Glass panels: what OS?

In article , Greg Copeland
wrote:

Having said that, if real
world use shows the device to be rock solid, then I think that speaks
volumes, clearly out weighing the voice of concern.

The trick is to truly show that the device was indeed rock solid, and for
that history to be actually applicable to future use.

--
Bob Noel

On Wed, 23 Jun 2004 22:59:18 +0000, Bob Noel wrote:

In article , Greg Copeland
wrote:

Having said that, if real
world use shows the device to be rock solid, then I think that speaks
volumes, clearly out weighing the voice of concern.

The trick is to truly show that the device was indeed rock solid, and for
that history to be actually applicable to future use.

Agreed.

"Greg Copeland" wrote in message
news
On Wed, 23 Jun 2004 08:15:41 -0700, C J Campbell wrote:


"Bruce Horn" wrote in message
news
I've been wondering...

What underlying OS do the various glass panels (Avidyne, Garmin, etc.)
use? It might actually make me think twice about buying a particular
system if I knew that (for example) it was Windows underneath.

The MX-20 runs plain vanilla Windows NT 4.0. I don't know why anyone
except
software bigots would have a problem with that. There have been no known
problems with the OS in this application.

Well, chances are, it's actually embedded NT, but I'll defer if you
factually know otherwise. The concern is, NT has a long history of
crashing and being less than stable. This is true of embedded NT, but to
a much lessor degree. So, for someone to have concern about the heart of
an important navigation tool, I think falls well outside of simple OS
bigotry.

The Apollo MX-20 boot-up screen says Windows NT 4.0

It has never failed. I have, however, seen the OS crash on Garmin handheld
GPS units. Frequently. To be honest, I would prefer the more stable Windows
OS.

All operating systems have a long history of crashing and being less than
stable. It would be interesting to know why you think Win NT would be
unstable on something like the MX-20. The device is dedicated to running one
program. It has no peripherals. It never runs for more than a few hours.
Basically, all the issues supposedly making Win NT unstable simply do not
exist on a closed box like this.

On Thu, 24 Jun 2004 07:44:49 -0700, C J Campbell wrote:

Well, chances are, it's actually embedded NT, but I'll defer if you
factually know otherwise. The concern is, NT has a long history of
crashing and being less than stable. This is true of embedded NT, but to
a much lessor degree. So, for someone to have concern about the heart of
an important navigation tool, I think falls well outside of simple OS
bigotry.

The Apollo MX-20 boot-up screen says Windows NT 4.0

I'm actually not sure that it makes the distinction during startup.
Technically, the embedded kernel is a slightly different animal from the
desktop/server brethren. No paging support, for example.


It has never failed. I have, however, seen the OS crash on Garmin handheld
GPS units. Frequently. To be honest, I would prefer the more stable Windows
OS.


Well, stability is always stated from a relative frame of reference.
Those that work higher up on the stability scale, tend to look down at NT
and consider it a toy OS.

All operating systems have a long history of crashing and being less than

That's simply not true. NT has a long history of having one of the worst
stability records in the entire modern history of IT, short only of DOS
and perhaps early MACs (which had no MMU).

stable. It would be interesting to know why you think Win NT would be
unstable on something like the MX-20.

That would be because the OS is known to have stability issues and is
often less than reliable.

The device is dedicated to running one
program.

Yes, but that says very little about what's actually going on under the
covers. I don't have those details so I can only say we're probably both
ignorant of what's going on there.

It has no peripherals. It never runs for more than a few hours.

This is probably one of the saving graces for it. One of the problems
common to NT, especially in the 3-4.x days, is a number of memory leaks in
the kernel. I believe I remember reading that even the embedded kernel
still suffered from memory leaks, but I would not be willing to walk out
on a limb with that assertion.

Basically, all the issues supposedly making Win NT unstable simply do
not exist on a closed box like this.

With all due respect, that's simply not true. MS has had a number of
issues with their OS, ranging from memory leaks to kernel crashes. The
important question, as it relates to this topic, does the application in
question trigger any of the known problems and/or bugs with the kernel?
Which is why I asserted that the real world performance should certainly
override the list of valid and well supported concerns. Notice that I am
not saying, never buy a device which has a MS OS in it. I'm simply
saying, use caution and let real world experience be your guide.

Personally, if I learn that a device is running a MS OS, I immediately
consider the device to be suspect until proven otherwise. That doesn't
mean that the alternative implementations (other devices) will always be
problem free. Just the same, the inclusion of a MS OS in a device should
always be treated as a yellow flag. Which means, use caution until proven
it's no longer needed.


Cheers,

Greg

Greg Copeland wrote in
news
On Thu, 24 Jun 2004 07:44:49 -0700, C J Campbell wrote:

Well, chances are, it's actually embedded NT, but I'll defer if you
factually know otherwise. The concern is, NT has a long history of
crashing and being less than stable. This is true of embedded NT,
but to a much lessor degree. So, for someone to have concern about
the heart of an important navigation tool, I think falls well
outside of simple OS bigotry.

The Apollo MX-20 boot-up screen says Windows NT 4.0

I'm actually not sure that it makes the distinction during startup.
Technically, the embedded kernel is a slightly different animal from
the desktop/server brethren. No paging support, for example.


It has never failed. I have, however, seen the OS crash on Garmin
handheld GPS units. Frequently. To be honest, I would prefer the more
stable Windows OS.


Well, stability is always stated from a relative frame of reference.
Those that work higher up on the stability scale, tend to look down at
NT and consider it a toy OS.

All operating systems have a long history of crashing and being less
than

That's simply not true. NT has a long history of having one of the
worst stability records in the entire modern history of IT, short only
of DOS and perhaps early MACs (which had no MMU).

stable. It would be interesting to know why you think Win NT would be
unstable on something like the MX-20.

That would be because the OS is known to have stability issues and is
often less than reliable.

The device is dedicated to running one
program.

Yes, but that says very little about what's actually going on under
the covers. I don't have those details so I can only say we're
probably both ignorant of what's going on there.

It has no peripherals. It never runs for more than a few hours.

This is probably one of the saving graces for it. One of the problems
common to NT, especially in the 3-4.x days, is a number of memory
leaks in the kernel. I believe I remember reading that even the
embedded kernel still suffered from memory leaks, but I would not be
willing to walk out on a limb with that assertion.

Basically, all the issues supposedly making Win NT unstable simply do
not exist on a closed box like this.

With all due respect, that's simply not true. MS has had a number of
issues with their OS, ranging from memory leaks to kernel crashes.
The important question, as it relates to this topic, does the
application in question trigger any of the known problems and/or bugs
with the kernel? Which is why I asserted that the real world
performance should certainly override the list of valid and well
supported concerns. Notice that I am not saying, never buy a device
which has a MS OS in it. I'm simply saying, use caution and let real
world experience be your guide.

Personally, if I learn that a device is running a MS OS, I immediately
consider the device to be suspect until proven otherwise. That
doesn't mean that the alternative implementations (other devices) will
always be problem free. Just the same, the inclusion of a MS OS in a
device should always be treated as a yellow flag. Which means, use
caution until proven it's no longer needed.


Cheers,

Greg



This all being said, I work with a medical application that runs on
windows and we have had a lot of machines running 4.0 and our app ( and
nothing else ) that have run for very long periods of time 7x24. I think
our record is 1 year and it did not crash, we rebooted it to load a
newer version of the app. On the other hand I've had a workstation
running NT 4.0 and a slew of other things that crashed on a real regular
basis. It depends on the apps. NT and Dos before it has to support a
slew of wild hardware from a bunch of venders and sometime things went
boom in the night. Macs have had a much better rep for stability because
Apple laid down the law as to what could be done in terms of hardware and
software. Apple may have had a more stable system but Dos/NT/Intel took
over the world.

"C J Campbell" wrote in message ...
All operating systems have a long history of crashing and being less than
stable.

That's too broad a statement. There have been very stable embedded
OS's meant for applications like this, with no history of crashing,
and that are certified for use in life-critical situations. I believe
those are what others are saying they'd rather rely on.

It would be interesting to know why you think Win NT would be
unstable on something like the MX-20. The device is dedicated to running one
program. It has no peripherals. It never runs for more than a few hours.
Basically, all the issues supposedly making Win NT unstable simply do not
exist on a closed box like this.

For something fairly simple like the MX-20, if the application is
proven stable, then I agree. One assumes that latency effects are
mitigated by using a faster CPU for example And locking pages in
memory, etc.

However, Microsoft themselves point out that Windows is not a hard
realtime OS, and should not be used in more demanding applications
such as fly-by-wire.

Best, Kev

Kevin Darling ) wrote:
:
: However, Microsoft themselves point out that Windows is not a hard
: realtime OS, and should not be used in more demanding applications
: such as fly-by-wire.
:
http://www.gcn.com/archives/gcn/1998/july13/cov2.htm
Software glitches leave Navy Smart Ship dead in the water |
GCN July 13, 1998

http://www.gcn.com/vol19_no27/dod/2868-1.html
Navy carrier to run Win 2000

"...Lockheed Martin officials chose Microsoft in part because of the
company's "experience in computers, networks and systems," Lockwood
said. "We felt that Microsoft had a lot of insight" that could help
Lockheed Martin stay current with commercial technology, he said.

"This is a new area for us," said Keith Hodson, a Microsoft Government
spokesman. "Windows-based products have not traditionally been
associated with Defense Department-specific mission-critical
applications..."


http://www.cnn.com/2000/TECH/computi...r.windows.idg/
CNN.com - Technology -
Futuristic Windows version to control aircraft carrier - August 8, 2000

"...The CVN-77 win is a key triumph for Microsoft in the defense
industry, because it sets the stage for the company's participation in
the Navy's long-term, three-phase future carrier design program. "This
is not just the one ship. It will decide the architectures for the
next three ships," Roach said. Microsoft's agreement also includes a
back-fit program for seven other carriers, bringing the total to 10."

At least some PLCs are also used, per this description of the "Smart Ship"
system...

http://www.e-d-i.com/products_control.html
L-3 Communications SPD Technologies - Control Systems

"...The original MPCMS was physically removed from the ship and
replaced by 15 COTS Programmable Logic Controllers (PLCs) and 12
Windows NT-based workstations. The workstations and PLCs are connected
via Ethernet to a fiber optic Local Area Network (LAN). The LAN
consists of five Automatic Transfer Mode (ATM) switches configured in
a 155 Mbps full mesh, backbone. To enhance reliability, the
workstations and PLCs have redundant Ethernet connections to two
different ATM switches..."


--Jerry Leslie
Note: is invalid for email

On Thu, 24 Jun 2004 23:16:44 +0000, leslie wrote:

Kevin Darling ) wrote:
:
: However, Microsoft themselves point out that Windows is not a hard
: realtime OS, and should not be used in more demanding applications
: such as fly-by-wire.
:
http://www.gcn.com/archives/gcn/1998/july13/cov2.htm
Software glitches leave Navy Smart Ship dead in the water |
GCN July 13, 1998


That's actually an application bug and not an OS bug. One interesting
note is that they also highlight that Unix would be a much more reliable
option, which would be true, excluding SCO, and including Linux.

It's also worth noting, that traditionally, all MS OS's have somehow
managed to sidestep the DoD qualification phases. Some cash and palms are
usually suspected to be the reasons. In fact, while I don't have a link
off hand, there is a fairly well known quote, by a DoD (IIRC) guy, which
can be paraphrase as, "If Windows had been forced to go through the same
channels as every other OS, it would of never qualified."

Cheers!

On Thu, 24 Jun 2004 07:44:49 -0700, C J Campbell wrote:

It has never failed. I have, however, seen the OS crash on Garmin handheld
GPS units. Frequently. To be honest, I would prefer the more stable Windows
OS.

A worthwhile question, which is only going to help to obfuscate the
issue, but what makes you so sure it is the OS which crashed and not the
application? For the end-user, granted, there isn't much of a
distinction. Just the same, technically speaking, there is a huge
difference.


Cheers!

On Thu, 24 Jun 2004 20:02:34 +0000, John Theune wrote:


This all being said, I work with a medical application that runs on
windows and we have had a lot of machines running 4.0 and our app ( and
nothing else ) that have run for very long periods of time 7x24. I think
our record is 1 year and it did not crash, we rebooted it to load a
newer version of the app. On the other hand I've had a workstation


This is actually a gray area. Having a long update is generally not a
problem for just about any OS. The problems occur when the system is
under load for extended periods of time. While uptimes of a year for NT
systems are not unheard of, they are a generally considered a minority.
In fact, most companies with serious business applications running on a MS
platform, generally make rebooting the systems part of their required
maintenance. Most companies attempt a reboot ranging from weekly to
bimonthly schedules. The reasons generally range from stability, latency
issues, memory leaks, locked resources (because of the predominately
threaded environment), and overall degraded performance (generally
associated with memory fragmentation).

For many commercial Unix systems, uptime is commonly measured in years
(multiples). For many highend IBM systems, stories of machines running
for 15-20+ years, non-stop, are not unheard of. And then, those systems
lost power only because they were decommissioned.

If you don't mind me asking, what does your medical application do?

boom in the night. Macs have had a much better rep for stability
because Apple laid down the law as to what could be done in terms of
hardware and software. Apple may have had a more stable system but
Dos/NT/Intel took over the world.

Keep in mind that early macs did not have MMUs (Memory Management Units),
which is what provided protection for one process against another.
Likewise, it's also what prevents OS corruption from applications bugs.

Just rambling on I guess...

Cheers.