News for hackers. Stuff that matters.

A comment on How do you compute the midpoint of an interval? (2014) [pdf]. Parent comment

Re:How do you compute the midpoint of an interval? (2014) [pdf] by adwn on Friday March 29, 2019 @05:39AM (#19520240)
> A rocket could blow up because the library function you wrote didn't account for floating point arithmetic correctly.

I would go so far as to say that using floating point in any software that might cause the rocket to blow up is probably Doing It Wrong (TM). That does not include the video encoder sending a livestream to the ground station, but it does include the µC that controls the engine turbopumps.

[ Parent | on HN ]
Re:How do you compute the midpoint of an interval? (2014) [pdf] by lhjuhafjbh on Friday March 29, 2019 @06:02AM (#19520315)
A microcontroller for a pump probably doesn't need floating point numbers, but flight control software uses it.

This went wrong before and an Ariane 5 was blown up (automatically as a safety measure), see https://en.wikipedia.org/wiki/Cluster_(spacecraft)#Launch_fa... Strictly speaking the problem was converting floats to 16-bit integers which ended in an overflow... So avoiding floats doesn't even help as it won't protect you from overflows.

[ Parent | on HN ]