Jump to content

KSP Weekly: Refactoring Kerbal


SQUAD

Recommended Posts

28 minutes ago, Mako said:

I believe a roadmap would do more good than harm.

I believe, Squad's investors have a kind of.
And unlikely with specific engines and decouplers todo list. 
For example, something like that: "2017q2 v1.3. Localization to expand user base."

What about specific game features:

Spoiler

20120212-223344.jpg

 

Edited by kerbiloid
Link to comment
Share on other sites

32 minutes ago, evileye.x said:

It's like Star Citizen?

Honestly, my wording was influenced just a little bit by Star Citizen. Good catch.

 

26 minutes ago, kerbiloid said:

I believe, Squad's investors have a kind of.
And unlikely with specific engines and decouplers todo list. 
For example, something like that: "2017q2 v1.3. Localization to expand user base."

What about specific game features:

  Reveal hidden contents

20120212-223344.jpg

 

I'm not interested in a roadmap with specific dates or specific parts like engines and decouplers. I don't think the community needs that.

I'm interested in what Squad would like to work on before they call KSP complete. I would guess, and it's been hinted at, that there is an internal plan. I believe sharing that plan would do more good than harm.

(Unless the plan was to do something that would not be in the customer's best interest. If that happened to be the case we'd be screwed either way, but I'd still welcome the advanced notice.)

[Also, just to be clear, the part in the parentheses is not meant to be taken seriously. I don't think Squad is "out to 'get' us."]

Edited by Mako
Link to comment
Share on other sites

13 minutes ago, Mako said:

an internal plan. I believe sharing that plan would do more good than harm

They already had been enough incautious to public their ISRU scheme or reveal their plans about Gas Planet 2.
And still are answering - "why".

Probably, they prefer to keep some freedom of maneuver. In case if, say, hyping for one feature they would realize that there is another one, more important, and there are so many bugs and so little time till the next month release.

P.S.
I would suggest them to make "GP2" and other recurring questions fee-based.
Either their support team will gain a penny, or these questions will get forgotten.

Link to comment
Share on other sites

1 hour ago, Mako said:

The reason that's always been given for why Squad doesn't share information regarding planned, specific, long-term features or release dates is they don't want to upset people by cancelling features or missing deadlines.

You missed the point. The information does not need to be specific, or be a release date. Even just a business reason would do as to why they are not, and have not been, saying anything. Also, I`m not even sure squad has said this, I`m especially sure the squad we have now have not said it, that reason has more been given by other people on the forum. Like yourself.

I`d love to be directed to where squad has said this officially.

Anyway, `Some people may be upset` is not a valid reason. It`s a cop-out.

Link to comment
Share on other sites

6 hours ago, kerbiloid said:

They already had been enough incautious to public their ISRU scheme or reveal their plans about Gas Planet 2.
And still are answering - "why".

Probably, they prefer to keep some freedom of maneuver. In case if, say, hyping for one feature they would realize that there is another one, more important, and there are so many bugs and so little time till the next month release.

P.S.
I would suggest them to make "GP2" and other recurring questions fee-based.
Either their support team will gain a penny, or these questions will get forgotten.

I guess we'll have to agree to disagree, and that's okay. But I do want to say I don't think Squad should say specifics of what they plan to release or give dates for anything, and I do think they should say what they hope to accomplish by continuing to work on KSP.

I'll be blunt and honest with you: I don't see the business sense in continuing to upgrade the current product for existing customers with no way to monetize it. We're not paying a monthly fee for a service, as some software tends to do these days, and there's no microtransactions or paid DLC, so I'm really not sure what Squad plans to do. I understand localization efforts because it can open the game up to new sales for a new audience. I can certainly understand it if those are the kinds of business-focused KSP updates Squad is working on from here on out. Unless Squad has some way to keep making money on games development, I'm not sure how they're planning to keep the lights on and employees paid. That's why I'm so curious to hear about what Squad wants to do with KSP. They need to work on things that will make them money, or they won't be working for very long.

The only way Squad can keep making major feature updates to KSP without a revenue source is if they've already made so much money that they can afford to keep hiring and paying people from the previous profits. That is definitely not a sustainable business model, but it would gain them a lot of loyalty with existing customers and probably earn them some new ones if they announced that as the plan.

The only way Squad loses by announcing what they want to do with KSP is if what they want is bad for KSP and/or its existing customers. And if that was the plan Squad would lose in the long run anyway. I doubt that's the plan.

 

6 hours ago, John FX said:

You missed the point. The information does not need to be specific, or be a release date. Even just a business reason would do as to why they are not, and have not been, saying anything. Also, I`m not even sure squad has said this, I`m especially sure the squad we have now have not said it, that reason has more been given by other people on the forum. Like yourself.

I`d love to be directed to where squad has said this officially.

Anyway, `Some people may be upset` is not a valid reason. It`s a cop-out.

Please don't take this the wrong way, but it seems like you haven't read anything else I've posted in this thread. I was only answering your question; I was not endorsing or defending Squad's choices. I do believe that a Squad employee at one time or another has said something along the lines of what I said, but I could be wrong. I don't have a quote to reference.

If you've read anything else I've said in this thread, you'll see that I'm advocating for Squad to share more information. In fact I've said several times that I think providing a general idea about what Squad wants to do with KSP from this point out would do more good than harm. I may not feel as strongly as you do about all of it, but I have only ever agreed with you so far in this thread. I suppose my earlier response by itself didn't make this explicitly clear, but in the context of every other post I've made in this thread you'll see exactly what I meant.

Link to comment
Share on other sites

On 12/03/2017 at 5:04 PM, 0111narwhalz said:

To do that, you'd need to bring all the affected vessels into physics, because the game has no way to know what groups do what and what the potential impact of activating them is. The game would also need to unpack all vessels within physics range of the affected vessels, to ensure that things like carriers and landing pads don't enter problematic states. I don't think the game could handle that kind of thing, even if it is doable code-wise.

Would work juste like macros, minus graphics (some logs on the screens would do.. I would LOVE a probe telling "Timing time...")... I can't see a problem... Stages could be easily cleaned too: a "smart part" (any AG) could optionnally remove the actions it will trigger from UI... Doable right now with a (miniaturized... 50kg for a basic clock...) SP timer set to 0, but with a big mess on the top on the actions list... With generic actions like "jettison" on any object, an extended possible actions on everything)(Also : I can't get why they is no 3D orientation reticles for cams, no 1° usable compass, no heading given to objects........... not complaining... I can write code too, but why C# ??? :'( + having too much fun with the game as is) (what about any number of namable AGs? Guys it's basic DBM...)

And, when you talk work you talk money. If you want people obey, you have to pay them. Valid for the 3 parts (users, modders, Squad) : who's *working* doing what? I think Squad don't owe me anything AT ALL. Complain if you can pay people to work for you more than the price you paid.

Edited by Kenny Kerman
Link to comment
Share on other sites

10 minutes ago, Kenny Kerman said:

Would work juste like macros... I can't see a problem... Stages could be easily cleaned too: a "smart part" could remove the actions it triggers from UI... Doable right now with a (miniaturized...) SP timer, but with a big mess on the top on the actions list...

The issue is not activating loads of AGs on multiple vessels. Plenty of mods already use that functionality (well, a couple, but that's enough). The thing is, those mods only work (to that capacity) when the affected ships are loaded into physics. Do you want landing gear spawning inside of terrain or fully compressed? How about SRBs set to fire on AG3? Or docking ports decoupling parts in a causally significant way? Would you forbid the remote activation of groups with such actions in them? Load the vessels (and their neighbors) and perform some simulation time until they "calm down?" Or is it simply too much to ask of stock, when it's so easy to push things into invalid states?

17 minutes ago, Kenny Kerman said:

(Also : I can't get why they is not 3D orientation reticle for cams, no 1° usable compass, no heading given to objects...........

Well, it's one or more of these three options:

1 SQUAD won't do a thing because they don't want to. Perhaps it doesn't conform to their vision. Maybe they tried it, but it wasn't fun. Maybe it's a nice trinket, but not worth the clock cycles and memory.

2 SQUAD simply hasn't gotten around to it yet. There's always a lot to do, and this thing is considered a lower priority than the current projects.

3 SQUAD tried to do a thing, but ultimately could not because it was impossible with the engine, couldn't be optimized enough to keep up with the rest of the game, or is just beyond their skill.

23 minutes ago, Kenny Kerman said:

not complaining...

Perhaps you need some more :sticktongue:s, because it looks a little bit like a complaint to me.

25 minutes ago, Kenny Kerman said:

I can write code too, but why C# ??? :'( + having too much fun with the game as is)

You have a problem with C#? Square up! :sticktongue:

Seriously, I don't know why C# makes you sad. Static typing and explicit, {}-based scope seem like good things to me. If you don't like it, I think any other CRLs will suffice. They might be kinda dodgy in their interface with Unity, but they work.

As for why C#: I don't profess to know why Unity chose C# as its language of choice, but I know it's a solid C derivative. It has clean, consistent syntax, it handles arrays well, it's pretty fast... It's certainly not a bad language. What would be your preference?

Link to comment
Share on other sites

3 hours ago, 0111narwhalz said:

The issue is not activating loads of AGs on multiple vessels. Plenty of mods already use that functionality (well, a couple, but that's enough). The thing is, those mods only work (to that capacity) when the affected ships are loaded into physics.

I have mixed up a bit the "issues". As is, KSP is a steampunk/realist/surrealist MYTHIC game... (Keep it mythic ! Localization is just secondary... Kerbal doc will never be fully translated... Space is english, as is air... As is engineering... )

Gonna try to be more structurated.

1) the Tracking Station, XXIth century: If I want do fire manually the same AG say on 10 probes, now I have to go manually to each of them, waiting physics, fire the AG, next probe and so on.

- It could be done without full graphic rendering : a small screen in the TS could log "probe 1: Timing Time... Making Physics..." and so on: that needs no more realism than "the real game", as far as I'm concerned : it's juste a way to render macro, without rendering hi-res action. Some small cathodic green-on-black screens with "off/on"-"choose vessel" - "fly" -" recover" + selecting vessels by AG name, then fire this AG for the list... Means: do what I would have do, but without switching display...

2) Basic navigation tools : I think modders could easily make a "F16-Falcon"-like HUD on the cams, any cam outside, and any IVA... With indication about the heading and angle of a reticle mouse cursor Or at least numerical display of these + a way to have ONE block with all gauges, altimeter, navball fom any view... IVAs generally are quiet unusable... And no map & GPS in a plane is quiet funny... One could at least make a map by writing on the one existing

3) Languages: I would prefer some real OOP like Python or C++, the problem wih C# is that it's MS, that explains some difficulties... Java for example would be fine to be multiplatform mods, and can use any compiled lib... Gonna use kOS one of these days... Anything workable with Eclipse or other generic IDE would do...

4) Generalized AGs : I don't get why there is not a single "new" button for AGs... With a save window like for subassemblies...Takes one hour to make for a noob...

Suggestions, needs, dreams, call it all but complains :):):):) <-- some stock :)

 

Edited by Kenny Kerman
Link to comment
Share on other sites

On 10-3-2017 at 8:38 PM, regex said:

Do you think Squad could benefit from a public

R O A D M A P
O A D M A P R
A D M A P R O
D M A P R O A
M A P R O A D
A P R O A D M
P R O A D M A

Why, it's a space program. Where they're going they don't need roads!

Link to comment
Share on other sites

On ‎3‎/‎13‎/‎2017 at 6:39 AM, Mako said:

I'm not interested in a roadmap with specific dates or specific parts like engines and decouplers. I don't think the community needs that.

I'm interested in what Squad would like to work on before they call KSP complete. I would guess, and it's been hinted at, that there is an internal plan. I believe sharing that plan would do more good than harm.

With this Roadmap conversation, I wondered why Squad would publish a development plan for others to see. For modders, perhaps, so they can plan updates? Not likely; even highly popular mod-supporting games like XCOM 2 and Fallout don't have public roadmaps. The only thing I got from a cursory search was some predatory company selling garbage documents; nothing from Bethesda or Firaxis. Roadmaps come up for large business applications, like Office 365. But for a game?

Folks are taking KSP way too seriously. It's being ruined FOREVER! supposedly.

And if you get trapped in TV Tropes going to that link, that's OK. Take a break from the KSP community for a little while.

Link to comment
Share on other sites

+ there is nothing like "KSP complete"... Bigger computers mean bigger simulations... I think one could really manage 100 vessels if there was a way to alleviate micromanagement. "Smart Parts" enhanced + AG is clearly a good and simple way to make macros. And then programming for users. And then a way to choose the level of management (nano-->giga). That needs sensors everywhere and cool roundings and "actions" everywhere. A lot. And encapsulating AG in subbassemblies, with names permitting to merge them when you plug subassemblies...

And, ahum, standard lists -_-

(Also: why the hell are some batteries physicless? For the modder: what about realism? A basic timer weighting 50kg? Really?! A sofisticated MechJeb Calculator weigthing around 10 grams (or is it ONE ?) ...?!)

(No zero speed, altimeter is not on radar I think, not showable there) (2 boxes resized)(uberprecision: this thing moves less than a dead parrot)

juk5

 

Edited by Kenny Kerman
Link to comment
Share on other sites

@Kenny Kerman

I think you're going way out of scope here. KSP cannot handle the possible performance pain of macromanagement. On some (decent) systems, it struggles with more than a couple hundred parts. Now, each vessel gets its own thread, but the impact of unpacking potentially dozens of craft is still more or less the same. You seem to think the graphical rendering is what consumes most of the system resources when a craft is unpacked. Except in extreme cases, this is probably false. The physics are the problem. The thermal and the aerodynamic and the main physics threads probably use far more each because each part has its own mass, force, heat, and so on, which must be communicated to adjacent parts in the tree.

I do think you could get some use out of kOS. Although, if you're complaining about C#, I don't know if you could stomach Kerboscript. It's a weird language.

I'm not sure C# is any less OO than C++. I've created classes and objects thereof many a time. You seem to dislike C# purely because it is related to Microsoft. I think this is shortsighted. Could you provide some specifics as to why you dislike C# so thouroughly so I might rectify my perception?

As to AG creation: You only have so many keys on your keyboard. Sixteen of them are availabe for action grouping, though three have side effects. If you have more AGs than that, I'm very curious as to how you plan to activate them. I don't really see how AG creation is that hard. You select a key, select a part, and select an action. Seems perfectly intuitive to me. How would you streamline it?

Why are some batteries physicsless? Because of that nasty thing whereby parts are simulated individually. Physicsless parts offload their mass and forces onto their parent part, effectively joining them as one. It cuts down heavily on complexity for parts that have little mass and tend to be spammed in large quantities.

There is no standard among modmakers because there is no standard within stock. And the last attempt at rebalancing was met with wails.

Link to comment
Share on other sites

46 minutes ago, 0111narwhalz said:

@Kenny Kerman

I think you're going way out of scope here. KSP cannot handle the possible performance pain of macromanagement. On some (decent) systems, it struggles with more than a couple hundred parts. Now, each vessel gets its own thread, but the impact of unpacking potentially dozens of craft is still more or less the same. You seem to think the graphical rendering is what consumes most of the system resources when a craft is unpacked. Except in extreme cases, this is probably false. The physics are the problem. The thermal and the aerodynamic and the main physics threads probably use far more each because each part has its own mass, force, heat, and so on, which must be communicated to adjacent parts in the tree.

I do think you could get some use out of kOS. Although, if you're complaining about C#, I don't know if you could stomach Kerboscript. It's a weird language.

I'm not sure C# is any less OO than C++. I've created classes and objects thereof many a time. You seem to dislike C# purely because it is related to Microsoft. I think this is shortsighted. Could you provide some specifics as to why you dislike C# so thouroughly so I might rectify my perception?

As to AG creation: You only have so many keys on your keyboard. Sixteen of them are availabe for action grouping, though three have side effects. If you have more AGs than that, I'm very curious as to how you plan to activate them. I don't really see how AG creation is that hard. You select a key, select a part, and select an action. Seems perfectly intuitive to me. How would you streamline it?

Why are some batteries physicsless? Because of that nasty thing whereby parts are simulated individually. Physicsless parts offload their mass and forces onto their parent part, effectively joining them as one. It cuts down heavily on complexity for parts that have little mass and tend to be spammed in large quantities.

There is no standard among modmakers because there is no standard within stock. And the last attempt at rebalancing was met with wails.

So, it wouldn't change the time taken to make the actions... Nor in the game neither IRL... But it would be less repetitive micromanagement, what computers (and tracking stations control rooms...) are made for. And a cool Tracking Station... (Maybe with some real map?)

To make macros in KSP is easy for flight (nearly done), if AGs are namable and unlimited, can hide actions they will fire from the general list (stages are AG, in a mof) + realistics sensors and smartboxes. "You only have so many keys on your keyboard. Sixteen of them are availabe for action grouping" You don't get my point : "Smart Parts" add-on parts fire AG. I quiet never fire custom AG manually. I don't need keyboard binding for all of them. 10 is enough. But I could need 20 more for boxes.

C++ is beter than any MS business, because it's free and runs on any platform easily. I know really well Java, and it's obvious that some technical problems spoken here, or some crappy as hell designs (a f*** list has always to have a f*** set of sorting and filtering, creating deleting, editing, always the same behaviour... Speaking about lists inside KSP... Why the hell can't I sort ALL by dates? It's just a 1st year OOP student exercise...). Plus : all ressources (icons, images..) can be easily extracted, modified... Plus JavaDoc... Plus Eclipse *-*

So, the stage actions list should contain names... If a box is a macro containing a list of actions, everything can be a a box, from a command point of view. I'm not sure I'm cristalclear :) so the stage list become a AG tree...

Localization is built in Java. I dont say this game should be done in Java, put it's compilable an can interpret Java. No need of DLL or other cryptic standards... Sources. Ideal for mods + everything built in for deployement and updates... Free. But well, not troll about languages because the main reason is : I foresee a PitA installing MS on a Linux system... And of course, MS is a parasitic entity I don't want to feed.

Anyway, there is nearly nothing to code : a reticle, roundings, smart boxes (nearly finished), "standard" AG list... And the game gets new heigths...

"Why are some batteries physicsless? Because of that nasty thing whereby parts are simulated individually. Physicsless parts offload their mass and forces onto their parent part, effectively joining them as one. It cuts down heavily on complexity for parts that have little mass and tend to be spammed in large quantities."

Got it :) thx. (Please drop smileys everywhere in my texts, just discussing)

Edited by Kenny Kerman
Link to comment
Share on other sites

 

20 minutes ago, Kenny Kerman said:

I foresee a PitA installing MS on a Linux system... And of course, MS is a parasitic entity I don't want to feed.

You don't need any Microsoft software to do C#. Monodevelop is free, lightweight, and supports a number of languages. If Microsoft's involvement in the .NET framework is the main issue, then there's no real discussion to be had on the subject.

I think I get your point with the AGs. You want to unify the action groups and the stage stack, right? I think that could be done. I'm not sure how the stack works in the backend, but the actions are assigned within each part. You could probably assign something like "stage 6 = toggle engine" to the actions block, and each stage would ping its respective group. Then you could drag conventional actions to the stack.

Link to comment
Share on other sites

Ok I could reconsider... For now I think there is no need to for me to program as the time to begin is too big for what I need...

The "generalized AG system" could have a "fire another AG" in it... Through devices... A landing strut could fire an event when it sense weight, a tank when empty and so on....

Including some hi-level controller (and a centralizing tracking station... :P )... Look: I dont event fire manually ONE stage or action group: only clicking a timer. 24 probes deployed... Other boxes in the cargo, with 3 parts probes (pod, chute, decoupler, 0 box) ... Could be 50 parts, 4 boxes each... that's 7 AG used (if all the probes are the same model.........), and a BIG MESS in the stages. See: I can launch a miniprobe anywhere, then send this (modular, around 400 parts when mounted for my current tests) guy automatically reach it to fire another bunch automatically when in range... If I want to drive I play Mario Cart, if I want to fly a Spitfire or a Raptor I play a fly simulator. KSP is to be a interplanetary-level exploration-planning rocket engineer and Space Center's CEO :)

BTW, a multiprobe space station could even be precomputed: you launch in the evening, and see the movie the next morning...So you make the cool job, then start the repetitive part before going to sleep, and see the result with your coffee... Could even do a SimCity-like paper with photo's.... You open the Tracking Station, you have the paper + screens with the log and videos recorded during the automatized mission, and sometimes oups, info lost.

+ It would give a more realistic role to antennas : transmit images at low resolution/fps)... --> you can limit the volume of infos to be retained, and that becomes a real operationnal aim... Could have a "no-cam no-see" rule.

+ you can sell images to KTV :)

 

 

Also: why isn't everything renamable? Image: how do I find this light in this action group?

http://www.mirari.fr/BvG5

Showing more about AGs and smart boxes... In english -_-

 

I'll make a "Geograph" night launch test video :)

Edited by Kenny Kerman
Morning coffee genius
Link to comment
Share on other sites

On 3/14/2017 at 8:34 AM, Gordon Fecyk said:

But for a game?

I've seen plenty of indie developers who have public roadmaps. Astroneer, recently released as early access, has a Trello page, as does Subnautica, just to name two bigger, higher profile titles. Then I've seen devblogs and roadmaps for a few other indie titles hosted on developer websites. So it most certainly happens for games, and unlike your example of Microsoft's Office, it's usually by smaller developers that I would put on par with Squad.

Public roadmaps don't usually exist for traditionally developed games since the public only ever interacts with the game upon completion. With the recent trend of games being released and playable while under active, heavy development, developers begin including their customers in the development process - including roadmaps.

Edited by Mako
Added second paragraph
Link to comment
Share on other sites

(BTW, why does the navball display numerically the heading but not the elevation? Anti-verticalism? :P )

(what about a massively multiplayer/multi[planetary]sytems server with sync time and management of multimissions (pysics in the Cloud...)? something with that : https://spacedock.info/mod/302/KRASH - Kerbal Ramification Artifical Simulation Hub (simulation mod for KSP) ) and some http://www.oolite.org/

You could have VIP non-free status (like for the price of half the game/year), giving rigths to events, this model is (has been?) used with success, with less players by http://www.gokgs.com/ for example. They have something like 3-5% VIPs ("members of the club").

Edited by Kenny Kerman
Link to comment
Share on other sites

5 hours ago, Kenny Kerman said:

(BTW, why does the navball display numerically the heading but not the elevation? Anti-verticalism? :P )

Do you mean 'pitch'? For the same reason reputation is not displayed numerically, even though science and funds are. Some kids are always picked last... :wink:

Link to comment
Share on other sites

Found the GPS, but well, speeds are quantic (sometimes to nm/s, maybe the speed of individual air molecules? Or nanoseismicity?) :) By rounding it to 6 digits for the computing, turbulence would be more realistic... ("Pitch" if you want... "Elevation in the navy, and well, "inclination" in the air, it seems. Pitch refers to me to the angle of the plane regarding his movement)... I'll find a way...

This comes from a plane on the runway. Still have to understand what means the relative velocity to the (unmoving) target...

http://www.mirari.fr/hB6j

Edited by Kenny Kerman
Link to comment
Share on other sites

This thread is quite old. Please consider starting a new thread rather than reviving this one.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...