NASA and OpenSource

ookabooka · 11-18-03, 11:33 PM

I was just talking to a friend of mine, and we both happened to be in the same astronomy class, somehow we got on debating how the simple mistake that caused the mars polar lander could have been avoided. Almost instantly, we thought of opensource. If NASA had released their AI as open source, millions would have gone over the code, undoubtably one would have caught the error, even if he/she wasn't a rocket scientist. So far I have found limited documentation on whether or not NASA is considering this.
This is a little info I found on the subject. My friend and I were thiking of drafting a paper to send to NASA about releasing some of their code. We came up with this list of pros and cons:
PROS:
Easily detect mistakes
Incredibly complex programs (an individual is responsible for a certain function, and not a section of the program)
Easy, free workforce

CONS:
Job security for NASA programmers(not too significant, gotta have someone who knows what they are doing)
Ability for someone to find a security hole in the program and exploiting it (We came to the conclusion that all I/O would be handled by NASA exclusively)
Integration(always a problem with opensource)

What do you all think, I was just brainstorming and thoguht they were pretty nifty ideas. I mean common, how cool would it to have the next lander navigating using one of your functions.

XWRed1 · 11-19-03, 12:47 AM

The problem with the probe was a logic error. Generally, NASA's code is supposed to be much more meticulus, their programmers only put out a 10th of the amount of code as someone at a normal company because there is so much analysis and reworking. They even say they have multiple groups work on the same thing so they have multiple solutions to choose from.

That said, the logic error in the probe was in converting between metric and imperial. Some components from Boeing was using imperial and the components from Nasa were using metric, the math didn't match up... and so the probe thought the surface of mars was a little further away.

If anything, it is an argument for the proliferation of the metric system.

engjohn · 11-19-03, 07:55 AM

Quote:

Originally posted by XWRed1

If anything, it is an argument for the proliferation of the metric system.

please....
OH DEAR GOD PLEASE can we just use the Metric system..... PLEASE....
The stupid system we have sucks! It was designed by and IDIOT!

Christoph · 11-19-03, 11:28 AM

Quote:

Originally posted by engjohn

please....
OH DEAR GOD PLEASE can we just use the Metric system..... PLEASE....
The stupid system we have sucks! It was designed by and IDIOT!

Worse than that, It wasn't even designed.

BTW XWRed1, where did you read about how NASA programmers code? This perhaps? Sounds like some fascinating stuff.

Ebola · 11-19-03, 12:27 PM

Quote:

Originally posted by engjohn

please....
OH DEAR GOD PLEASE can we just use the Metric system..... PLEASE....
The stupid system we have sucks! It was designed by and IDIOT!

True so true. Good luck converting joe sixpack.

XWRed1 · 11-19-03, 06:06 PM

I'm not sure where I read about NASA's coding practices.

I did read about that probe failure on the prime95 forums, a NASA engineer posts there and he wrote a pretty lengthy explanation. He might have mentioned NASA's coding practices also, it was back in June/July when I read this.

ookabooka · 11-19-03, 06:49 PM

I know that the error was in a conversion, but I think that it could have been avoided if the source had been public, granted, most people would trust nasa's numbers and work on other areas of the program, but there would be one guy (with way too much time on his hands) that would research the numbers, and find the error. How many times do you sit debugging a program for hours, when someone walks by and points out a typo you made, and then proceeds to make fun of you. I did this to my friend (including the make fun of part) when I saw he said that there were 356.25 days in a year. This happens all the time in programming, finding your own errors can be difficult, while teams of programmers can solve this problem, I still think the best bet would be to release the source. If you wanted different ways of going about programming something, open source is definately the way. You could easily have thousands of different versions of a function written by people. I think that NASA should give open source a shot, with some sort of test trial. I believe the UK did this with a satellite they have in orbit, though I cannot remember the name.

edit: Since NASA typically has such perfect software, I think the most benefit would be from creating a more complex program, rather than finding errors (as there ought to be few that slip past NASA)

jajmon · 11-19-03, 07:27 PM

"If NASA had released their AI as open source, millions would have gone over the code, undoubtably one would have caught the error, even if he/she wasn't a rocket scientist"

With the above statement, I suppose you would have the Armed forces of the USA, CIA, FBI etc, have all of their programs released to open source with all of the hackers and terrorists out there. SHEESH, come on here, this is national security. It is possible that if NASA released the code to opensource, it may have been found. Hind sight is 20/20 and NASA knows the problem and will make policy/procedure changes to correct, so as not to have it happen in the future. At least no lifes were lost - (silver lining in a cloud?)

Anyone remember 9/11? and all of the security put in place as a result?

Just take some time to ponder what you are suggesting and the results.

ookabooka · 11-19-03, 10:31 PM

Good thing we don't trust opensource to encrypt all of our data, or run computers than have peoples social security info on them and credit card numbers etc, oh wait, we do (OpenSSL,OpenSSH, Linux) Open source is not as big of a security risk as you would think. True a "hacker" could see a security hole and exploit it, but it is also possible for a good programmer to see it and fix it, in the end, I think it balences out. Besides, like I said, some parts like I/O would be private to NASA, wouldn't want anyone hijacking the lander and making it kill without moral compunction, or go al masacistic(sp?) and take a welding torch to its solar panels. I'm talking about the software that takes in images, and makes a 3d-map of the area and then navigates the area, or one that effectively compresses images to send back home. Besides, some sectors of the government are turning to linux, because it is a good os, and cuts taxes (always a good thing).

SickBoy · 11-19-03, 10:51 PM

Quote:

Originally posted by jajmon
Anyone remember 9/11? and all of the security put in place as a result?

Amen. This is why Apple can't sell G4's and G5's outside the United States.

My company works with DoD contract companies. You think ISO 9001 was strict? Try DoD contractor rules. Everything has to be AUDITABLE, 100% TRACEABLE and REPEATABLE. The DoD contract companies require the same standards out of the companies they work with. Sorry, but Open Source would never fly here.

You may say "If the code were open source, someone would have found the problem." What are you going to tell the families of the dead if, conversely, somebody's code checkin to an OSS project caused the International Space Station to crash back to earth? Sorry. OSS in high risk or defense uses won't fly. Period.

And as much as I am a fan of the open source community, I prefer it that way.

XWRed1 · 11-20-03, 04:09 AM

I think it is just a matter of utility. There isn't much utility in open sourcing your internal software. The stuff that goes into Nasa's probes would count as internal software.

Stuff that is publically released in the first place is usually better when open sourced though.

I don't care about terrar concerns or security by obscurity. Those would be some of the silliest reasons not to open source something. Its just a matter of utility.

Stedeman · 11-20-03, 04:30 AM

You guys missed another point. Who will sort all the incoming reviews? You know 10K idiots think they found an error and now NASA has to spend 3 years sorting through the fakes (non-error-ed errors

) they just need to stay of the cheaper = better bandwagon and all will be fine.

ookabooka · 11-20-03, 10:36 PM

Ok, I want to clarify what I am thinking of. I am not thinking of developing software in open source that could potentially kill people, no open source life support systems, no way. I am talking about software that runs probes, or satellites. Also, I'm not suggesting the entire project be open source, just some areas, such as imaging, compresssion, navigation, etc. Sorting the propsed code could be difficult, this would probably become the number one job of NASA programmers. You could have civillians/comittees, that have demonstrated trustworthiness, be inbetween the average user and NASA, filtering out worthless suggestions. Standards could become an issue too, though I am confident that NASA can optimise the code to make it integrate better, though opensource typically adapts to new standards pretty quickly.

jajmon · 11-21-03, 11:46 PM

Thank you for clarifying. I still think it is prudent that our US government angencies hold dear to their heart 'all' of their software and not offer 'any' aspect to open source, no matter how trivial it may seem. Sorry, I'm republican, don't own any guns, not a vet and I'm not a redneck. I fly the US flag 24/7 and proud of it.

ookabooka · 11-22-03, 02:40 AM

I don't see how relesing some NASA code as open source, or creating some open source projects for NASA, could be seen as a bad thing. I thought that NASA was a organization with human exploration and curisoty at heart, and teamwork at its core (ISS comes to mind). It seems like it would be a good thing for a unified effort on space exploration, instead of a top secret governemnt institution, with all security beaches promtly stomped out.

XWRed1 · 11-22-03, 05:23 PM

I just don't think there is really any point if it is code that only NASA would really use. People would only look at it then, not many people would be deploying it themselves because they don't have a use of it.

The usefulness of open source is when it is something alot of people use that is open.

11-18-03, 11:33 PM	Thread Starter #1
ookabooka Member Join Date: Sep 2002 Location: root@ localhost.localdomain	NASA and OpenSource I was just talking to a friend of mine, and we both happened to be in the same astronomy class, somehow we got on debating how the simple mistake that caused the mars polar lander could have been avoided. Almost instantly, we thought of opensource. If NASA had released their AI as open source, millions would have gone over the code, undoubtably one would have caught the error, even if he/she wasn't a rocket scientist. So far I have found limited documentation on whether or not NASA is considering this. This is a little info I found on the subject. My friend and I were thiking of drafting a paper to send to NASA about releasing some of their code. We came up with this list of pros and cons: PROS: Easily detect mistakes Incredibly complex programs (an individual is responsible for a certain function, and not a section of the program) Easy, free workforce CONS: Job security for NASA programmers(not too significant, gotta have someone who knows what they are doing) Ability for someone to find a security hole in the program and exploiting it (We came to the conclusion that all I/O would be handled by NASA exclusively) Integration(always a problem with opensource) What do you all think, I was just brainstorming and thoguht they were pretty nifty ideas. I mean common, how cool would it to have the next lander navigating using one of your functions.
	QUOTE Thanks

11-19-03, 12:47 AM	#2
XWRed1 Senior Member Join Date: Oct 2001	The problem with the probe was a logic error. Generally, NASA's code is supposed to be much more meticulus, their programmers only put out a 10th of the amount of code as someone at a normal company because there is so much analysis and reworking. They even say they have multiple groups work on the same thing so they have multiple solutions to choose from. That said, the logic error in the probe was in converting between metric and imperial. Some components from Boeing was using imperial and the components from Nasa were using metric, the math didn't match up... and so the probe thought the surface of mars was a little further away. If anything, it is an argument for the proliferation of the metric system.
	QUOTE Thanks

11-19-03, 06:06 PM	#6
XWRed1 Senior Member Join Date: Oct 2001	I'm not sure where I read about NASA's coding practices. I did read about that probe failure on the prime95 forums, a NASA engineer posts there and he wrote a pretty lengthy explanation. He might have mentioned NASA's coding practices also, it was back in June/July when I read this.
	QUOTE Thanks

11-19-03, 06:49 PM	Thread Starter #7
ookabooka Member Join Date: Sep 2002 Location: root@ localhost.localdomain	I know that the error was in a conversion, but I think that it could have been avoided if the source had been public, granted, most people would trust nasa's numbers and work on other areas of the program, but there would be one guy (with way too much time on his hands) that would research the numbers, and find the error. How many times do you sit debugging a program for hours, when someone walks by and points out a typo you made, and then proceeds to make fun of you. I did this to my friend (including the make fun of part) when I saw he said that there were 356.25 days in a year. This happens all the time in programming, finding your own errors can be difficult, while teams of programmers can solve this problem, I still think the best bet would be to release the source. If you wanted different ways of going about programming something, open source is definately the way. You could easily have thousands of different versions of a function written by people. I think that NASA should give open source a shot, with some sort of test trial. I believe the UK did this with a satellite they have in orbit, though I cannot remember the name. edit: Since NASA typically has such perfect software, I think the most benefit would be from creating a more complex program, rather than finding errors (as there ought to be few that slip past NASA) __________________ Linux motto: "Linux-give yourself a hard time " Microsoft motto: "Windows-You pay us, and give yourself a hard time " Obviously Linux > Windows From root (/) : mount -t proc proc proc "Listen to me boy, I've made a living out of being a failure, and you sir are not a failure!" -Mantis, Disney's A Bugs Life One minute he's washing grapes, the next minute he's eating them. . . I guess that's logical.
	QUOTE Thanks

11-19-03, 07:27 PM	#8
jajmon Member Join Date: Apr 2002 Location: Burnsville, Minnesota	"If NASA had released their AI as open source, millions would have gone over the code, undoubtably one would have caught the error, even if he/she wasn't a rocket scientist" With the above statement, I suppose you would have the Armed forces of the USA, CIA, FBI etc, have all of their programs released to open source with all of the hackers and terrorists out there. SHEESH, come on here, this is national security. It is possible that if NASA released the code to opensource, it may have been found. Hind sight is 20/20 and NASA knows the problem and will make policy/procedure changes to correct, so as not to have it happen in the future. At least no lifes were lost - (silver lining in a cloud?) Anyone remember 9/11? and all of the security put in place as a result? Just take some time to ponder what you are suggesting and the results.
	QUOTE Thanks

NASA and OpenSource

Internet Services

Website Development

Tech Hardware

11-19-03, 10:31 PM	Thread Starter #9
ookabooka Member Join Date: Sep 2002 Location: root@ localhost.localdomain	Good thing we don't trust opensource to encrypt all of our data, or run computers than have peoples social security info on them and credit card numbers etc, oh wait, we do (OpenSSL,OpenSSH, Linux) Open source is not as big of a security risk as you would think. True a "hacker" could see a security hole and exploit it, but it is also possible for a good programmer to see it and fix it, in the end, I think it balences out. Besides, like I said, some parts like I/O would be private to NASA, wouldn't want anyone hijacking the lander and making it kill without moral compunction, or go al masacistic(sp?) and take a welding torch to its solar panels. I'm talking about the software that takes in images, and makes a 3d-map of the area and then navigates the area, or one that effectively compresses images to send back home. Besides, some sectors of the government are turning to linux, because it is a good os, and cuts taxes (always a good thing). __________________ Linux motto: "Linux-give yourself a hard time " Microsoft motto: "Windows-You pay us, and give yourself a hard time " Obviously Linux > Windows From root (/) : mount -t proc proc proc "Listen to me boy, I've made a living out of being a failure, and you sir are not a failure!" -Mantis, Disney's A Bugs Life One minute he's washing grapes, the next minute he's eating them. . . I guess that's logical.
	QUOTE Thanks

11-20-03, 04:09 AM	#11
XWRed1 Senior Member Join Date: Oct 2001	I think it is just a matter of utility. There isn't much utility in open sourcing your internal software. The stuff that goes into Nasa's probes would count as internal software. Stuff that is publically released in the first place is usually better when open sourced though. I don't care about terrar concerns or security by obscurity. Those would be some of the silliest reasons not to open source something. Its just a matter of utility.
	QUOTE Thanks

11-20-03, 04:30 AM	#12
Stedeman The Half Asleep Member Join Date: Aug 2002 Location: Lewiston Maine	You guys missed another point. Who will sort all the incoming reviews? You know 10K idiots think they found an error and now NASA has to spend 3 years sorting through the fakes (non-error-ed errors ) they just need to stay of the cheaper = better bandwagon and all will be fine. __________________ My HEATWARE My eBay
	QUOTE Thanks

11-20-03, 10:36 PM	Thread Starter #13
ookabooka Member Join Date: Sep 2002 Location: root@ localhost.localdomain	Ok, I want to clarify what I am thinking of. I am not thinking of developing software in open source that could potentially kill people, no open source life support systems, no way. I am talking about software that runs probes, or satellites. Also, I'm not suggesting the entire project be open source, just some areas, such as imaging, compresssion, navigation, etc. Sorting the propsed code could be difficult, this would probably become the number one job of NASA programmers. You could have civillians/comittees, that have demonstrated trustworthiness, be inbetween the average user and NASA, filtering out worthless suggestions. Standards could become an issue too, though I am confident that NASA can optimise the code to make it integrate better, though opensource typically adapts to new standards pretty quickly. __________________ Linux motto: "Linux-give yourself a hard time " Microsoft motto: "Windows-You pay us, and give yourself a hard time " Obviously Linux > Windows From root (/) : mount -t proc proc proc "Listen to me boy, I've made a living out of being a failure, and you sir are not a failure!" -Mantis, Disney's A Bugs Life One minute he's washing grapes, the next minute he's eating them. . . I guess that's logical.
	QUOTE Thanks

11-21-03, 11:46 PM	#14
jajmon Member Join Date: Apr 2002 Location: Burnsville, Minnesota	Thank you for clarifying. I still think it is prudent that our US government angencies hold dear to their heart 'all' of their software and not offer 'any' aspect to open source, no matter how trivial it may seem. Sorry, I'm republican, don't own any guns, not a vet and I'm not a redneck. I fly the US flag 24/7 and proud of it.
	QUOTE Thanks

11-22-03, 02:40 AM	Thread Starter #15
ookabooka Member Join Date: Sep 2002 Location: root@ localhost.localdomain	I don't see how relesing some NASA code as open source, or creating some open source projects for NASA, could be seen as a bad thing. I thought that NASA was a organization with human exploration and curisoty at heart, and teamwork at its core (ISS comes to mind). It seems like it would be a good thing for a unified effort on space exploration, instead of a top secret governemnt institution, with all security beaches promtly stomped out. __________________ Linux motto: "Linux-give yourself a hard time " Microsoft motto: "Windows-You pay us, and give yourself a hard time " Obviously Linux > Windows From root (/) : mount -t proc proc proc "Listen to me boy, I've made a living out of being a failure, and you sir are not a failure!" -Mantis, Disney's A Bugs Life One minute he's washing grapes, the next minute he's eating them. . . I guess that's logical.
	QUOTE Thanks

11-22-03, 05:23 PM	#16
XWRed1 Senior Member Join Date: Oct 2001	I just don't think there is really any point if it is code that only NASA would really use. People would only look at it then, not many people would be deploying it themselves because they don't have a use of it. The usefulness of open source is when it is something alot of people use that is open.
	QUOTE Thanks

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search