Beagle ... alas

On Sat, 27 Dec 2003 23:50:05 -0500, Kevin Willoughby
wrote:

Personally, I'm a big fan of a reviewer submitting his comments to the
review team by email several days in advance of the face-to-face
meeting. Most review comments are trivial. Some are as simple as
pointing out a grammatical or typographical error. Any reviewer who
wastes the time of a review team by pointing out these issues in
committee needs to learn more about how to review.

And the review chair needs to learn more about how to run a review. I
used to threaten grammarians with physical injury (usually withholding
doughnuts) when I was a chair. Only if meaning is obscured are such
issues worth discussing. Otherwise, mark your copy, because you'll be
giving it to the author after the review.

Mary

--
Mary Shafer Retired aerospace research engineer

Kevin Willoughby wrote:
In article ,
says...

FBC seems almost inevitable so long as Moore's Law continues to hold.


?? A spacecraft is a lot more than just some chips.



Prices drop as technology becomes adopted. Generally, the lead
time and general conservative engineering in space projects tends
to make aerospace late adopters.



Every year electronics become more powerful, less expensive, less
massive, less power hungry and take up less volume.


One corollary of Moore's Law that is well known within the software
industry: as the capabilities of computer hardware increase, the
difficulty of writing the software to exploite these capabilties
increases exponentially. A more powerful computers can be a liability as
it encourages a much more complex system with many more points of
failure.

Wierdly enough, you generally don't *need* all the capabilities
of the new chip. The important part is the pricing. The core technical
requirements have all stayed pretty much the same. For example,
the Shuttle has an 18ms response time to send a shutdown command
to the main engines. The computer for this is a multimillion dollar
machine. The work could easily be done by an embedded 386 CPU now,
which I believe is priced in the $10-$20 range. A quick poke around
and I am actually having a hard time finding just the CPU available
anywhere. The standard purchase unit seems to be a PC/104 with
all the trimmings.


Software requires that they exploit the new advances to make up for
the usually kludge nature of the software. When you are aiming at
a specific physical interaction with specific timing, then advances
in added features means little. You are shooting to meet an unmoving
target. Your primary concern is cost and maintaining the same
capabilities. It's not rocket science. If you can order it from
a catalog, it's engineering. (I will concede Error Correction Code
as a very needed improvement, especially for the poor *******s who
bought Sun hardware).

Vincent D. DeSimone wrote:
Yes, FBC does not work. Spend the proper amount of money on missions
and TEST, TEST, TEST. When you are done testing, then test some more.
Then test a little more, and then test some more for the hell of it.

FBC works just fine. There is no contradiction between FBC and adequate
testing.

(And it's not like slower/worse/costlier has a conspicuously better track
record, especially at Mars...)


I agree with your response, but there have been too many examples brought up
in this newsgroup, as well as the news feeds, that FBC is just plain flawed.

There have been examples, yes. And, in many cases, a closer
examination will show a certain consistancy to the problems
that are the sort of things that can, and have, happened in
any project of any size.

My belief that the opinion voiced earlier this year that you can get get two
of these options by only sacrificing the third, is the way to go. It was
called "FBC: Pick 2". I like to rhyme it by saying "FBC: 2 Out Of 3".


There isn't anything endemic about FBS that separates
this from any other engineering project.

Correction. I seem to recall reading a Dilbert cartoon where they
are interviewing a prospective employee. Dilbert mentions that they
try to "empower their employees". The prospective employee asked why
they would have a slogan for that. If that had it, they would not
have to preach it. The same applies here. If you have to create a
slogan for lower costs, you have the wrong corporate philosophy to
actually achieve lower costs.

The errors and flaws you are pointing out all occurred in what
is the second generation of vehicles under this philosophy. They
are to be expected. No one expected anything from the first wave
of projects. So, they have some fairly fixed goals and targets.
Pathfinder had a budget of about $200 million for the lander and
rover and that includes also the launch vehicle. The follow-up
mission Mars '98 came in at about $190 million and it included
an orbitor, lander, and about 3 times as many science packages.

Someone went overboard with trying to cram items into the second
generation vehicles. That is not uncommon. It is a generally accepted
fact in the building industry that the second house that someone
builds is their worst house as they try to do too much for what they
have. The first house is always conservative. It's the second when
they think they know what they are doing that they go overboard.
After that, it is just a process of trimming back and building
appropriately for the job. By spending all their money on the
specfic hardware, they pushed the envelope on having an adequate
budget for training, command and control, and testing. The ground
crew was not as familiar with the mission as they should have been,
they did not follow up on errors, and the testing completely missed
the inclusion of spurious signals from the landing sensors.

When you get down to it, Mars '98 had roughly the same elements
as the Viking mission in terms of complexity of the interaction
of the parts and number of major articles.

snip
You can always let time and technology advance until the
mission becomes more feasible under your 2 goals. That's why money spent of
basic R & D is not "wasted".

Finally, schedule monthly reviews to ensure that the project is not
"wandering" away from the two goals that you have chosen. If it does, don't
be afraid to acknowledge it and make hard decisions to bring it back in
line, or even kill it. But NEVER sacrifice testing. Giving up testing to
balance a budget is a false savings. A lost mission is nothing less than a
100% waste of total allocated time (F), manpower (B), and funds (C).


The problem with a lot of the failed missions is that they held to
the standards without the means to actually accomplish those goals.
They were FBC in name only and were budgetted outside the range of
what they were aiming for. Until they get a much better cost/return
decision matrix, they will keep trying to add to much for the budget.
The core of FBC is a fixed low budget.

On 28 Dec 2003 13:03:54 GMT, rk
wrote:

Mary Shafer wrote:

And the review chair needs to learn more about how to run a
review. I used to threaten grammarians with physical injury
(usually withholding doughnuts) when I was a chair. Only if
meaning is obscured are such issues worth discussing.
Otherwise, mark your copy, because you'll be giving it to the
author after the review.

Surprised you didn't get in trouble with the union. Withholding
donuts is definitely cause for a ULP.

Dryden didn't have one for the engineers (maybe still don't).
Besides, no shop steward would dream of messing with me. I'd stop
bringing in homemade goodies, like sticky rolls, if they did. Being a
good baker confers a certain immunity.

Kidding aside, there's nothing worse than three engineers arguing
about whether the subject and predicate numbers match. The classic
example is "one of these is/are", which is easy in this stripped form
but gets harder as clauses and adjectives are added. I was in a
review that must have spent a quarter of an hour on one of those once.

And then there's the use of "criteria" as the singular, when it should
be "criterion". I almost smacked a fellow reviewer with a dictionary
over that one once. Fortunately, the editor restrained me as she
fixed the offender with a basilisk eye and said, "'Criteria' is
plural, 'criterion' is singular, live with it." The guy subsided in
his chair, mumbling discontentedly to himself until I gave him "the
look". (OK, it went more like me saying that the editor would catch
things like that, and her nodding agreement, but the desire, although
fleeting, was there.)

You know, I was a very popular review chair and I think a lot of it
was that I didn't let people dwell on grammar and typographical errors
and other picayune issues. We'd mostly review by the paragraph, not
the sentence, and this really did keep people's minds on meaning, not
mechanics. Not that mechanics aren't important, but that's why we
have manuscripts to mark up and editors to be the final word. And I
truly did want the review committee to help the author improve the
document, which set the right tone from the beginning. Collegial, not
adversarial.

Mary

--
Mary Shafer Retired aerospace research engineer

On 28 Dec 2003 13:03:54 GMT, rk
wrote:

Mary Shafer wrote:

And the review chair needs to learn more about how to run a
review. I used to threaten grammarians with physical injury
(usually withholding doughnuts) when I was a chair. Only if
meaning is obscured are such issues worth discussing.
Otherwise, mark your copy, because you'll be giving it to the
author after the review.

Surprised you didn't get in trouble with the union. Withholding
donuts is definitely cause for a ULP.

Dryden didn't have one for the engineers (maybe still don't).
Besides, no shop steward would dream of messing with me. I'd stop
bringing in homemade goodies, like sticky rolls, if they did. Being a
good baker confers a certain immunity.

Kidding aside, there's nothing worse than three engineers arguing
about whether the subject and predicate numbers match. The classic
example is "one of these is/are", which is easy in this stripped form
but gets harder as clauses and adjectives are added. I was in a
review that must have spent a quarter of an hour on one of those once.

And then there's the use of "criteria" as the singular, when it should
be "criterion". I almost smacked a fellow reviewer with a dictionary
over that one once. Fortunately, the editor restrained me as she
fixed the offender with a basilisk eye and said, "'Criteria' is
plural, 'criterion' is singular, live with it." The guy subsided in
his chair, mumbling discontentedly to himself until I gave him "the
look". (OK, it went more like me saying that the editor would catch
things like that, and her nodding agreement, but the desire, although
fleeting, was there.)

You know, I was a very popular review chair and I think a lot of it
was that I didn't let people dwell on grammar and typographical errors
and other picayune issues. We'd mostly review by the paragraph, not
the sentence, and this really did keep people's minds on meaning, not
mechanics. Not that mechanics aren't important, but that's why we
have manuscripts to mark up and editors to be the final word. And I
truly did want the review committee to help the author improve the
document, which set the right tone from the beginning. Collegial, not
adversarial.

Mary

--
Mary Shafer Retired aerospace research engineer

On Sun, 28 Dec 2003 06:35:04 -0700, Charles Buckley
wrote:

The work could easily be done by an embedded 386 CPU now,
which I believe is priced in the $10-$20 range. A quick poke around
and I am actually having a hard time finding just the CPU available
anywhere. The standard purchase unit seems to be a PC/104 with
all the trimmings

....I have a 386 sitting in my desk drawer, still in the antistat
casing. If NASA wants it for a Shuttle upgrade, I'll gladly donate it
and skip the tax writeoff.

OM

--

"No ******* ever won a war by dying for | http://www.io.com/~o_m
his country. He won it by making the other | Sergeant-At-Arms
poor dumb ******* die for his country." | Human O-Ring Society

- General George S. Patton, Jr

On Sun, 28 Dec 2003 06:35:04 -0700, Charles Buckley
wrote:

The work could easily be done by an embedded 386 CPU now,
which I believe is priced in the $10-$20 range. A quick poke around
and I am actually having a hard time finding just the CPU available
anywhere. The standard purchase unit seems to be a PC/104 with
all the trimmings

....I have a 386 sitting in my desk drawer, still in the antistat
casing. If NASA wants it for a Shuttle upgrade, I'll gladly donate it
and skip the tax writeoff.

OM

--

"No ******* ever won a war by dying for | http://www.io.com/~o_m
his country. He won it by making the other | Sergeant-At-Arms
poor dumb ******* die for his country." | Human O-Ring Society

- General George S. Patton, Jr

On 28 Dec 2003 13:03:54 GMT, rk
wrote:

Surprised you didn't get in trouble with the union. Withholding
donuts is definitely cause for a ULP.

....Yeah, but here in Texas, as long as it's not construction, the
general policy is "**** them mafia union scumbags!"

OM

--

"No ******* ever won a war by dying for | http://www.io.com/~o_m
his country. He won it by making the other | Sergeant-At-Arms
poor dumb ******* die for his country." | Human O-Ring Society

- General George S. Patton, Jr

On 28 Dec 2003 13:03:54 GMT, rk
wrote:

Surprised you didn't get in trouble with the union. Withholding
donuts is definitely cause for a ULP.

....Yeah, but here in Texas, as long as it's not construction, the
general policy is "**** them mafia union scumbags!"

OM

--

"No ******* ever won a war by dying for | http://www.io.com/~o_m
his country. He won it by making the other | Sergeant-At-Arms
poor dumb ******* die for his country." | Human O-Ring Society

- General George S. Patton, Jr

OM wrote:
On Sun, 28 Dec 2003 06:35:04 -0700, Charles Buckley
wrote:


The work could easily be done by an embedded 386 CPU now,
which I believe is priced in the $10-$20 range. A quick poke around
and I am actually having a hard time finding just the CPU available
anywhere. The standard purchase unit seems to be a PC/104 with
all the trimmings


...I have a 386 sitting in my desk drawer, still in the antistat
casing. If NASA wants it for a Shuttle upgrade, I'll gladly donate it
and skip the tax writeoff.

OM


Well, the main problem with a 386 and Intel in general
was the number of IRQ's, IIRC. You'd have to design the
system around those limitations, which means a 386 would
work fine in anything else, but it is incompatible with
the basic design concepts of the Shuttle computer. You
could design something new into orbit using a 386, but
back engineering is basically impossible, especially when there
is no possibility of porting code.