The Google Lunar X Prize’s Race to the Moon Is Over. Nobody Won • Hacker News Board

Re: The Google Lunar X Prize’s Race to the Moon Is Over. Nobody Won

#91

Post by Retric » Thu, Jan 25, 2018, 10:27 AM UTC

melling wrote: post #26

JumpCrisscross wrote: post #23

Earlier quoted context omitted.

> Should we just stay at home? The legacy of Scaled Composite's air launch approach, aerodynamic re-entry, and piloted design--all fair decisions for a quick daredevil run to space--cost Virgin Galactic and its investors hundreds of millions of dollars. Their delays, overruns and mistake, in turn, made life difficult for SpaceX, Blue Origin, Sierra, and a handful of other space start-ups. Customers, Congress and inve…

Look, if we're talking manned space flight, I was arguing against it 3 years ago in favor of robotic exploration. HN said I was wrong. As for X-Prizes, in general, I think they're a great idea. My blog is http://h4labs.org I took h4 from the first "X-Prize". H4 was the version of the watch that won the money. It has been explained repeatedly why they're a great idea. Should I rehash it on my blog, because I'm not goi…

Yea, you completely misunderstood my point.

Even just building and testing rocks has killed people because they are fucking dangerous. However, so is putting solar panels on people roofs. That does not mean nobody should build rockets, just that safty is a major consideration when building them.

IMO, astronauts are almost irrelevant from an overall safty standpoint. Rocketry has killed more people on the ground than it has people trying to get to space, though this may change over time as we get away from those extremely dangerous early days.

Ex: Just VLS-1 an unmanned rocket killed 21 peopel more than the total number of US astronauts killed so far.

Re: The Google Lunar X Prize’s Race to the Moon Is Over. Nobody Won

#92

Post by skgoa » Thu, Jan 25, 2018, 11:19 AM UTC

osense wrote: post #74

skgoa wrote: post #64

Earlier quoted context omitted.

This is typical for embedded software in cars, planes etc. Most of that C code will be generated though.

What would you be generating the C from? Are there any open-source tools that generate this kind of C?

Features will almost always be modelled in a tool like Matlab/Simulink, which has a C generator specifically aimed at automotive software. Similar tools that are used by EEngs, MechEngs etc. have the same functionality.

The support software underneath (equivalent to the OS kernel, OS libraries, frameworks etc. you would use to run desktop apps) is typically configured in one or more AUTOSAR[1] tools. The exchange format for AUTOSAR is .arxml, which is a horrible adaption of XML. These tools will also take .armxl files as input and generate the source files that contain the standard components, the drivers, the glue code for inter-component communication etc.

Only the software components that are not covered by the AUTOSAR standard or handed down by the car manufacturer as standard components will typically be written by hand, though even for those we will increasingly often write generators in e.g. Python. That's because different versions of the same car, different target plattforms (that is car models) etc. will often use different subsets of features, different internal networking, different messages etc. and since everything has to be set statically before compile time, we need to generate different versions.

TBH I doubt that much of that is open-source. There is a reference implementation of AUTOSAR and there are some code generators that are FOSS as part of the genIVI project. But practically all tools you would actually want to use are closed source and super expensive. That's one of the major reasons why the push for open-sourcing automotive software is not going to lead to anything.

[1] AUTOSAR is an international standard for automotive software architectures and components/APIs/etc. that is aimed at making software modules more or less plug'n'play between different SW vendors, car manufacturers etc. All of this is statically "linked" at generation time! The software will not even compile if things don't match exactly. The first version of a Linux-based Adaptive AUTOSAR that basically works like Android has been released recently, though.

Re: The Google Lunar X Prize’s Race to the Moon Is Over. Nobody Won

#93

Post by plafl » Fri, Jan 26, 2018, 5:32 PM UTC

osense wrote: post #56

plafl wrote: post #55

Earlier quoted context omitted.

I'm not sure for space, for aircraft it's usually C or Ada and lot's of testing. Sometimes Scada. I have not seen static analysis applied. You could use any language of your choosing but you then would need to demonstrate test coverage on the generated machine code, which you could avoid maybe by using a certified C compiler. I would expect space applications to have less stringent requirements, unless human lifes ar…

It seems unfortunate to have to run tests at all, all the more running them on the generated machine code. What about having to satisfy some real-time guarantees? Is this checked by tests as well?

I think, but not remember too well, that tests and some human analysis can suffice. But in the end there is testing, more tests and whole teams dedicated to testing.

Re: The Google Lunar X Prize’s Race to the Moon Is Over. Nobody Won

#94

Post by slededit » Sat, Jan 27, 2018, 9:25 PM UTC

jk4930 wrote: post #43

greenhouse_gas wrote: post #38

Earlier quoted context omitted.

Just curious (not a competitor :), what business model is there in placing micro-sats on the moon?

Considering that there's little time left, your payload (and business model) has severe constraints. Unless you have something pre-developed and nearly ready to launch (e.g. at a university), aim for something simple. That leaves you mostly with media attention (sponsors), technology demos (e.g. for third parties), paraphernalia/vanity items that wealthy individuals or communities pay for (no remains or religious art…

A constellation of cube sats orbiting the moon would provide a nice synthetic arepture radio antenna that would be shielded from earth noise. It's a lot more economical than a giant dish on the moon itself.