If you worry about software updatates going wrong..

[Wayne]

NASA upgrades Mars Curiosity software ... from 350M miles away

http://www.computerworld.com/s/artic...50M_miles_away

[secipolla]

If the Martians don't want it to work, they turn it off, if they do want it to work, they even fix the upgrades, if needed.
Of course we need to do our part by doing our best.

[stevea]

Think they have a gpg signature on that upload ?

[Wayne]

Curiosity's CPU is also interesting, it's a 200 megahertz BAE PowerPC RAD750... Speed demon, eh?

http://news.cnet.com/8301-11386_3-57...uilt-for-mars/

[DBelton]

Actually, Curiosity is going to find methane. It will also find evidence of water. They might even show a building or two.

NASA needs to kick up public support so they can demand the gooberment to continue their funding.

But, it does seem odd that they would need to do an update just days after the rovers mission starts.

[Wayne]

According to the article, it's more of a replacing flight and landing software with scientific software...

[DBelton]

I missed that in the article. I guess I need to read closer.

Now that does make sense. I'm sure that memory/storage is having to be conserved wherever possible, and the flight/landing software is no longer needed, so replace it with something that is useful.

[Gareth Jones]

Originally Posted by Wayne

Curiosity's CPU is also interesting, it's a 200 megahertz BAE PowerPC RAD750... Speed demon, eh?

Given the conditions they're exposed to in space (and on Mars), probes generally have slower more fault-tolerant CPUs, rather than the smallest domestic transistor technology which is sensitive to radiation etc. Given the delay between communication due to the speed-of-light limit, there isn't much need for good interactive performance either...

[droidhacker]

There are two very powerful motivations for keeping things simple in situations like that one;
1) Power consumption. A 140 Watt desktop CPU won't run for very long on solar hybrid.
2) Reliability. As systems increase in complexity, they become drastically more complex to test and debug, and the chances of leaving in some hidden fault increases. If I recall correctly, it is exponential. It would really suck to spend millions of dollars to send something over to Mars and have it self destruct the moment it touches ground.

[John the train]

Originally Posted by droidhacker

There are two very powerful motivations for keeping things simple in situations like that one;
1) Power consumption. A 140 Watt desktop CPU won't run for very long on solar hybrid.
2) Reliability. As systems increase in complexity, they become drastically more complex to test and debug, and the chances of leaving in some hidden fault increases. If I recall correctly, it is exponential. It would really suck to spend millions of dollars to send something over to Mars and have it self destruct the moment it touches ground.

I seem to recall that one Mars probe was lost because of a simple bug in its code ( divide by zero? ).

[billybob linux]

There is ome info about it on the NASA website here:

http://www.nasa.gov/mission_pages/msl/index.html

Curiosity Preps for a Planned 'Brain Transplant'
Sat, 11 Aug 2012 12:34:28 AM GMT+0100

A healthy Curiosity spent Sol 4, its fifth day on Mars, preparing for this weekend's planned "brain transplant"--transitioning to a new version of flight software on both of Curiosity's redundant main computers. The new software is better suited for Mars surface operations, such as driving and using Curiosity's robotic arm. The "brain transplant" will take place during a series of steps beginning this evening and continuing through Aug. 13. The new software was uploaded to the rover's memory during the Mars Science Laboratory spacecraft's flight from Earth. Key capabilities in the new software enable full use of Curiosity's powerful robotic arm and drill, and advanced image processing to check for obstacles while driving. This will ultimately allow Curiosity to make longer drives by giving the rover more autonomy to identify and avoid potential hazards and to drive along a safe path that the rover identifies for itself.



The flight team elected to defer further science activities until after the flight software transition is completed. Several Mastcam color images were downlinked. The mission's science team is busy analyzing images Curiosity has taken of its surroundings, and is discussing features of interest that they will investigate in a few weeks once initial checkouts of the rover and the landing site are completed.

From reading that it would appear that the software had already been loaded during flight and they are "transitioning" to the new version. Looking at the dates it should be well underway by now or possibly already completed.

[Gareth Jones]

I remember one was supposedly lost because of metric/imperial unit confusion in the code...

[Penguinclaw]

Originally Posted by Gareth Jones

I remember one was supposedly lost because of metric/imperial unit confusion in the code...

Also why Hubble needed "a pair of spectacles"! The mirror was ground to the wrong specs because of a miss calculation on conversion which required NASA to use a corrective lens to focus the light correctly. Once done though what a magnificent satellite.

I am always in awe of the original Apollo missions to the moon and the hardware that enabled them to land.... it just proves that it's not what you've got, it's how you use it!

[John the train]

Originally Posted by Gareth Jones

I remember one was supposedly lost because of metric/imperial unit confusion in the code...

Shades of the ' Gimli glider ' - an airliner ran out of fuel mid-air because the airline was changing from lbs to kgs for fuel measurement and someone used the wrong conversion factor. Using two systems in parallel is fraught with problems.

[billybob linux]

More about the RAD750 and quite an interesting article here:

http://www.wired.com/wiredenterprise...8/cosmic-rays/