[ale] [OT] Mars Lander!

JD jdp at algoloma.com
Tue Aug 7 07:03:52 EDT 2012


On 08/06/2012 10:33 PM, Michael H. Warfield wrote:
>  and they had to exploit and unintended backdoor in the
> machine to regain control  

I know they've patched the shuttle using manual inputs in hex.  Those computers
had that facility built-in and fully documented.  It definitely was not a back
door. The hex patch capability is like a deadly pill carried by spies.  Nobody
wants to use it, but there are thousands of scenarios where it could be useful.
It is really hard to correctly patch a system with dynamic memory allocation.
The programming language used on the shuttles didn't do dynamic memory
allocation. Every bit in memory was known. Heck, I had to get a waiver to use a
pointer to access elements in a huge array because it was larger than the HAL/S
compiler supported.

In the mid-90s, I had a different job at JSC working on the shuttle and station
flight control rooms writing software and managing an app.  They were hundreds
of DEC Alphas used, with a few HP-UX (2 or 3) and a few AIX (1-2) and 2 SunOS HA
NFS servers.  PCs were forbidden on the network.  If you ever see the blue
consoles at JSC on TV for space station support, those were the control rooms
that my company installed - everything from networking, hardware and almost all
software was included in the contract.  I think they moved one of the FCRs
somewhere else in Bldg-30-S in 2004, but can't figure out where. There were 4
FCRs when I was there. 2 were Apollo era and 2 were brand new.  I thought the
old FCRs would never be touched, left for history.  They weren't sealed, I've
sat in the Apollo FD chair when the room was empty.  What I remember thinking is
"I"m not worthy."

The app my team wrote was deployed to every POCC around the world and on the
laptops for shuttle and space station flights. I installed it on all the
workstations everywhere, POCCs, FCRs, launch rooms ... but only once on the
astronaut laptops.  I rubbed someone the wrong way during an initial training
session. One of them asked a question about the storage on disk, which was
pretty tiny, except the data for the program was huge and not stored on the
disk. He didn't like my answer, yelled at me and I was out.  I figure he was
having a bad day, one of the others told me that as I left.  Very high stress
job and being on call 24/7 sucks. I wasn't allowed to leave town. Even when a
mission wasn't flying, the software and FCRs were used constantly for training.

I was only there 7 yrs, but it was pretty amazing work and my job had some
pretty cool access.  Some of it sucked too, like all the disconnected networks
that needed user account maintenance all the time ... I'd spend half a day
walking around changing passwords every 28 days.  Most of the developers worked
100% in secured environments and networks. My team worked on the internet and I
had to "introduce" the code to the more secure networks. Our project had a way
to bring data in that didn't go through the same cumbersome method with lots of
paperwork. That was my job too.

Sorry to go on so long. I don't want to bore you with JSC stories.


More information about the Ale mailing list