Rover brains?

As soon as I hear VxWorks I get the impression that they went out and
hired some engineers with C programming experience. These people do
not want to know what is really going on so they count on others to
write them an operating system and the finger pointing can begin. The
people who wrote the software probably have not read the actual code
written by the compiler and they have not read or do not understand
the real time system. They work in an abstract world of
multiprocessing, inter process messaging, semaphors, etc.

My experience has shown that I could get a factor of 100 improvement
over this above method and I could get absolute reliability -
updateability by switching over to assembly language and writing all
the code myself. I even wrote my own editor, assembler, development
system, operating system etc. Of course this is a labor intensive
method not suited for viewgraphs. The ancient art of assembly and
machine code programming is dying out and so nobody can make reliable
electronic systems anymore. No wonder I keep hearing of yaw dampers
crashing airplanes and thrust reversers deploying in flight.

Zoltan

Zoltan Szakaly wrote:
As soon as I hear VxWorks I get the impression that they went out and
hired some engineers with C programming experience. These people do
not want to know what is really going on so they count on others to
write them an operating system and the finger pointing can begin. The
people who wrote the software probably have not read the actual code
written by the compiler and they have not read or do not understand
the real time system. They work in an abstract world of
multiprocessing, inter process messaging, semaphors, etc.

My experience has shown that I could get a factor of 100 improvement
over this above method and I could get absolute reliability -
updateability by switching over to assembly language and writing all
the code myself. I even wrote my own editor, assembler, development
system, operating system etc. Of course this is a labor intensive
method not suited for viewgraphs. The ancient art of assembly and
machine code programming is dying out and so nobody can make reliable
electronic systems anymore. No wonder I keep hearing of yaw dampers
crashing airplanes and thrust reversers deploying in flight.

Zoltan


JPL switched from roll your own to VxWorks with the Mars Prospector, as
the budget didn't support the roll your own, not to mention the
nightmare that the typical cyclic executive creates from a maintenance
perspective every time you need to change the timing.

As far as those yaw dampers you heard about, those airplanes had analog
computers on them--at least the yaw damper was analog.

Not sure about the thrust reversers. Those airplanes may have had
digital engine controllers, which were probably written in Jovial or
assembly--with home-grown executives.

Realistically, writing with assembly and rolling your own executive is
fine for simple systems, but when you start getting to the levels of
complexity that require 32 bit hardware, writing your own and using
assembler may very well be more error-prone and certainly more costly
than using a well developed OS and language, whether commercial or open
source.

On Sun, 25 Jan 2004 17:44:48 GMT, John Ahrens
wrote:

As far as those yaw dampers you heard about, those airplanes had analog
computers on them--at least the yaw damper was analog.

The yaw dampers in many aircraft are mechanical or electromechanical,
not with analog computers.

Not sure about the thrust reversers. Those airplanes may have had
digital engine controllers, which were probably written in Jovial or
assembly--with home-grown executives.

The first production digital engine control system was probably the
spike controller retrofit for the SR-71, some time in the '80s. The
first digital FADEC/HIDEC/DEEC was flown on the Dryden F-15 about
then.

Thrust reversers are electromechanical. They don't use analog or
digital computers. They also don't use feedback. They're bang-bang
systems, so there's no need for proportional control.

Mary

--
Mary Shafer Retired aerospace research engineer