NASA's Bug Free Code?

[Pyrotex]

...Question is, does IBM actually bother to manufacture 10-bit Ram? In correction, you actually want to have some programmatic control over the correction process in some cases, so the cheap way to do this is simply alter your compiler so that it just (wastefully) uses 16-bits for each byte: having to double your RAM requirements would be a small fraction of the cost in off-the-shelf hardware compared to custom-built-stuff-that-nobody-but-NASA-wants hardware....

You are so correct. But in the early days of Space Station, when IBM managed to "control" so much of the program behind the scenes, the master design of the GPC (and the spinoff modem) was totally their baby. And something else: The GPC RAM (but not the modem) was designed to incorporate hard-wired Object Oriented Design!! :shrug: :eek_big: :evil: :eek_big: :D

 

Yeah, baby! RAM was super-grouped into (I believe) 64 byte OBJECTS, each of which contained the value of ONE variable. ONE!! The rest of the object contained hard-wired logic for inserting typing info, units, last change date, previous value, owner, security level, enable/disable states, blah blah blah. I think the value field was 24 bytes long. This was all IN the HARDWARE structure of the RAM itself!!

 

Now can you understand why IBM was "kicked" out of the Johnson Space Center community? And why we never used their GPC, even though they had a (one) working prototype by 1993? That was the end of the Space Station Freedom project. The ISS project began something like a year later.

 

John Neuman was right all along. Keep your objects in software.

[Leo]

Very interesting and funny thread. We should learn from the bacteria that thrive in nuclear power plants without upgrading to Hypercubical Polyrelational Metaobjectified SQL DNA .