4 ms·

> "The Shuttle suffers from its own Y2K problem"... "A software fix is possible, but it has never been worth the millions of dollars necessary to fix it." MILL

> "The Shuttle suffers from its own Y2K problem"... "A software fix is possible, but it has never been worth the millions of dollars necessary to fix it." MILLIONS of dollars?!

Klinky 15y ago
Well, if it's a mission critical system with millions of lines of code that need inspecting by highly trained engineers, I don't think that's so far fetched. Some Y2K problems might require significant refactoring. Then pile on code auditing & certification before it can go back into production systems. There were some comments here just a few days ago about NASA software: http://news.ycombinator.com/item?id=2736107 http://news.ycombinator.com/item?id=2736107
megablast 15y ago
I imagine for something like the shuttle, there would need to be an awful lot of testing to ensure nothing goes wrong. If a lot of systems rely on the Date code, they all need to be tested.
konstruktor 15y ago
The software for the space shuttle flight computers is produced using an incredibly disciplined process. Here's a very idealistic article about it http://www.fastcompany.com/magazine/06/writestuff.html http://www.fastcompany.com/magazine/06/writestuff.html The downside is that every change is prohibitively expensive. I didn't bookmark the article but remember to have read that the astronauts are taught lots of rules like "if the computer displays x and y, this means that z", stuff that would usually be a simple software feature. The new year bug seems to be quite similar.
ak217 15y ago
Yes. But I'm not sure if I understand why they would need the flight control computers to stay on if it were to be permanently docked. Of the reasons he listed, only the obstruction of docking ports strikes me as valid. In any case, I think it would have been really cool if they spent a few millions to retrofit the shuttle into a last-chance emergency crew return vehicle (so enable avionics to survive an extended attachment to external power). Then a few years later, if it's not needed, have it land on autopilot in Mojave (or ditch into the Pacific if something goes wrong).
- smithian 15y ago
  There are other reasons than the ones he listed. For instance, one reason that the Shuttle can remain docked to the ISS without damaging the docking connector is that the two vehicles employ their respective Reaction Control Systems (RCS) to keep precisely in alignment. The Shuttle RCS is Hypergolic and has limited fuel. You would need to totally redesign the system to keep the shuttle attached to the station long term. There are lots of other reasons as well. The Shuttle was not designed to be in orbit for more than a few weeks. For a crew return vehicle, it would probably be a lot cheaper to finish up an x-38 than retrofit the Shuttle. (and cheaper still to just keep doing what we are doing now, which is use Soyuz capsules)
dereg 15y ago
Gov't programs...
- Klinky 15y ago
  Private industry never had to spend that much on a Y2K fix. http://books.google.com/books?id=QVAEAAAAMBAJ&lpg=PA17&pg=PA17#v=onepage&q&f=false http://books.google.com/books?id=QVAEAAAAMBAJ&lpg=PA17&#...
kemayo 15y ago
Feynman's "Personal observations on the reliability of the Shuttle" [1] might help understand why this is so. Down in the "avionics" section. [1] http://www.fotuva.org/feynman/challenger-appendix.html http://www.fotuva.org/feynman/challenger-appendix.html

> "The Shuttle suffers from its own Y2K problem"... "A software fix is possible, but it has never been worth the millions of dollars necessary to fix it." MILLIONS of dollars?!

Klinky 15y ago

Well, if it's a mission critical system with millions of lines of code that need inspecting by highly trained engineers, I don't think that's so far fetched. Some Y2K problems might require significant refactoring. Then pile on code auditing & certification before it can go back into production systems. There were some comments here just a few days ago about NASA software: http://news.ycombinator.com/item?id=2736107 http://news.ycombinator.com/item?id=2736107

megablast 15y ago

I imagine for something like the shuttle, there would need to be an awful lot of testing to ensure nothing goes wrong. If a lot of systems rely on the Date code, they all need to be tested.

konstruktor 15y ago

The software for the space shuttle flight computers is produced using an incredibly disciplined process. Here's a very idealistic article about it http://www.fastcompany.com/magazine/06/writestuff.html http://www.fastcompany.com/magazine/06/writestuff.html The downside is that every change is prohibitively expensive. I didn't bookmark the article but remember to have read that the astronauts are taught lots of rules like "if the computer displays x and y, this means that z", stuff that would usually be a simple software feature. The new year bug seems to be quite similar.

ak217 15y ago

Yes. But I'm not sure if I understand why they would need the flight control computers to stay on if it were to be permanently docked. Of the reasons he listed, only the obstruction of docking ports strikes me as valid. In any case, I think it would have been really cool if they spent a few millions to retrofit the shuttle into a last-chance emergency crew return vehicle (so enable avionics to survive an extended attachment to external power). Then a few years later, if it's not needed, have it land on autopilot in Mojave (or ditch into the Pacific if something goes wrong).

smithian 15y ago
There are other reasons than the ones he listed. For instance, one reason that the Shuttle can remain docked to the ISS without damaging the docking connector is that the two vehicles employ their respective Reaction Control Systems (RCS) to keep precisely in alignment. The Shuttle RCS is Hypergolic and has limited fuel. You would need to totally redesign the system to keep the shuttle attached to the station long term. There are lots of other reasons as well. The Shuttle was not designed to be in orbit for more than a few weeks. For a crew return vehicle, it would probably be a lot cheaper to finish up an x-38 than retrofit the Shuttle. (and cheaper still to just keep doing what we are doing now, which is use Soyuz capsules)

dereg 15y ago

Gov't programs...

Klinky 15y ago
Private industry never had to spend that much on a Y2K fix. http://books.google.com/books?id=QVAEAAAAMBAJ&lpg=PA17&pg=PA17#v=onepage&q&f=false http://books.google.com/books?id=QVAEAAAAMBAJ&lpg=PA17&#...

kemayo 15y ago

Feynman's "Personal observations on the reliability of the Shuttle" [1] might help understand why this is so. Down in the "avionics" section. [1] http://www.fotuva.org/feynman/challenger-appendix.html http://www.fotuva.org/feynman/challenger-appendix.html