Future of KSP, still a game it used to be?

[rkman]

1 minute ago, regex said:

There's always tried and true do-it-by-hand math, and KSP gives you the information for that in the VAB/SPH, and it's not at all hard to do. 

The market for a game that requires doing actual rocket science is very small. 

It's hard enough that most people don't even try. And you know that, all you're doing here is bragging about you math skills. 

 

[Guest]

Just now, rkman said:

And you know that, all you're doing here is bragging about you math skills.

Bragging about the fact that I know how to use a calculator and can read and interpret a simple equation?  Oooooooookaaaaaaaay...

[MiniMatt]

25 minutes ago, regex said:

...use a calculator and can read and interpret a simple equation?

But a PC is just a calculator with prettier graphics  Why work it out on a separate pocket computer when you can work it out on the same computer that's launching little green kerbonauts to a fiery doom?

Don't get me wrong, there is great value in knowing and understanding Tsiolkovsky - knowledge of it informs construction, resulting in a more rapidly designed, more efficient craft. Knowledge of it also allows you to realise when the computer has got it wrong - if you've a rough idea in your head of the correct figure, the computer giving a wildly different result lets you know something's squiffy: perhaps a fuel line is incorrectly plumbed.

Were I to work out each stage on a calculator I'd soon tire and resort to saving the equation as a stored function - and at that stage, I might as well have the function stored in KSP itself.

I'd bet teaching contacts in the KSP Education edition would be great collaborators to come up with a method to teach the "how & why" of the equation within KSP. Present step by step instruction/experiment/result to build and demonstrate the formula from the ground up, and allow students to automate calculations as they grasp them and repeated practice becomes tedious rather than instructional.

I'm not particularly forceful for stock dv calculators while the option through mods is available, but I do think KSP has an opportunity to do something better - something other than "here's the answer" that mods give you, or the "go read the equation on wikipedia" option left to those who don't know it by heart.

 

[Guest]

1 minute ago, MiniMatt said:

But a PC is just a calculator with prettier graphics  Why work it out on a separate pocket computer when you can work it out on the same computer that's launching little green kerbonauts to a fiery doom?

I have several problems with stock delta-V calculators because they will never be as good as mod software.  This is not because Squad is comprised of terrible programmers or can't handle the challenge but because they have limited time to spend on a delta-V calculator.  ENG and MJ use some pretty complex, optimized code that took a long time to get right.  It also takes work to maintain that code.  Squad can't simply lift that for their own use due to licensing restrictions, and writing and maintaining new code will take a lot work.  Furthermore, Squad's paradigm for KSP has always been to starve the player for information so people will pretty much always end up installing ENG or MJ in the end because they will always provide more features.

Much as I think a delta-V calculator would be good for the game, I don't think people realize what they'll end up getting as a stock feature.  And a stock feature has to be pretty much perfect on the first round, as opposed to mod software that can constantly iterate and address customer concerns and innovative ways of breaking the software.

Which is why I suggest using the tools we already have or installing a mod to handle it.

[MiniMatt]

3 minutes ago, regex said:

I have several problems with stock delta-V calculators because they will never be as good as mod software.

Replace "good" with "extensive" and I definitely agree.

4 minutes ago, regex said:

Which is why I suggest using the tools we already have or installing a mod to handle it.

Yep, broadly agree. Whilst I don't think adding a stock calc would take anything away from these options the status quo is acceptable. I just think there are possibilities to do more.

7 minutes ago, regex said:

...they have limited time to spend on a delta-V calculator.  ENG and MJ use some pretty complex, optimized code that took a long time to get right.  It also takes work to maintain that code.... ...writing and maintaining new code will take a lot work....

Here I think you're on shakier ground. Firstly with regard to the focus and man-hours a paid 9-5 team can put in versus part-time unpaid hobby work honed and tweaked over years. But more importantly it kinda torpedoes the "you can always simply work this out on a calculator" argument - you can't simultaneously say that the player can simply use a calculator to interpret a simple equation and say these are complex optimised calculations beyond the wit of a professional software development house.

Now I do get your point that whilst the underlying equations are relatively simple, the optimisation needed and bug potential comes from the horrendously plumbed multi-onion skinned asparagus Eve mission designs. But whilst you could calculate those by hand I think it's fair to say most people would rather throw that problem at computer function, whether mod or stock.

Interesting thought experiment there: is a dv mod effectively compulsory for a crewed Kerbin-Eve return design?

 

But hey, we agree on two out of three points. Which, I think, for stock/mod dv calc discussion might be an internet record

[Guest]

2 minutes ago, MiniMatt said:

Here I think you're on shakier ground. Firstly with regard to the focus and man-hours a paid 9-5 team can put in versus part-time unpaid hobby work honed and tweaked over years.

Squad doesn't have the luxury of devoting all their 9-5 time to getting a delta-V calculator correct, nor do they release often enough to address all the issues that will come up with a delta-V calculator in KSP.

2 minutes ago, MiniMatt said:

you can't simultaneously say that the player can simply use a calculator to interpret a simple equation

Yes I can.  The human mind is much better at interpreting human intention and creativity than a computer will be.  Much of the problem with writing a delta-V calculator in KSP is dealing with human intention and creativity.

2 minutes ago, MiniMatt said:

complex optimised calculations beyond the wit of a professional software development house.

Never said that.  What I did say was that mod software can do it better because the release schedule is much more free and can rapidly address issues, as opposed to KSP's development cycle which may let issues sit for months while they are addressed for the next version.  When you're talking about a delta-V calculator that works for a game with part interactions and combinations as complex as KSP's you need to account for all sorts of crazy contraptions that players will come up with.  ENG and MJ have to constantly evolve to the playerbase's expectations and methods of building; Squad doesn't release often enough to do that.

2 minutes ago, MiniMatt said:

the optimisation needed and bug potential comes from the horrendously plumbed multi-onion skinned asparagus Eve mission designs.

That's precisely the issue delta-V calculators have to deal with in KSP, human intention and creativity, and mod authors will do it better than Squad mainly because of reaction time.

[MiniMatt]

And here I think we've reached the "agree to disagree" stage

15 minutes ago, regex said:

The human mind is much better at interpreting human intention and creativity than a computer will be.

That's actually a damn good point though, and one I hadn't considered. Chapeau.

On the wider points re Squad's ability to meet the feature set and reaction time of mod authors I do accept your argument but don't give it the same weight. I see it as largely similar to the aero/heat models pre v1 (or was it 0.9x? can't remember) - FAR/Deadly were significantly more extensive, objectively and subjectively better than the stock cold soup model, highly complex, accounting for infinite variety in craft design, and rapidly tweaked to suit player needs. FAR (and possibly Deadly, no recent experience) remains far more extensive, and to many is still subjectively better, but Squad were quite capable of coming up with a much better aero model than their initial placeholder.

If Squad deploy a stock DV mod I will almost certainly continue to use Engineer or similar because those mods will undoubtedly be more extensive, more suited to my needs and react quicker to playerbase whims. But I think Squad have an opportunity to do better than their current placeholder, and I think they can do it. Because they have the opportunity to do something mods don't (currently) do - not just provide a basic model that's surpassed by Engineer or MJ, but demonstrate how and why those equations work. Not just provide a cut down Engineer, but bring to the underlying maths the accessibility that hallmarks KSP's approach to basic orbital mechanics.

[CobraA1]

Patched conics: I don't notice anything terribly bad with them right now. My biggest beef is really with the maneuver node editing. Way too fiddly, and far to easy to lose track of it.

Rocket stability / part attachment node strength: It's actually very much improved since the early days of KSP, and IMO is at a good place currently. Part of the fun of KSP is overcoming design challenges with your spacecraft. Having to use struts is part of the price you pay for designing large rockets.

Mission and timewarp management: I'd agree with this.

Planetary exploration: Agreed. That said - they have made some progress in that area. There are missions to build bases, and you can get resources now. But it can still be improved.

Contracts and rewards: I'm actually pretty dissatisfied with this system right now. I'd like more missions focused on science - either by rewarding me with a lot more science, or by taking somewhere where I haven't done much science yet.

Overall, though, I think KSP's direction is pretty positive. I've liked all of the features they've added so far.

[Sirad]

Well as for my 5 Cents:

 

It looks like the Game is going the way it goes the last months. More Parts to play around at Kerbin and Kerbins SOI (Plane stuff and so on) Less or absolutely no debugging of stuff more far off from Kerbin (like Pol, who will never get fixed until the game gets dropped and several other bugged stuff)

So Kerbal Space Program comes back to the roots of KSP itself. Kerbin and Kerbins SOI. This is nothing bad, its just a economic 'Grab it' decision to sell the game. All else, we have to rely on the Modding Community . Im quite

satisfied with the actual Modding community and would even play KSP without any original content so far. As long as it works.

Well, all else will show the future, we have to look forward to 1.1

[rkman]

17 hours ago, regex said:

 The human mind is much better at interpreting human intention and creativity than a computer will be.  Much of the problem with writing a delta-V calculator in KSP is dealing with human intention and creativity.

I can see how the user's intention affects required delta-v, but how is delta-v available on a vessel dependent on the user's intention? There can be staging configurations that can confuse the calculator, but it would still be very useful in a majority of cases. Even in cases where it is not accurate enough it'll likely be in the right ballpark and thus would reduce the number of trials that the user needs to do. 

[I_Killed_Jeb]

18 hours ago, regex said:

I have several problems with stock delta-V calculators because they will never be as good as mod software.  This is not because Squad is comprised of terrible programmers or can't handle the challenge but because they have limited time to spend on a delta-V calculator.  ENG and MJ use some pretty complex, optimized code that took a long time to get right.  It also takes work to maintain that code.  Squad can't simply lift that for their own use due to licensing restrictions, and writing and maintaining new code will take a lot work.  Furthermore, Squad's paradigm for KSP has always been to starve the player for information so people will pretty much always end up installing ENG or MJ in the end because they will always provide more features.

Much as I think a delta-V calculator would be good for the game, I don't think people realize what they'll end up getting as a stock feature.  And a stock feature has to be pretty much perfect on the first round, as opposed to mod software that can constantly iterate and address customer concerns and innovative ways of breaking the software.

Which is why I suggest using the tools we already have or installing a mod to handle it.

I agree with all of your points. However, it took KER a LONG time to get updated to 1.0.5 and I think most advanced players consider dV information to be pretty critical. Yes we can calculate it manually (it's just arithmetic folks, it's really not any more complicated than putting two numbers into a single equation) but it takes up valuable time. A stock version is important so that it updates in a timely manner since you can't fairly expect modders, who are volunteering their time freely, to be 100% on the ball.

[Guest]

3 hours ago, rkman said:

Even in cases where it is not accurate enough it'll likely be in the right ballpark and thus would reduce the number of trials that the user needs to do. 

Free software can afford to think that way, paid software cannot.

[troyfawkes]

I thought this post was going to be a negative rant, but I found it really constructive and thought provoking

I agree on all points, and while I understand the "no one can say which mods should be stock" idea being put forth, I think we could all agree on some basics E.g., KAC, KJR and some implementation of an editor/in flight dV display.

I'd love to see more love put into rocket parts, specifically engine variations, and a complete rethink around planets. Like if they dedicated a large update exclusively targeting things to do on planets/moons, I'd be stoked. Like adding on a mission system rather than just contracts, e.g. land on Duna, set up a base with X criteria, and then follow the progression of 12 contracts around testing parts, bringing scientists or experiments to locations, and returning within a certain window (or something like that). Or really just spend a good chunk of timing thinking of what could be done to make me want to spend a year exploring Duna rather than just launching another random probe or what-have-you.

[Nightshift83]

Honestly, I hope they don't change anything apart from what is allready known to come. For me KSP is very similar to 'Mount & Blade' wich I played (as for a lot of other games too) like 20 minutes before I started to look for mods. It's the perfect framework to play it just as I like it to play (except for my low-end system restrictions). For now I play mostly vanilla with mods mostly for aesthetic and/ or performance issues, but I'm very glad that I could change the whole game in a couple of minutes thanks to all the mods avaiable.. (and have done so in the past)

 

Edited March 7, 2016 by Nightshift83
*cough* multiplayer