Hacker News Reader

Comment by torpfactory | Hacker News Reader

The really crazy part is that they never did a full up integrated test of all those systems together and in sequence before arriving at Mars.

View on HN

I was at a local radio telescope for the “live” event when it landed. Another crazy part was due to the time delay, by the time they were receiving info about entering the Mars atmosphere, it was was already over in reality and we all just had to wait those “minutes of hell” to find out what had happened.

There's no such thing as "in the same time" at distances where light propagation delays are comparable to the duration of the event.

For perceiving the events happening, sure, that was my point. These time differences are nearly always imperceptible but at distances of planets it truly becomes a way of looking into the past.

But to be clear, events happen concurrently, eg NASA knew _when_ relative to earth time and _perceived_ events, each stage was "happening" local to Mars vs Earth (assuming it all went well).

They made videos about this "7 minutes of terror", I forgot the name of them but linking here [0]

[0] https://youtu.be/Ki_Af_o9Q9s

Earth and Mars are practically stationary, and Occam's razor would say to assume that light travels at the same speed in all directions, even if we can't prove it. So the time of events on Mars has an "obvious", if technically not unique, solution.

That light travels at the same speed in all directions has been experimentally verified extremely rigorously.

Relevant video: https://www.youtube.com/watch?v=pTn6Ewhb27k

As motherboard designers and HFT firms will tell us.

Those guys have supernatural powers.

Each time I by pass integration tests it always fails at the integration points :-)

Agreed, the amount of validation work that goes into these projects must be amazing, and even then the engineers have to essentially cross their fingers that no issues were missed. I think I’m spoiled in my job in that our software can at least be rolled back if we detect an issue after the fact (also much lower stakes).

Keep in mind that a lot of investments are made in ensuring that these things don't fail. Not saying its not a great achievement. But it helps to remember that for most software shops, reliability improvements are often treated as 2nd class compared with new feature development, which reflect the reality that most software shops don't really need quite that perfect amount of reliability.

This is a good point, which I think was reflected by in nickpeterson's comment as well. There's still a bit of a culture of "move fast and break things" in my line of software engineering that I don't think ever existed in space travel. (or at least hope not)

Absolutely spoiled. My team does a lot of testing in production. We can find real world faults quickly and easily, roll back, and save ourselves time and effort by failing in production intentionally. It’s about as easy as it gets, haha.

Tesla and Boeing appear to do the same...

I have a theory that difference isn’t just quality level of the engineers, but the shared interest and stakes. Most things we attribute to engineering incompetence are really just carelessness because the stakes don’t appear high.

It's probably comparing apples to oranges. There are very talented software engineers and very talented non-software engineers. Regardless of the field, I think it's worth mentioning how extraordinary it is to take on an engineering feat that comes with so much risk. And while this isn't the case with the Curiosity rover, in my head I suppose I was thinking more of missions where human lives are at stake. I'm just a bit in awe at the courage and confidence someone must have to write code or design something and be comfortable that it will keep another human safe.

Reply on news.ycombinator.com