Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

NASA can't figure out what's causing computer issues on the Hubble telescope

npr.org

Post reply on HN
Print view View original

81–90 of 333 posts

Re: NASA can't figure out what's causing computer issues on the Hubble telescope

#81

Post by setug »

beerandt wrote: post #67
setug wrote: post #61

Earlier quoted context omitted.

It's a 30/40-yr old code and hardware. Now that I think about it, the problem could be related to some "exotic" decisions made by that time...

Lots of 30/40 year old hardware is still classified, especially spy technology.

Then hiding it when it's on Earth could make sense, actually attacking Hubble is much more complicated that anything on earth, given that you can actually put your hands on it, connecting through a JTAG and understanding what's wrong (besides spying).

Re: NASA can't figure out what's causing computer issues on the Hubble telescope

#82

Post by throw_away »

tomrod wrote: post #11
Koshkin wrote: post #6

They should have moved to the cloud. /ducks

This is a truly stellar wit. But in all honesty, how would an internet across the solar system work?

https://en.wikipedia.org/wiki/Interplanetary_Internet

Re: NASA can't figure out what's causing computer issues on the Hubble telescope

#84

Post by voldacar »

etskinner wrote: post #51
fouric wrote: post #17

I know that, at different times, NASA has used Forth[1] and Lisp[2] in some of their space applications. Both of these languages offer REPLs that generally accelerate the debugging process, and while your "average" Lisp might be unsuitable for hard real-time applications (due to the presence of a garbage collector, usually without the hard real-time constraints that you can get out of garbage collectors with extreme…

> Wait, then why don't they just switch over each component in turn? The "divide and conquer" debugging strategy. My guess would be that they want to try that method only if this debugging doesn't work. Imagine that there's an electrical issue in item 1 that fries item 2. If you switch over to item 2b, then you fry item 2b too! This is exactly what happened with the Soviet Salyut 7 station. They tripped an over-curre…

Wow, that's an amazing story. Thanks for posting, i had no idea something like that ever took place

Re: NASA can't figure out what's causing computer issues on the Hubble telescope

#87

Post by slver »

sharkweek wrote: post #46
mcc1ane wrote: post #43

Earlier quoted context omitted.

everything's in space

“It’s all in your head, but so is everything.”

Technically the universe is projected into our brain and we only perceive that projection. The problem is that it's a very shitty projection.

Re: NASA can't figure out what's causing computer issues on the Hubble telescope

#88

Post by AnIdiotOnTheNet »

beerandt wrote: post #38
fouric wrote: post #17

I know that, at different times, NASA has used Forth[1] and Lisp[2] in some of their space applications. Both of these languages offer REPLs that generally accelerate the debugging process, and while your "average" Lisp might be unsuitable for hard real-time applications (due to the presence of a garbage collector, usually without the hard real-time constraints that you can get out of garbage collectors with extreme…

>"The rule of thumb is when something is working you don't change it," Hertz said. "We'd like to change as few things as possible when we bring Hubble back into service."

A lesson not taught to any modern software developer. Instead they change things all the time for no real reason other than that they want to change things.

Re: NASA can't figure out what's causing computer issues on the Hubble telescope

#89

Post by tomrod »

JumpCrisscross wrote: post #76
tomrod wrote: post #11

Earlier quoted context omitted.

This is a truly stellar wit. But in all honesty, how would an internet across the solar system work?

> how would an internet across the solar system work? Realistically, each ~150 ms sphere would have its own cloud infrastructure. Those systems would then bridge with one another. So idk AWS on Earth and DogeNet on Mars. I would love for a distributed model as much as the next guy, but it's unlikely to happen for the same reasons that it isn't happening today.

But if each node communicated within its 150ms sphere, and you place nodes 10ms (3,000 km?), could this serve as a mesh network?

Re: NASA can't figure out what's causing computer issues on the Hubble telescope

#90

Post by bshep »

qzw wrote: post #44
Koshkin wrote: post #14

Earlier quoted context omitted.

Debugging on a computer that is down is not a very exciting process.

In the movie Hero [0], two kung fu masters fight a battle purely in their minds. And when the mental fight was over, they only execute one physical move to finish the battle. Think of this as the computer equivalent of that scene. [0] https://www.imdb.com/title/tt0299977/?ref_=fn_al_tt_1

In case anyone wants to watch a clip of the fight:

https://youtu.be/AeeoEpmyb2Y

Post reply on HN