it's really simplistic and well-designed, and being a Lisp, elegant as well. Now I've been kinda stuck with Lisps... Haven't even heard of "dependently typed" languages though.
BadRocketsCo has quit [Ping timeout: 383 seconds]
<zilti>
Hmm so weird. My moon rocket is much bigger and has about the same part count as my LEO rocket I use for polar orbits, it has stronger engines, and so on. Yet it's only half the price.
<awang_>
zilti: Sounds like quite the adventure!
<awang_>
I've done sort of the same thing, but probably not to the depth that you have
<awang_>
zilti: You should try SICP and see how fast you can go through the exercises, given you already know lisp and lisp-ish thinking
<awang_>
zilti: Have you tried any of the more "popular" functional languages, e.g. Haskell, OCaml, etc.?
<awang_>
Dependently typed languages are complicated
<awang_>
I don't understand everything about the theory or how to use them
<awang_>
But they promise the ability to prove at compile time that certain parts of your program match a spec
<awang_>
Or something along those lines
Shoe17 has joined #RO
<zilti>
awang_: Yea, I thought about SICP... but I'm so much more the tinkerer, I have to be in the mood to follow a path to that extent :D But it's definitely something I should work through
<zilti>
I was thinking about Haskell around the time I started Scala, and again when I started Clojure. But at that time, I wasn't "functional" enough, plus it seemed like a chore to even get into installing the entire tooling and to work into the Syntax. And now I first want to see some new shores ^^
<zilti>
Hmm proving a program follows a spec sounds a bit like https://clojure.org/guides/spec - definitely an interesting topic. Clojure seems to "steal" all those good features, like go-channels :P
Technicalfool has joined #RO
<zilti>
aand my KSP crashed again
Starwaster has quit [Read error: Connection reset by peer]
<schnobs>
hmmmmm.
<schnobs>
Can it be that Krash doesn't revert to previous game settings?
<schnobs>
Just did a sim, now TF is gone and I cannot load savegames.
BadRocketsCo2 has quit [Ping timeout: 383 seconds]
schnobs has quit [Ping timeout: 186 seconds]
<awang_>
zilti: Ah, fair enough. Sort of been down that road, but haven't had the time to keep it up
<awang_>
Heh, new shores are fair enough
<awang_>
Check out K
<awang_>
Some array-based programming language
<awang_>
Or MUMPS
<awang_>
(not for suggestions of good programming language design)
<awang_>
Hmmm
<awang_>
Does sound a bit like spec
<awang_>
Seems spec is runtime, though?
<awang_>
dependently typed programming languages let you prove things like "concat(vec<5>, vec<4>) = vec<9>"
<awang_>
Or proving that a variable cannot be zero, so you can safely divide by it
<awang_>
All at compile time
<awang_>
Or proving that a list is sorted
<awang_>
So you know your implementation must be correct according to your spec
<awang_>
(I'm bad at examples, so I'm looking them up -_-)
<awang_>
!tell schnobs That definitely sounds like a bug. Sounds like whatever event is needed to get KRASH to realize that the sim is over didn't fire, so TF is still disabled, I think.
<Qboid>
awang_: I'll redirect this as soon as they are around.
<awang_>
!tell schnobs I want to say it's not my fault, since none of my changes were to code that governs how sims start/end
<Qboid>
awang_: I'll redirect this as soon as they are around.
<awang_>
!tell schnobs Did you quit KSP normally, or did it crash (pun totally intended)? Or did you try to start a simulation, but then KRASH/KCT refused due to a missing part or something, then you had to buy that part and try again?
<Qboid>
awang_: I'll redirect this as soon as they are around.
<awang_>
!tell schnobs Not that those are the causes, but I've experienced some odd behavior in the past related to those. One of KRASH's advantages was that it's supposed to roll back state when the sim ends regardless of how the sim ends, so if it isn't doing that I'm a bit concerned
<Qboid>
awang_: I'll redirect this as soon as they are around.
qwertyy has quit [Ping timeout: 207 seconds]
qwertyy has joined #RO
Probus has joined #RO
<Probus>
o/
Shoe17 has quit [Quit: Connection closed for inactivity]