Faulty hardware found on shuttle

"Derek Lyons" wrote in message
...
(My father-in-laws XP machine is far less stable than my
Win98 machine, the key difference? He keeps installing crap, never
optimizes his registry, never uses the tools I've tried to teach him
to use etc...)

Optimising the registry is one thing the home user should never have to
bother with. It is such an esoteric, geeky thing to do, for the average
user. The fact that it can be necessary on a home system is not something
that counts in Windows' favour.

"Derek Lyons" wrote in message
...

Early inertial systems weren't considered particularly stable. It was
not uncommon to have *both* systems installed. Civilian navigators
were also helped out the door by the increasing availability and
reliability of LORAN/DECCA/OMEGA etc...

Air NZ flights to Antarctica in the late '70s carried a human navigator, but
this did not save ZK-NZP despite its state-of-the-art INSs, which were found
to be only as safe as their programmers.

John Doe wrote:

What what will it be ? Building an updated model of proven technology with
many improvements, or continue to dab into new esotheric technologies in the
hopes of finding something radically cheaper ?

A lot depends on exactly *which* path to replacing/upgrading the STS
you choose. I have identified 3, in (parentheses) are the names I use
when thinking about them. (These are listed in no particular order).

(Orbiter MK II) - a plug, moldine, and trunion pin compatible orbiter.
Built from derived components or from a clean sheet with the bounds of
existing compatibility requirements and specifications.

(Shuttle MK II) - An entirely new shuttle type vehicle with the same
basic specifications. It may be built from derived or new components,
but overall from a clean sheet and revised specifications as compared
to the current ones.

(STS II) - Something completely different. Most likely a capsule
style crew vehicle and a seperate heavy lifter. (Though there is no
reason cargo cannot be carried in the equivalent of a SLA or in a
MOL-like configuration.) May or may not replace the whole range of
current Shuttle capabilities, most importantly in the areas of tended
assembly and the ability to return hardware to Earth.

D.
--
Touch-twice life. Eat. Drink. Laugh.

John Doe wrote:

Lets look at the ET for the shuttle. I realise they are a technical problem
for Columbia. But from a business point of view, aren't those as close to
"commodity product" as one can get ?

No.

D.
--
Touch-twice life. Eat. Drink. Laugh.

In article ,
(Derek Lyons) wrote:

Hmm... My Windows box hasn't crashed outright in over a week, hasn't
had a program freeze in two or three days, and only needs rebooting
when one particular memory picky game is run.[1] The only people to
whom that is not acceptable are those with a pathological hatred of
Redmond, or an unrealistic standard of performance.

Come on, Derek, you're setting your standards too low. My OS X laptop
hasn't crashed outright three times in 14 months, hasn't had a program
freeze in several weeks and only needs rebooting when I update system
software. My current uptime is about an hour under 5 days - remember,
this is a laptop, by the way, which goes into hibernation at least once
each night and wakes up rarin' to go without needing to reboot. It's
gone as long as about 22 days without restarting. My office Linux box
once went 47 days without a restart, and then it did only because of a
building power glitch. My current XP box has NEVER BSOD'd *except* when
I have driver issues: a USB audio interface I own can kill it in seconds
(go figure . . .) and video drivers used to be the bane of my prior
AMD-based system with Nvidia cards.

However, I can't get 98 to work routinely for nearly the same length of
time. A couple of days to about two weeks, tops, is the most I can get
the office machines to go. Some do much better than others, depending
on the apps they run. Quickbooks Pro, for instance, seems to have a
serious memory leak that Win98 doesn't take care of properly. One of
our accounting machines gets screwy if QB runs all day, while the other
(running XP) hasn't rebooted in several weeks (since the last MS
updates).

--
Herb Schaltegger, B.S., J.D.
Reformed Aerospace Engineer
Columbia Loss FAQ:
http://www.io.com/~o_m/columbia_loss_faq_x.html

LooseChanj wrote:

On or about Mon, 29 Mar 2004 22:08:37 -0500, Peter Stickney
made the sensational claim that:
And MS_Word 2204 will require a semitrailer full of terabyte
SVHDDVDs. run like a Sloth on your MeraHertz Anthill processor, and
provide no more useful functionality than Word 6.

But by then that semitrailer of storage will be smaller than a shoebox.
--


It already is...

In article ,
(Derek Lyons) writes:
(Peter Stickney) wrote:

(Derek Lyons) writes:
Marvin wrote:

Microsoft has a great sales department, second only to their legal staff.
But quality control is waaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaay down on the
priority list.

*Right*. That's why millions of people around the world use it on a
daily basis. It's far from perfect, but it does work, people hold it
to an unrealistic standard and them complain when it fails to meet
that standard.

But nobody used Windows in a situation where Somebody COuld Die or Go
To Jail.

Certainly. But "not reliable enough situations where Somebody Could
Die or Go To Jail" does not mean "not reliable for average day to day
use". I don't hold my $5.99 Kmart cooking knife to the same standards
I hold my $159.90 Wusthof-Trident chef's knife, nor do I reach for
them interchangeably.

One of the biggest problems with Windows-based systems is that you
don't ever really know what's on it. Every bit of code, every
application, every little widget has to insert its own little bit of
cruft into the fundamental operation of the machine. With that many
cooks whizzing into the soup, you aren't going to get Lobster Bisque.

Not a problem when one doesn't install every damm thing in the
universe. (My father-in-laws XP machine is far less stable than my
Win98 machine, the key difference? He keeps installing crap, never
optimizes his registry, never uses the tools I've tried to teach him
to use etc...)

Which, for a Vanilla home system, that won't be doing much, is almost
possible. But - when you start needing to use various, eay
Engineering tools, which require their very own ActiveX Controls, and
display their output through Flash, or whatever other Shiny Stuff
attracted the Programmers eye, it goes downhill real fast. (Then you
add in the people who Just Have to Have the wonky screensavers, or the
Joke of the Day Calendar, and, and...) I'll admit I'm a bit Testy on
the subject. I've spent a couple of days trying to delouse one of my
Carographic Servers. It was the initial server in a project that I
took over. (Win 2K Server, Autodesk Apps. I'm about ready to call
out the Peasants with their Pitchforks and Torches) Somehow, the
software got installed twice, and wedged itself in a way that it won't
uninstall in anything resembling a coherant manner.

(Which leads to another big gripe about MS-Window-oid code - the UI is
shotgunned throughout the OS Kernal, and through all of the app &
middleware. This makes layering, and optimizing the function of any
particular bit of code an almost unworkable nightmare, for any sort of
serious project.)

That depends on your definition of 'serious'. The vertical app my
wife used at her old accounting office was certainly serious, to
them, the IRS, and their customers. And it worked just fine.

Well, to me, serious is a data acquisition system (distributed among
several host systems, capable of running more than 5,000 test
stations, each one taking 4 types of measurement at sub-millisencond
intervals, while controlling dynamic electrical loads (Constant
Resistance, Constant Current, or Constant Power, or a mix of all 3) at
microsecond intervals, With better than 2 mV and 1 Ma accuracy. With
the minumum risk of losing data to any cause, for test durations
ranging from 1 second to over 6 months. Oh, yeah - the Test
Stations may be in environments ranging from Ambient to -40C to 50C,
and the test subjects may be in the act of exploding at the time.
You can't do that with Windows.

It works out to 1250 discrete microcontrollers running in-house
developed code, 5 redundant data collection/filter/control systems
(VentureCom Real TIme UNIX SYSVR4), a pretty danged serious Sun
running Oracle, and a set of Windows Apps for teh users to get to adn
process their data. All in all, 4 people, 3 years, including custom
hardware development, and a shade over a Million and a Quarter lines
of code, GUI included. Normal test throughput was over 130,000
cells/year, with a "No-Test" rate of 1%.

One of the big problems with Windows anything is dealing with large
projects. The integration tools are poor, the APIS are as fluid as
Slime Mold, and as a Project Leader, you end up having to spend all your time
being the Coding Standards Enforcement Nazi in order to keep everyone
from stepping on each other's genitalia. In a good system, the tools
are able to handle the integration far, far better.

--
Pete Stickney
A strong conviction that something must be done is the parent of many
bad measures. -- Daniel Webster

In article ,
(Derek Lyons) writes:
(Henry Spencer) wrote:

It did, as did other military aircraft (and even some cruise missiles)
dating back well into the 50s. Both the military star trackers and the
civilian navigators were swept away by the advent of aircraft-sized
inertial navigation systems.

Early inertial systems weren't considered particularly stable. It was
not uncommon to have *both* systems installed. Civilian navigators
were also helped out the door by the increasing availability and
reliability of LORAN/DECCA/OMEGA etc...

For aircraft, there were a lot of Doppler Nav systems, as well. The
same sort of integration as an INS, for the most part, but mostion
sensing was done with a set of 4 radar beams. Change of postion was
measured by the Doppler Shifts of the beams. With a good alignment,
it was rather accurate, but, also rather unstealthy, and susceptible
to weather problems.

--
Pete Stickney
A strong conviction that something must be done is the parent of many
bad measures. -- Daniel Webster

Derek Lyons wrote:
Dave Michelson wrote:

Perhaps you should preface your remark, "In my uninformed opinion, people
hold it to an unrealistic standard...."

If my opinion was uninformed, you'd have a point. But I'll give you a
free clue: "uninformed" != "disagrees with yours".

If the standard is "unrealistic", it couldn't be met by people developing
other desktop operating systems (Linux, Solaris, BSD, etc.) under similar
constraints.

Since they do so handily, one can only conclude that the standard *is*
realistic. QED.

(To myself: Hmmm. I wonder how Derek is going to twist this one around!)

--
Dave Michelson

"LooseChanj" wrote ...
On or about Mon, 29 Mar 2004 22:08:37 -0500, Peter Stickney
made the sensational claim that:
And MS_Word 2204 will require a semitrailer full of terabyte
SVHDDVDs. run like a Sloth on your MeraHertz Anthill processor, and
provide no more useful functionality than Word 6.

But by then that semitrailer of storage will be smaller than a shoebox.
.... for sufficiently large values of 'shoe'.