IV&V as a tool for dealing with "mistakes" and uncertainty
Ron Graham
I thought it might be timely, considering the recent thread on the
subject of mistakes. I have cross-posted it to sci.space but directed
follow-ups to sci.engr. Suitable for clipping, just in case ;-).
Posted as a public service by RG
Philosophy of Independent Validation and Verification
Lewis Research Center personnel are involved in control
dynamics analysis for several projects from the standpoint
not of primary design, but of independent validation and
verification (IV&V). This section contains a description of
the philosophy of IV&V.
The IV&V effort is a means of ensuring the quality of the
final control [or structure, or software] system's design
and performance. The prime contractor has responsibility
for the delivery of the overall system (whether a spacecraft,
launch vehicle, or space structure), including control system
design. NASA, either through its own personnel or through
another contractor, reviews the work of the prime contractor
as part of IV&V.
The IV&V effort is also a means of assisting the contractor
during the analysis phase, while keeping the contractor
responsible for system performance. This assistance takes
on two forms: (1) helping to ask "what-if" questions that
define the system's performance limits, and (2) helping to
extend analysis, when the contractor is constrained by time,
funding, manpower, or experience. Ideally, either form of
assistance increases the contractor's (and therefore, NASA's)
confidence in the final design.
Informally, the term "validation" refers to a check of the
actual numeric results of a contractor's analysis. The term
"verification" refers to a check of the contractor's
methodology. In practice, the IV&V effort consists of some
combination of both, depending on the system under consideration
and the personnel available.
Exchange of Information
For control dynamics analysis IV&V, the following
information is provided by the contractor to NASA:
o system dynamics (plant model),
o control system model,
o starting assumptions,
o input data (parameter values) and
o contact(s) at the contractor's location;
and NASA supplies the following to the contractor:
o requirements and specifications,
o environmental data,
o system integration information and
o questions/concerns/comments.
Validation
The validation portion of the effort may consist of the IV&V
team taking the math model(s) and control law(s) prepared by
the contractor and actually running simulations on-site,
with the goal of obtaining the same (or very nearly the same)
numeric results as those obtained by the contractor. When
this fails, the NASA-contractor team takes steps to locate the
cause of any discrepancy. When causes are isolated and corrected,
the goal of validation is achieved.
Validation is sometimes nothing more than doing exactly what
the contractor did on a different computer: a trivial case.
At other times, validation can involve a detailed analysis
of many subsystems of both the control system and plant: a
tedious case, but one that gives the team a great deal of
confidence.
Verification
The verification portion of the effort may consist of the
IV&V team developing a mathematical model of the system
under consideration from scratch. The method used by the
contractor in model development may not be necessary, since
extra confidence and insight are often gained from obtaining
the same results using two or more different methods.
Verification is necessary at times when a contractor
considers portions of its model to be proprietary, or when
the physical design of a system is changing frequently
enough, or significantly enough, that a totally independent
model may give the team results more quickly, or with greater
confidence, than the original model.
Extending IV&V
Having convinced themselves that both the contractor's model
and methodology are sound, the IV&V team may extend the
concept into potential problem areas which have not been
examined by the contractor. This will happen when the IV&V
team is actually on the "critical path" - that is, the
contractor must commit to a basic design and move on toward
a known launch, and any design changes must either be minor
or possibly delay the launch. In that case, when concerns
arise, the IV&V team must respond to them.
Here are some examples of extending the concept:
o add features to the model, such as
- nonlinear behavior,
- discrete components or
- dispersions of uncertain parameters; or
o simulate the model under extra conditions, such as
- mechanism failures and partial failures,
- center-of-gravity offsets,
- actuation misalignments or
- previously unexpected disturbances.
These exercises can be used to define the limits of system
stability and performance, and as tools to suggest minor
design changes even late in a program. Some examples of such
changes would be
o gain adjustments,
o addition of baffles to a propellant tank and
o filtering.
References
"Independent Stability and Control Analyses of Orbital, Re-entry,
and Launch Vehicles," SAMSOP 800-4, Department of the Air Force/AFSC,
February 1975.
Deming, W. E. _Out of the Crisis_. Cambridge, MA: MIT Center for
Advanced Engineering Study, 1986.
Ron Graham
>The IV&V effort is a means of ensuring the quality of the
>final control [or structure, or software] system's design
>and performance. The prime contractor has responsibility
>for the delivery of the overall system (whether a spacecraft,
>launch vehicle, or space structure), including control system
>design. NASA, either through its own personnel or through
>another contractor, reviews the work of the prime contractor
>as part of IV&V.
I might point out here that the contractor's response to NASA
taking on such a role is not always positive. And why should
it be? At first glance, IV&V almost looks like NASA is saying
"We don't trust you guys to do the job right. We're going to
watch over your shoulder."
Of course, this is not really the case. Unless the NASA personnel
involved in the effort do not understand the goal of the effort.
What you want to do here is to establish a NASA/prime contractor/
other contractor *team* that works together to put something in
space (or in the air).
If you read Deming's work, listed in the original post's references,
you get the idea that simply checking someone else's work after
the work is done (which is the way many folks apply "quality control")
has nothing whatsoever to do with really adding quality to the final
product. Quality has to be built in at the beginning.
>The IV&V effort is also a means of assisting the contractor
>during the analysis phase, while keeping the contractor
>responsible for system performance. This assistance takes
>on two forms: (1) helping to ask "what-if" questions that
>define the system's performance limits, and (2) helping to
>extend analysis, when the contractor is constrained by time,
>funding, manpower, or experience. Ideally, either form of
>assistance increases the contractor's (and therefore, NASA's)
>confidence in the final design.
One other problem that NASA occasionally runs into with
contractors in the IV&V relationship is that the contractor
considers the design of the [spacecraft, rocket, component,
pick one] to be proprietary. So what information can they
turn over to NASA for inspection during the analysis and
design phases that will get past their own security regulations?
That's not a trivial question. Sometimes, NASA and the
contractor may even reach a temporary "impasse" until together
they can compromise on information exchanges.
Of course, even that compromise is a form of IV&V, IMPO.
You see how the IV&V relationship has the potential for being
a "win-win" situation?
RG
"engineering mystic"
Henry Spencer
>If you read Deming's work, listed in the original post's references,
>you get the idea that simply checking someone else's work after
>the work is done (which is the way many folks apply "quality control")
>has nothing whatsoever to do with really adding quality to the final
As witness the fact that it takes 1.5 million signatures to launch a
shuttle mission, and they *still* manage to do things like leaving support
beams inside the orbiter accidentally...
Max Hunter once commented that if he designed a new launcher, it would be
designed not to need a gantry tower with access platforms, because "if the
platforms are there, then there will be a man with a clipboard standing
on every one".
--
Programming graphics in X is like | Henry Spencer @ U of Toronto Zoology
finding sqrt(pi) using Roman numerals. | he...@zoo.toronto.edu utzoo!henry
Doug McDonald
>>final control [or structure, or software] system's design
>>and performance. The prime contractor has responsibility
>>for the delivery of the overall system (whether a spacecraft,
>>launch vehicle, or space structure), including control system
>>design. NASA, either through its own personnel or through
>>another contractor, reviews the work of the prime contractor
>>as part of IV&V.
>
>taking on such a role is not always positive. And why should
>it be? At first glance, IV&V almost looks like NASA is saying
>"We don't trust you guys to do the job right. We're going to
>watch over your shoulder."
>
>Of course, this is not really the case. Unless the NASA personnel
>involved in the effort do not understand the goal of the effort.
>What you want to do here is to establish a NASA/prime contractor/
>other contractor *team* that works together to put something in
>space (or in the air).
>
The point is that the FINAL USERS of the system are the ones who should be
the final quality controls. In the case of, for example, the Hubble
Space Telescope, that means the astronomers. There should be someone
(or, in the case of a big project like that, several people) among
the final user commumity who watch every step of the design and construction,
looking for possible problems. I do science small enough that I can
do such a thing myself - and often find things to check or change.
I seriously doubt that I would have accepted the method of construction
and verification used on Hubble's main optical system. I know absolutely
that I would have, had I had the authority, stopped the launch of
a space ship with people in it with frozen O-rings. Had I not had the
authority, and no one listened to me, I would have called the three
main TV networks and told them to be on the lookout for disaster.
Doug McDonald
Ron Graham
mcdo...@aries.scs.uiuc.edu (Doug McDonald) writes...
>In article <20SEP199...@ariel.lerc.nasa.gov>
> eca...@ariel.lerc.nasa.gov (Ron Graham) writes:
>>In article <19SEP199...@ariel.lerc.nasa.gov>, I previously wrote...
>>>The IV&V effort is a means of ensuring the quality of the
>>>final control [or structure, or software] system's design
>>>and performance. The prime contractor has responsibility
>>>for the delivery of the overall system (whether a spacecraft,
>>>launch vehicle, or space structure), including control system
>>>design. NASA, either through its own personnel or through
>>>another contractor, reviews the work of the prime contractor
>>>as part of IV&V.
>>What you want to do here is to establish a NASA/prime contractor/
>>other contractor *team* that works together to put something in
>>space (or in the air).
>The point is that the FINAL USERS of the system are the ones who
>should be the final quality controls. In the case of, for example,
>the Hubble Space Telescope, that means the astronomers. There
>should be someone (or, in the case of a big project like that,
>several people) among the final user commumity who watch every
>step of the design and construction, looking for possible problems.
>I do science small enough that I can do such a thing myself - and
>often find things to check or change.
I won't dispute this statement, but I will point out one of the
drawbacks of IV&V from a management standpoint: it costs extra
money up-front to build quality into the analysis and design phases
like that. Deming states in his book that if management commits to
quality at the beginning of a project like that, you may make a
financial investment at that point, and then you must follow it up
by continuing to work toward quality at all steps. The final design
will (theoretically) save us all money because we took this approach,
but we may never know how much, because (as Deming states), there are
factors "unknown and unknowable" involved.
Further, as applies to the end users, it is true that they should be
involved at the design and analysis phases. But you could easily take
your team of astronomers and overwhelm them with engineering. Manage-
ment must take upon itself to decide where the astronomers (or whoever
the users are) can make a contribution without taking them away from
their primary work. Astronomers know what sort of performance they
need, and engineers know the systems and mechanisms that will deliver
it. They have to be able to talk to each other and not get in each
others' way to make the system work.
>I seriously doubt that I would have accepted the method of construction
>and verification used on Hubble's main optical system. I know absolutely
>that I would have, had I had the authority, stopped the launch of
>a space ship with people in it with frozen O-rings. Had I not had the
>authority, and no one listened to me, I would have called the three
>main TV networks and told them to be on the lookout for disaster.
I will not enter into a discussion of whistle-blowing: that's outside
the scope of the thread I was trying to introduce.
I also cannot say what sort of quality measures were used with Hubble
or with the O-rings, since I don't work in those areas or even at those
field centers. But you do bring up examples of cases where simply
examining some finished product would not serve for introducing quality
in design. Deming uses a lot of examples in industry of products that
have a failure rate: you put quality into the system and bring the rate
down. In the cases of the SRM or the HST optics, you just can't tolerate
any failure rate at all. So you have to decide: what points are crucial
for introducing quality? How much do we have to spend? Some of the
ideas Deming proposes can fly in the face of budgetary decisions.
In _Men From Earth_, Buzz Lukens (oops! I mean, "Aldrin" ;-)) reminds us
of the design of the lunar module, which was carried on concurrently
(and competitively) by three organizations, with NASA ultimately deciding
which one goes to the moon at the end of production. You do something
like that in today's dollars (like with Fred's electrical power system,
for instance - that's an activity we have here), and it costs you A TON.
But it gives you a great deal of confidence. Compared to the lunar module
design competition, IV&V for some of Fred's systems, for instance, ends up
looking like a patch: but it is a small step in Deming's direction.
RG
who became an engineer because he thought "free-body diagrams" would
be interesting ;-)
Doug McDonald
In article <21SEP199...@ariel.lerc.nasa.gov> eca...@ariel.lerc.nasa.gov (Ron Graham) writes:
>In article <1991Sep20....@ux1.cso.uiuc.edu>,
> mcdo...@aries.scs.uiuc.edu (Doug McDonald) writes...
[lots of stuff about quality control]
>Further, as applies to the end users, it is true that they should be
>involved at the design and analysis phases. But you could easily take
>your team of astronomers and overwhelm them with engineering.
>Manage-
>ment must take upon itself to decide
NO! This is exactly what I am arguing against.
>where the astronomers (or whoever
>the users are) can make a contribution without taking them away from
>their primary work.
It would seem that their primary work during a construction phase would be
to see that the final product worked!!!
>Astronomers know what sort of performance they
>need, and engineers know the systems and mechanisms that will deliver
>it. They have to be able to talk to each other and not get in each
>others' way to make the system work.
>
That last section is of course quite true.
The END USERS (i.e. in this case the astronomers) should be - in
the final decision - superior to "management". They should have have the
final say in all important decisions - including the unfortunate but
sometimes necessary one of asking Congress for more money.
Doug McDonald
Ron Graham
mcdo...@aries.scs.uiuc.edu (Doug McDonald) writes...
>In article <21SEP199...@ariel.lerc.nasa.gov>
> eca...@ariel.lerc.nasa.gov (Ron Graham) writes:
>>In article <1991Sep20....@ux1.cso.uiuc.edu>,
>> mcdo...@aries.scs.uiuc.edu (Doug McDonald) writes...
>>Further, as applies to the end users, it is true that they should be
>>involved at the design and analysis phases. But you could easily take
>>your team of astronomers and overwhelm them with engineering.
>>Management must take upon itself to decide
>NO! This is exactly what I am arguing against.
Then I guess I have to disagree with your argument. Roles of
management must be (I am not saying "are") to advocate for quality
on behalf of all the workers, and to make the others *able* to do their
jobs. If you oppose this, then exactly what do you think managers
should be doing? Bean-counting?
>>where the astronomers (or whoever
>>the users are) can make a contribution without taking them away from
>>their primary work.
>It would seem that their primary work during a construction phase would be
>to see that the final product worked!!!
Don't over-simplify this. To see that the final product worked means,
"to see that the final product met requirements." This is even more
than simply working. The requirements have to be defined (which is more
than simply pulling them out of the air), they have to be communicated
to the design team in such a way that they can be translated into a
system that makes sense to all, and they have to be verified through
[analysis, demonstration, inspection, similarity, testing - pick any
that are needed]. All these functions fall within the role (IMPO) of
the scientist without forcing the scientist to play the role of engineer
as well.
>The END USERS (i.e. in this case the astronomers) should be - in
>the final decision - superior to "management". They should have have the
>final say in all important decisions - including the unfortunate but
>sometimes necessary one of asking Congress for more money.
If they are superior to management, then management should be protecting
them from being taken away from scientific investigation. This is part
of establishing quality in the design system. Management still has the
job of asking Congress for enough money to get the job done. Everybody -
management, engineers, astronomers, etc. - works together to find out how
much that sum is. I am not saying this is the way it is - I am saying
this is the way it should be. The end users can't (and shouldn't have
to) do everything.
RG
"engineering mystic"
Henry Spencer
>The point is that the FINAL USERS of the system are the ones who should be
>the final quality controls. In the case of, for example, the Hubble
>Space Telescope, that means the astronomers. There should be someone
>(or, in the case of a big project like that, several people) among
>the final user commumity who watch every step of the design and construction,
I think Doug has the right stick but possibly the wrong end here. Charles
Sheffield, in a recent issue of Analog, talked about HST's problems and
made a very interesting observation. A lot of the highly-successful
big trailblazing high-tech projects have had a common element in their
management: one man who was on top of everything. Not necessarily a
final user, mind you, and not necessarily the ultimate boss, but *one*
man who kept aware of *everything*.
The Manhattan Project had a lot of technical people working on a lot
of technical problems, but Robert Oppenheimer was aware of it all,
understood it all, and paid attention to it all, at a level of detail
that amazed his subordinates. The Saturns were the most successful
launchers the US has ever built, partly because Wernher von Braun spent
a decade making sure they were. The Hale telescope is named after the
man who spent his whole life obsessed by all the minutiae of it and
several of its predecessors. In each of these projects, there was
one specific man who was with the project from inception to operation,
was involved in everything, and spent all his time and energy making
sure that every last detail of His Project was done right.
With HST, there was no such person. It was managed by a large assortment
of committees with constantly changing memberships. There was nobody
who identified with the project enough to babysit it from beginning
to end, nobody who was willing to say "this will be my triumph if it
works and my disgrace if it fails".
It's not enough to have the user community watching. There has to be
some specific person, either a user or a developer, who cares enough
about the project to make it the primary focus of his life, year after
year, from inception to completion.
As far as I know, there isn't such a person for AXAF either... and it's
being built by much the same management setup as HST. Cross your fingers.
The AXAF mirrors should be okay -- they're on the lookout for that one
now -- but heaven only knows what *other* problems have been missed.
While we're on this... It occurred to me last night that there's a strong
analogy between the lack of astronomer involvement in HST (and AXAF) and
the lack of astronaut involvement in the Challenger launch decision. This
goes deeper than the surface, because in both cases the alleged victims
were largely responsible for the alleged crime.
The lack of astronaut involvement in shuttle decision-making was not
because Big Bad NASA Management kept them out. It was because John Young,
the boss astronaut, didn't push them in. His predecessor demanded astronaut
involvement, and generally got it. Young made it a low priority, so it
gradually slipped off the astronauts' busy schedules. I don't know all
the details, but I get a strong impression that the HST case paralleled
this. All the astronomers are now enthusiastically pointing the finger
at Big Bad NASA Management, but in fact none of them roared, pounded on
tables, and spent time and effort and political brownie points insisting
that they be involved. Instead they went limp and submitted to being
shuffled off to one side, assigned mostly to placating Congress rather
than keeping an eye on the hardware. If this is not to happen again,
the initiative is going to have to come from them.
Ron Graham
he...@zoo.toronto.edu (Henry Spencer) writes...
>In article <1991Sep20....@ux1.cso.uiuc.edu>
> mcdo...@aries.scs.uiuc.edu (Doug McDonald) writes:
>>The point is that the FINAL USERS of the system are the ones who should be
>>the final quality controls. In the case of, for example, the Hubble
>>Space Telescope, that means the astronomers. There should be someone
>>(or, in the case of a big project like that, several people) among
>>the final user commumity who watch every step of the design and construction,
>>looking for possible problems...
>I think Doug has the right stick but possibly the wrong end here. Charles
>Sheffield, in a recent issue of Analog, talked about HST's problems and
>made a very interesting observation. A lot of the highly-successful
>big trailblazing high-tech projects have had a common element in their
>management: one man who was on top of everything. Not necessarily a
>final user, mind you, and not necessarily the ultimate boss, but *one*
>man who kept aware of *everything*. [examples deleted...]
>While we're on this... It occurred to me last night that there's a strong
>analogy between the lack of astronomer involvement in HST (and AXAF) and
>the lack of astronaut involvement in the Challenger launch decision. This
>goes deeper than the surface, because in both cases the alleged victims
>were largely responsible for the alleged crime.
One piece of evidence that supports this is the greatly-increased role
played by the astronauts in some subsequent decisions (e.g. the demise
of Shuttle/Centaur and the current representation of astronauts in many
technical working groups within Fred). You would just expect these folks
not to want anything like that to happen again, and to do anything possible
to prevent it. Their participation in these areas is an example of what
Deming is looking for in quality improvement.
RG
"engineering mystic"
herr...@iccgcc.decnet.ab.com
> In article <1991Sep20....@ux1.cso.uiuc.edu>,
> mcdo...@aries.scs.uiuc.edu (Doug McDonald) writes...
>
>>should be the final quality controls. In the case of, for example,
>>the Hubble Space Telescope, that means the astronomers. There
>>should be someone (or, in the case of a big project like that,
>>several people) among the final user commumity who watch every
>>step of the design and construction, looking for possible problems.
>>I do science small enough that I can do such a thing myself - and
>>often find things to check or change.
>
> I won't dispute this statement, but I will point out one of the
> drawbacks of IV&V from a management standpoint: it costs extra
> money up-front to build quality into the analysis and design phases
> like that. Deming states in his book that if management commits to
If there had been an amateur mirror grinder with authorization to
wander around the Hubble project, he would have pulled a knife out
of his pocket just to admire the perfection of the figure. Didn't
I read here that that test would have discovered the problem?
The o-ring problem is far less clear. Someone who understood that
they were frozen could have taken some o-ring material and frozen
it and then broken into a high ranking office and performed the
Feynman experiment on the mucky-muck's desk. However, such a career
limiting move is only worth while if it can be reasonably expected
to succeed in changing the wrong-headed policy. Again, it needs an
outsider with imagination and curiosity authorized to wander around
and talk to people and poke into things.
How do you put someone in parallel with the Space Station project
authorized to poke into things and with the standing in the relevant
community that makes it possible to make the appropriate people
understand the folly of the EVA numbers?
Quality assurance people in the Hubble project believed what the
technical people told them. You need an outsider. You need someone
who does not starve after he gets fired.
There have been a few whistle blowers who were crying wolf. If a
bottom level project engineer had prevented Challenger from flying,
he would have been remembered as the engineer who delayed a launch,
not as the hero who prevented a disaster.
dan herrick d...@NCoast.org
Ron Graham
herr...@iccgcc.decnet.ab.com writes...
>Quality assurance people in the Hubble project believed what the
>technical people told them. You need an outsider. You need someone
>who does not starve after he gets fired.
>There have been a few whistle blowers who were crying wolf. If a
>bottom level project engineer had prevented Challenger from flying,
>he would have been remembered as the engineer who delayed a launch,
>not as the hero who prevented a disaster.
This is part of what Deming talks about when he refers to "causes (or
results) unknown and unknowable." If the disaster never occurred, no-one
could know the whistle-blower was a hero. What really adds to the disaster
is that *someone must know* that he/she/they could have prevented it, had
they tried, and they have to live with the knowledge. Of course, if they
didn't know the o-rings would indeed fail under the given conditions, and
they had no way to quickly measure the probability, they might just have
hesitated to make a guess.
Those kinds of guesses are difficult to encourage engineers to make. How
often you see a young engineering student willing to guess *out loud*, in
front of prof or class, what result a certain process might have?
RG
Kelly Carney
>they were frozen could have taken some o-ring material and frozen
>it and then broken into a high ranking office and performed the
>Feynman experiment on the mucky-muck's desk. However, such a career
>limiting move is only worth while if it can be reasonably expected
>to succeed in changing the wrong-headed policy. Again, it needs an
>outsider with imagination and curiosity authorized to wander around
>and talk to people and poke into things.
In Feynman's book he tells "the rest of the story". He was kind of
set up. Someone in NASA passed Rodgers enough info to do the famous
experiment. Rodgers gave the hint to Feynman. (Feynman's book is
a pretty good examination of what happened with Challenger.) The
bottom line is that all the technical info needed to make the proper
decision was known. An outsider with imagination and curiosity would
have made no difference.
>How do you put someone in parallel with the Space Station project
>authorized to poke into things and with the standing in the relevant
>community that makes it possible to make the appropriate people
>understand the folly of the EVA numbers?
The EVA numbers were unveiled by "the insiders". Space Station is
so big and complex, it just took a long time to be considered and
solved.
>Quality assurance people in the Hubble project believed what the
>technical people told them. You need an outsider. You need someone
>who does not starve after he gets fired.
I know of no one at NASA who is worried about being fired or starving.
We all have enough integrity to try to do the right thing.
John Stevenson
(Kelly Carney) writes:
> [stuff deleted]
> I know of no one at NASA who is worried about being fired or starving.
That's a sure recipe for disaster.
John Stevenson
hang...@spf.trw.com
Ron Graham
(John Stevenson) writes...
>In article <10OCT199...@venus.lerc.nasa.gov> lvk...@venus.lerc.nasa.gov
>(Kelly Carney) writes:
>> [stuff deleted]
Even though it was relevant.
>> I know of no one at NASA who is worried about being fired or starving.
>__________________________________^^^^^^^^^^^^^^^^^^^^^^^^^
>That's a sure recipe for disaster.
Sez you. Next time, consider the context.
RG
herr...@iccgcc.decnet.ab.com
> In article <1991Oct8.0...@iccgcc.decnet.ab.com>, herr...@iccgcc.decnet.ab.com writes...
>>
>>technical people told them. You need an outsider. You need someone
>>who does not starve after he gets fired.
>
> I know of no one at NASA who is worried about being fired or starving.
> We all have enough integrity to try to do the right thing.
I'm sorry, neighbor (I'm over in Mentor, on the other side of Cleveland),
I did not express myself well. And my clumsiness was insulting in
a way that was not intended or justified:
Most of the people in NASA cannot go storming into Truly's office
and declare that the shuttle cannot fly today, it is too cold.
There are times when information has to jump four or five levels
without being filtered. An insider does not do such a thing twice
because he is not an insider after the first time.
This is not a criticism of NASA in particular, it is an attribute
of any organization where there are four or five levels in between.
Another topic from the same article I posted:
Someone sent email pointing out that the knife edge test would not
have detected the fault in the Hubble mirror. The correct design is
two complementary mirrors with very strange figures. I'm no better
qualified to judge his accuracy than I was to judge the original
posted assertion about the knife edge test, but maybe the response
to this will create a correct record in sci.space.
dan herrick d...@NCoast.org
THEODORE FABIAN
>In article <10OCT199...@venus.lerc.nasa.gov>, lvk...@venus.lerc.nasa.gov (Kelly Carney) writes:
>> In article <1991Oct8.0...@iccgcc.decnet.ab.com>, herr...@iccgcc.decnet.ab.com writes...
>>>
>>>Quality assurance people in the Hubble project believed what the
>>>technical people told them. You need an outsider. You need someone
>>>who does not starve after he gets fired.
>>
>> I know of no one at NASA who is worried about being fired or starving.
>> We all have enough integrity to try to do the right thing.
>
>
>I'm sorry, neighbor (I'm over in Mentor, on the other side of Cleveland),
>I did not express myself well. And my clumsiness was insulting in
>a way that was not intended or justified:
>
>Most of the people in NASA cannot go storming into Truly's office
>and declare that the shuttle cannot fly today, it is too cold.
>There are times when information has to jump four or five levels
>without being filtered. An insider does not do such a thing twice
>because he is not an insider after the first time.
>
you might be right about most folks at NASA not being able to storm into
Admiral Truly's office... but I think that anyone with email capability
could send him a short (or long) note explaining their concerns...
I'm not going to be foolish enough to post his address, but almost anyone
with some ingenuity could find it if he tried hard enough...
>This is not a criticism of NASA in particular, it is an attribute
>of any organization where there are four or five levels in between.
>
at least in my work group at NASA, we've bridged the gap in communications
between various levels by using electronic mail.. folks are responsible
for reading their mail daily, and if an invitation to a meeting is there,
and they don't show up at the meeting, they have no excuse...
as an example of how the culture has taken hold, I once send a message
with a read receipt on it to 350 people... within four hours I had about
300 read receipts back... the other 50 or so filtered in during the next
48 hours...
>
>dan herrick d...@NCoast.org
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Thanks,
Ted Fabian NASA Lewis Research Center
Cleveland, Ohio 44135
tpf...@tm0006.lerc.nasa.gov
/pn=theodore.fabian/admd=telemail/prmd=lerc/c=us/@x400.msfc.nasa.gov
Disclaimer: My Opinions are My Own, not NASA's.....
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -