Criminal incompetence at the FAA

"zatatime" wrote in message ...
].

I doubt it will happen, but I hpe they go back to using a real OS like
UNIX for this stuff, else progress will become an oxymoron.

Oh, like UNIX never crashes...

"Ron Natalie" wrote in message om...
"zatatime" wrote in message ...
].

I doubt it will happen, but I hpe they go back to using a real OS like
UNIX for this stuff, else progress will become an oxymoron.

Oh, like UNIX never crashes...

I guess it can. However, I can say that for the 10 years I've sat in
front of a Solaris workstation I have never once had it lock up do to
an OS issue. I have had it die and say a disk was bad or a memory SIMM
was bad, but it has never locked up or even hung for no reason. The
$4K Win XP Pro lap to they issued me is another story... Having
said that, I've long since given up being a UNIX bigot and spend most
of my time working in front of the Windows box. If Sun could ever
figure out a way to make an Ok Java engine for Solaris I might switch
back.

-Robert

Shows the damage that can be done when people who don't understand the
system try to read about it in the paper. I write software to manage
telecommunications infrastructure and I've never seen software that
can run forever without maintenance. We talk in term of the number of
"9's". 2 "9"s is 99% up time, 3 "9s" is 99.9%, etc. Usually 4 "9s" is
the best you can do. When my software fails, new telephone services
(new DSL, call waiting ,etc) do not get activated, wireless companies
cannot detect a tower outage, etc. Even in that environment 4 "9s" is
consider exceptional. The solution is to have backup systems. The
FAA's system in question was the backup system.

-Robert


(No Such User) wrote in message ...
Whoever made this purchase should be in jail:

http://software.silicon.com/applicat...9124122,00.htm

It's shameful that they would trust people's lives to a computer
system that had to be "reset every thirty days" just to keep from
crashing.

Microsoft delenda est.

Funnily enough, I've also worked on telecom infrastructure, mostly for
keeping international traffic flowing around the world. UNIX is what
is used for those machines. We also strive for 99.99% up time and
know Microsoft can't give it to us through stress testing. Now I'm
talking about tracking 100 million plus phone calls a day so I'll
admit the scale is something out of the ordinary. Its my opinion that
real time systems should be cautiously reviewed.

As far as my "facts" on Microsoft; the only thing I can offer is
directly from a friend who was one of the leads on building the NT 4.0
kernel. He very candidly told me that everyone within Microsoft's NT
architecture group knew the proclaimed 127 year up time would never be
achieved, and the it was a purely theoretical number. You have
experienced some of the reasons it can't be proven i.e electrical
failure. What the designers worked toward was having a machine that
could run for 30 days without a re-boot. We have also spoken about
real time applications from a medicinal use point of view, and his
take was that it would be a few generations down the road before the
Microsoft OS was ready for such a thing. I'm sure you can easily
discount this if you chose, but I'm speaking from a personal source
who spent at least a half day a week with Mr. Bill himself working
through all that was required to build the NT platform. For me that's
alot better than was I read in any computing rag (especially knowing
how thorough he is).

Now I don't know how ATC applications compare with real time medicinal
computing, but I have to think they are just as critical as each other
and should be treated with an overly adequate computing platform in
both the OS and the program design. I don't see the OS or programming
tools available (with the exception of C) ready to do that for apps
requiring significant up time for real time data analysis. Please
don't misunderstand me, I am not against Microsoft. They have
afforded me to make a good living for many years, but I don't think
their place is in real time computing......yet.

z
On 22 Sep 2004 18:13:30 -0700, (Robert M. Gary)
wrote:

Shows the damage that can be done when people who don't understand the
system try to read about it in the paper. I write software to manage
telecommunications infrastructure and I've never seen software that
can run forever without maintenance. We talk in term of the number of
"9's". 2 "9"s is 99% up time, 3 "9s" is 99.9%, etc. Usually 4 "9s" is
the best you can do. When my software fails, new telephone services
(new DSL, call waiting ,etc) do not get activated, wireless companies
cannot detect a tower outage, etc. Even in that environment 4 "9s" is
consider exceptional. The solution is to have backup systems. The
FAA's system in question was the backup system.

-Robert


(No Such User) wrote in message ...
Whoever made this purchase should be in jail:

http://software.silicon.com/applicat...9124122,00.htm

It's shameful that they would trust people's lives to a computer
system that had to be "reset every thirty days" just to keep from
crashing.

Microsoft delenda est.

zatatime wrote:

Funnily enough, I've also worked on telecom infrastructure, mostly for
keeping international traffic flowing around the world. UNIX is what
is used for those machines. We also strive for 99.99% up time and
know Microsoft can't give it to us through stress testing.

And I wrote system requirements for telecom software that had the same (or worse)
uptime requirements. We found the MS servers to be quite capable of delivering what
we needed.

George Patterson
If a man gets into a fight 3,000 miles away from home, he *had* to have
been looking for it.

G.R. Patterson III wrote:
zatatime wrote:
Robert M. Gary wrote:

I write software to manage telecommunications infrastructure ...

Funnily enough, I've also worked on telecom infrastructure ...

And I wrote system requirements for telecom software ...

And there's a fair bit of my code out there, primarily in the UK's
PSTN.

zatatime wrote in message . ..
Funnily enough, I've also worked on telecom infrastructure, mostly for
keeping international traffic flowing around the world. UNIX is what
is used for those machines.

All of our customers run their servers on Solaris or HPUX. However,
clients are almost always run on PCs (usually Win 2K). Our customers
are most of the cell phone companies and all the long distance
companies. Sometimes we manage faults, sometimes activation, etc so we
don't necessarily do the same thing for each co. However, if you call
your phone company an order DSL service anywhere in the U.S. I can
almost guarantee you that its my code that's actually connecting to
the switch to turn on the service.

-Robert