Comment by devindotcom | Hacker News Reader

Comment by devindotcom | Hacker News Reader

Parent Full thread

Boy, where to start. First of all the conditions around various planets and bodies are very different. Jupiter for example is oozing with ugly radiation so we can't just put a "stock" probe into orbit around Europa, it would fry.

Second, these missions aren't actually that frequent. It takes months or years for a spacecraft to arrive at its destination, and its design had to be finalized and tested rigorously years before its launch. That means by the time we even get a close up picture of Ceres, for example, the science of spacecraft building will have advanced by several years and we now can include new sensors and improve fuel efficiency and so on.

Third, what do you think science is? Taking photos and readings. That is a huge, huge part of basic science, especially when it comes to astrophysics, astrogeology, astrochemistry, and so on. Landing on a planet provides further opportunities to take photos and readings, but it's phenomenally difficult to do.

Finally, if you are bored by cutting-edge space travel and the study of distant worlds that may hold the possibility of life... I don't even know. I'm mystified by your perspective on the universe we live in.

To amplify your comment, here's a web page with documents from the science definition team -- the panel that decides what exactly the mission will study:

https://solarsystem.nasa.gov/europa/sdt2013.cfm

A key document is the very granular "Science Traceability Matrix" which is very focused on "can instruments be made to satisfy these science goals" (zoom in to notice a lot of requirements flow-down from science goals):

https://solarsystem.nasa.gov/europa/docs/Clipper_Trace_Matri...

And regarding Europa, one particular item of concern has been, can markers of life be detected in a plume as sampled by an orbiter. Here's a recent workshop on the subject:

https://astrobiology.nasa.gov/calendar/europa-plume-workshop...

It's a colossal understatement to say that a standard design plus tailoring isn't really going to lift the burden of figuring out if detecting life in the plume is possible.

But that said, there are re-usable components. Some of the spectral radiometers ("take photos") are largely re-used, also communications devices, etc.

Apart from being extremely interesting from space perspective, this matrix seems like a great instrument to help manage a project of this complexity and uncertainty. I wonder what other things software industry can learn from NASA about this stuff.

Well, NASA did run a software engineering laboratory for 25 years https://www.cs.umd.edu/~basili/publications/proceedings/P94....

> I wonder what other things software industry can learn from NASA

they take their software seriously. This is just one shallow article describing their process: https://rhodecode.com/blog/code-review-learn-nasa-codes/

But i recall reading that they do simulation testing of their rocket firing software - by exhaustively running through all possible input ranges from all their sensors, and add in invalid ranges to boot, and their software _musts_ pass it. Takes days to run i heard.

But what's stopping us from making exact copies of this probe/sat (say... 10) and launching them all at once?

Or at an interval of say 6 months a piece? That way we will have low cost since the r&d will be basically 0 for all the extra models and we get 10x the data and larger time span of coverage. Plus we won't be putting all our eggs in the same basket.

The day (or month, or year) you want to send a probe to Pluto and the day you want to send one to Mercury, not to mention the direction, are going to be totally different, and aren't easy to schedule — you might have a 3-day window to be sure something can get to Saturn, since you're relying on a gravity assist from Mars and a configuration of the planets that only happens every couple decades. Also, even if such a thing could be organized, there's simply no way the budget will support 10 missions launching at once. They have to be tracked, developed, and funded separately for a lot of really good reasons. Stuff that's shared (launch vehicles, software) can be co-developed but each mission has wildly different timelines and requirements.

Are launch costs still high?

It will be a long while before launch costs are brought down. They aren't all that different than what they were a few decades ago.

Because of budget.

The closest thing to what you are describing that NASA has done recently was sending 2 rovers to Mars at the same time (Spirit and Opportunity). It cost nearly $1B dollars in 2003 money to do it [1].

Nasa's entire budget for 2015 is $17.5B

In contrast, NASA only spent a further ~$130M keeping the rovers going for the last decade. R&D, as well as getting stuff into space, is incredibly expensive.

[1] - https://en.wikipedia.org/wiki/Mars_Exploration_Rover

Having two rovers turned out to be insanely popular. Yes, it was expensive and NASA's budget is too small. But let's separate out the marginal cost here, because I'm pretty certain that if we had (for example) sent 4 rather than 2, it wouldn't necessarily have cost twice as much.

Many of us abide in disbelief of NASA's inability to leverage economies of scale. But like SpaceX and other private agencies have taught us, maybe this inability is a governmental-culture issue, not some sort of weird space-related issue.

Other responses nearby are good. I'd just like to point out that the instruments on the craft are unique and hand-assembled, and hand-tested through several phases (thermal, vacuum, radiation), so costs don't really go down with small multiples.

Here's a picture to get the flavor: http://www.jpl.nasa.gov/spaceimages/details.php?id=PIA12021

It's interesting to look at India's Mangalyaan Mars probe, which cost $74M, or about a tenth of the NASA orbiter sent at the same launch window. This was largely because Mangalyaan was made out of completely standard parts; it was a tech demo for ISRO's satellite bus. (And it only carried 15kg of instruments.) It also only took 15 months to build.

So it's certainly _possible_ to build cheap spacecraft using mass production. You won't get anything specialised, so they're unlikely to work beyond the inner solar system, but you'll get cheap, simple and tested probes.

Whether this is _worthwhile_ is a very difficult question. Spacecraft cost money to run, and you're going to have to do very careful cost-benefit analysis as to what's the most efficient use of your money long-term. Plus, standard parts only work in standard environments; Jupiter's radiation belts would most likely kill Mangalyaan stone dead, even if had enough delta-V to get there, even if there was enough sunlight to run it.

But it's certainly worth considering designing a standardised long distance spacecraft bus, especially for missions to the outer solar system. Maybe someone's already done that...

(Mangalyaan is still in orbit, still collecting science data, still has years' worth of propellant, and AFAICT from the internets, has been an utterly textbook mission. They're sending another one at the next transfer window.)

You could build and launch probes at regular intervals to L4 and L5, and dispatch them from there via low-energy transfer.

It would take months for each probe to get there, but once one arrives, you can transfer it from there to anywhere else in the universe far more easily than from the Earth's surface. And since those points are more stable than L1, L2, and L3, you can just leave the probe there for months or years until a good transfer opportunity comes up.

Reply on news.ycombinator.com