What doesn't KSP teach about Rocket Science

[ValleyTwo]

That computers control MOST space flights.  I mean people talk about manned and unmanned flight but most of it is computer controlled.  We should call them 'normal' or 'with a side of meat bags'.  Well, that is what Bender would say. 

[PB666]

7 hours ago, ValleyTwo said:

That computers control MOST space flights.  I mean people talk about manned and unmanned flight but most of it is computer controlled.  We should call them 'normal' or 'with a side of meat bags'.  Well, that is what Bender would say. 

As far as i can tell humans don't have built in gyroscopes or accelerometers or telemetry. Even if we did we would have to be a math savant and still only be able to interpret the flight data 100 times slower than a computer could. 

BTW most commercial pilots fly most of the time in autopilot modes. Particularly now you can program waypoints into the GPS system, its actually safer if all the craft are flying with the same equipment along the same GPS waypoints.(except when the pilots are surfing the internet and not responding to center). The newer Airbus are only exceptionally manual, its all flight computers unless you have flight system failure like AF447. 

I don't think ive ever heard of a computer controlled aircraft entering a deadmans spiral, its what they mean when they say for IFR, trust your instruments. 

[Camacha]

On 25/05/2016 at 10:57 AM, Creature said:

What we do in KSP is nothing like reality. Everything in real world is thousands of times harder, more tedious and takes a massive amount of work. I don't know much about aerospace engineering, but I bet that every single nut and bolt has been thoroughly tested, there has been years of materials studies behind it and there's probably a company somewhere that employs a dozen people who primarily manufacture these bolts. Of course everything looks nice and efficient once it's done, but there's a huge amount of work done behind the scenes. What KSP gives us is a nice dream, but it doesn't prepare anyone for working in engineering in any meaningful way.

It is supposed to be a game, not a job Everything in life that you take for granted turns out to be exponentially more involved when you start looking into it. Even the cheapest toy you have, made of smelly plastic, was produced using techniques that has been developed over many decades, centuries or even millennia, on intricate and complex machines designed by huge teams and set up by people that can talk for hours about just the temperature the plastic needs to be, or about how plastic flows in the mould, mould maintenance or any of the many other areas that need to be dialled in just right to make it all tick. Your very life depends on armies of these people and those are just the basic necessities. For a bit of luxury, you need a few armies more.

If you want a game that takes all that into account, there happens to be one out there that is perfect. It is called Real Life® and has amazing graphics, though the story line is poor at times.

3 hours ago, PB666 said:

BTW most commercial pilots fly most of the time in autopilot modes. Particularly now you can program waypoints into the GPS system, its actually safer if all the craft are flying with the same equipment along the same GPS waypoints.(except when the pilots are surfing the internet and not responding to center). The newer Airbus are only exceptionally manual, its all flight computers unless you have flight system failure like AF447. 

Pilots tend to fly planes manually regularly to keep their skills sharp, or merely because they enjoy it more than baby sitting a machine

 

Quote

I don't think ive ever heard of a computer controlled aircraft entering a deadmans spiral, its what they mean when they say for IFR, trust your instruments. 

 

There have been plenty of incidents and crashes caused by pilots trusting their instruments too much or because the computer responded in ways a pilot did and/or could not expect. A computer often executes without judgement and that can be as dangerous as a pilot making mistakes, or a computer can simply generate faulty output because the input (in the form of sensors) was polluted.

[PB666]

9 minutes ago, Camacha said:

There have been plenty of incidents and crashes caused by pilots trusting their instruments too much or because the computer responded in ways a pilot did and/or could not expect. A computer often executes without judgement and that can be as dangerous as a pilot making mistakes, or a computer can simply generate faulty output because the input (in the form of sensors) was polluted.

Mainly human error. Af447 was cause by the pilots failing to trust their instruments, but it wasn't the flight computers ultimately that failed, but the pitot tubes.

[Camacha]

18 minutes ago, PB666 said:

Mainly human error. Af447 was cause by the pilots failing to trust their instruments, but it wasn't the flight computers ultimately that failed, but the pitot tubes.

You cannot cherry pick what is part of the system and what is not Otherwise we can say the pilots did not fail, only their brains/hands/eyes did. Pilots rarely completely fail and neither do computers, due to them both being redundant. Humans mainly fail because of their inherent weaknesses (which is mostly menial tasks, communication and a few other things), computers fail because of their inherent flaws (executing without questioning, operating in unexpected ways, often due to the complexity of the different interoperating systems).
 

Using the strengths of both, you get a pretty reliable system, but depending solely on computers would likely cause as many issues as it would solve. Even the two main aircraft builders cannot seem to agree on whether a pilot should be able to override the computer (Boeing) or it is the computer that should protect the aircraft from the pilot (Airbus).

Edited May 28, 2016 by Camacha

[PB666]

When you go into a dead stall because you don't believe the IAS and altitude are correct, and they are, its pilot error. Read the AF447 wiki. For the most part most aircraft and space craft have redundant systems. On AF447 only when all three disagreed momemtarily was a manual mode forced, once the craft had lost altitude the pitot tube reopened up but the pilots did not nose down, and by the time they did it was too late. You can look at the majority of crashes over the last few year, it if was not intentional, it was often pilot error, there are exceptions like the Alaska Air incident, so I'm not cherry picking.

 

[Camacha]

4 minutes ago, PB666 said:

When you go into a dead stall because you don't believe the IAS and altitude are correct, and they are, its pilot error. Read the AF447 wiki. For the most part most aircraft and space craft have redundant systems. On AF447 only when all three disagreed momemtarily was a manual mode forced, once the craft had lost altitude the pitot tube reopened up but the pilots did not nose down, and by the time they did it was too late. You can look at the majority of crashes over the last few year, it if was not intentional, it was often pilot error, there are exceptions like the Alaska Air incident, so I'm not cherry picking.

 

See my edit. If the major manufacturers are not agreeing on the matter, I am certainly not going to argue about it. Computers prevent errors, but introduce them as well. The exact mechanism through which that happens is not relevant.

Edited May 28, 2016 by Camacha

[kBob]

I'd say it doesn't teach about the time of space travel.  That is it takes a long time to get anywhere.  In KSP you can't even travel to the nearest star, but if you could you'd just time warp there.  Sure people could do everything without time warp (and some do) but for most, time warp gives no real sense of the real time involved (wait 6 years for a transfer no thanks I'll warp).  Kind of like watching Star Trek, all the waiting and time (even at warp 9.9 to get places is often glossed over...but without that it would be an RTS and for most, including me, it would get tedious fast so we end up skipping through our space problems with only a token nod to the number of years that have passed by. 

Also for basically the same reasons it doesn't teach about the true vastness of space (and even if more solar systems were reachable it probably still wouldn't ), and the unlikelihood of running into friends, foes or much of anything while traveling out there (and certainly not Amelia Earhart ).

And time and distance are part of the science of rockets I should hope.

Edited May 28, 2016 by kBob
minor clarification

[Creature]

2 hours ago, Camacha said:

It is supposed to be a game, not a job Everything in life that you take for granted turns out to be exponentially more involved when you start looking into it. Even the cheapest toy you have, made of smelly plastic, was produced using techniques that has been developed over many decades, centuries or even millennia, on intricate and complex machines designed by huge teams and set up by people that can talk for hours about just the temperature the plastic needs to be, or about how plastic flows in the mould, mould maintenance or any of the many other areas that need to be dialled in just right to make it all tick. Your very life depends on armies of these people and those are just the basic necessities. For a bit of luxury, you need a few armies more.

If you want a game that takes all that into account, there happens to be one out there that is perfect. It is called Real Life® and has amazing graphics, though the story line is poor at times.

 

Of course and for the most part it's far better off being simplified. There's no need for hyperbole, I certainly wouldn't want a game about testing materials for nose cones or anything like that and nobody would. But as for the topic at hand, it is one of the big things that KSP doesn't teach on any level at all.

I believe it could (and should) take into consideration at least some aspects of engineering that are now completely ignored without losing the fun, but that's obviously just a personal preference. Seeing as how insanely complicated rocket science and space flight in real life is and how incredibly simple it is in KSP, I'd dare to say that the spectrum is so wide the position we're at now could be nudged a bit towards complicated.

[Camacha]

26 minutes ago, Creature said:

I believe it could (and should) take into consideration at least some aspects of engineering that are now completely ignored without losing the fun, but that's obviously just a personal preference. Seeing as how insanely complicated rocket science and space flight in real life is and how incredibly simple it is in KSP, I'd dare to say that the spectrum is so wide the position we're at now could be nudged a bit towards complicated.

As an avid fan of technology and someone getting into the nitty-gritty of material properties and force calculations, I would not be against a little more detail, but how would you imagine that? Mods seem to have filled most gaps pretty well - different types of fuel and engines, for example. What else would you consider a valuable lesson, without making it overly complicated?

[RenegadeRad]

Soundtracks don't play while you're in space

[Creature]

14 minutes ago, Camacha said:

As an avid fan of technology and someone getting into the nitty-gritty of material properties and force calculations, I would not be against a little more detail, but how would you imagine that? Mods seem to have filled most gaps pretty well - different types of fuel and engines, for example. What else would you consider a valuable lesson, without making it overly complicated?

Like you said, mods cover a lot. Apart from what the best known mods already do, something I personally would like to see would be more emphasis on electrics since typically a solar panel or two and a couple of weightless batteries seem to be enough to power almost anything. I don't know, maybe it's a trivial thing in real life, but somehow it just seems far too easy. I never ever pay attention to power drain and generation unless it's a mining operation. Different materials could have different environmental tolerances, in BTSM for example had this fun experience of trying to land a probe on Eve, but I hadn't realised that the core couldn't handle the pressure and it exploded before reaching the surface. This actually made the barometer a useful instrument. I'm playing with Kerbalism right now and it's in a pretty sweet spot where there are more things I need to consider but it still feels like a fun game and not like work.

Also the way science is done could be more involved, after all that's one of the primary motivations for spaceflight. I wouldn't mind something fairly easy to show how the instruments and scientific payloads come to being. You could for example have general purpose instruments and then after you've done some studies on a planet or biome you could design experiments for that particular phenomenon. But these are just things that pop into my mind right now. 

Overall though I'd say that with realism overhaul installed the game is already complicated enough for my tastes and sometimes a bit too complicated which is why I'm playing without it. It's a difficult thing to balance, bringing in too much detail is most likely far worse than not having enough detail. I bet if we had RO-levels of complexity in stock, the game would be much, much less popular.

[Camacha]

1 minute ago, Creature said:

Like you said, mods cover a lot. Apart from what the best known mods already do, something I personally would like to see would be more emphasis on electrics since typically a solar panel or two and a couple of weightless batteries seem to be enough to power almost anything. I don't know, maybe it's a trivial thing in real life, but somehow it just seems far too easy. I never ever pay attention to power drain and generation unless it's a mining operation. Different materials could have different environmental tolerances, in BTSM for example had this fun experience of trying to land a probe on Eve, but I hadn't realised that the core couldn't handle the pressure and it exploded before reaching the surface. This actually made the barometer a useful instrument. I'm playing with Kerbalism right now and it's in a pretty sweet spot where there are more things I need to consider but it still feels like a fun game and not like work.

Yeah, that is one thing I have though about in the past: some sort of need to wire your components up. No more trivially sticking things together, you actually need to consider internal paths. That would certainly impact the design of things. I do not know whether anyone is familiar with the music program Reason, but there you can wire up things in a pretty involved fashion. Something similar would actually be pretty sweet, even though it would still not equate to an engineering simulation in any way.

[PB666]

2 hours ago, Creature said:

Like you said, mods cover a lot. Apart from what the best known mods already do, something I personally would like to see would be more emphasis on electrics since typically a solar panel or two and a couple of weightless batteries seem to be enough to power almost anything. I don't know, maybe it's a trivial thing in real life, but somehow it just seems far too easy. I never ever pay attention to power drain and generation unless it's a mining operation. Different materials could have different environmental tolerances, in BTSM for example had this fun experience of trying to land a probe on Eve, but I hadn't realised that the core couldn't handle the pressure and it exploded before reaching the surface. This actually made the barometer a useful instrument. I'm playing with Kerbalism right now and it's in a pretty sweet spot where there are more things I need to consider but it still feels like a fun game and not like work.

Also the way science is done could be more involved, after all that's one of the primary motivations for spaceflight. I wouldn't mind something fairly easy to show how the instruments and scientific payloads come to being. You could for example have general purpose instruments and then after you've done some studies on a planet or biome you could design experiments for that particular phenomenon. But these are just things that pop into my mind right now. 

Overall though I'd say that with realism overhaul installed the game is already complicated enough for my tastes and sometimes a bit too complicated which is why I'm playing without it. It's a difficult thing to balance, bringing in too much detail is most likely far worse than not having enough detail. I bet if we had RO-levels of complexity in stock, the game would be much, much less popular.

More of the more of should go into the game, some sort of flight engineer, more parts. Off-world launch pads, the   silly part lockers that allow you to carry things.