Engineering zero margin for error

[hanelyp]

http://www.rhjunior.com/QQSR/QQSR0000.html#Comic=123

So in your professional opinion as a shuttle engineer, how much margin for error do the shuttle systems have?

None. None whatsoever.

Waiting for whoever signed off on this zero margin design to be charged in this court. For those who don't want to follow back, the episode started with a stowaway on a shuttle of this design.

[Jccarlton]

http://www.rhjunior.com/QQSR/QQSR0000.html#Comic=123

So in your professional opinion as a shuttle engineer, how much margin for error do the shuttle systems have?

None. None whatsoever.

Waiting for whoever signed off on this zero margin design to be charged in this court. For those who don't want to follow back, the episode started with a stowaway on a shuttle of this design.

That would be incredibly stupid, if true. You always design for redundancies and fail safes if you can, even in high reliablility vehicles. On the other hand, this was an emergency situation where you may have to rely on those redundancies and use your margins. In this case it seems that there were no magins to be used. That's fine for nonmanrated spacecraft and non critical things, but if your are in space, more than likely, it's critical.

[Schneibster]

I have no intention to exposing myself to those scripts, sorry.

As far as the Space Shuttle the reason it was so expensive was that it had triple redundancy in every system possible. The most notable was its computers; this is well known, but I'll be happy to provide references if you happen not to have heard of it.

[93143]

You guys have read "The Cold Equations", right?

No real-life system is ever engineered with zero margin. Ever.

Not deliberately, anyway...

[Schneibster]

You guys have read "The Cold Equations", right?

No real-life system is ever engineered with zero margin. Ever.

Not deliberately, anyway...

I never read that, but I was taught very early by a teacher I really trust that 2:1 overdesign is still risky; we don't know everything, nor can we reasonably expect to find out before we build it. Really important items are overengineered as much as 5:1.

ETA: Actually even that's an oversimplification. Try this:

Fast.
Good.
Cheap.
Choose two.