Date: Sat, 25 May 2002 08:10:23 -0700
Mr. Rod Welch
rowelch@attglobal.net
The Welch Company
440 Davis Court #1602
San Francisco, CA 94111 2496
..
Subject:
|
Boeing Case Study, SDS Requirements
|
Rod,
Comments interspersed below [in relation to excerpts from your letter on May
22, 2002 2051... ]
..
[
Rod Welch
wrote on May 22, 2002 2051... ]
Glad you are thinking about SRS for SDS. Not sure, but I think I have
the code Morris has created for SDS. The past few years, when Morris
has worked on Medit, he uses code on my computer, because it is
important for me to maintain it to support SDS, and this has become
less of a priority for Morris, because he does not use the code for
anything else. On the other hand, if there is other stuff you need
from Morris, it is mostly a matter of making a request. He has always
been fairly accommodating in providing this kind of support.
..
Having said that, I am at a bit of a loss to understand why the
assembly code for Medit is needed to get started on SRS for SDS, since
all of the functionality is created above the underlying code it would
seem there is a level of product description that says what are the
pieces, what do that do, how are they used and how will this improve
upon what is being used now, or something like that. Then at a lower
level, there is an engineering description of how to write code that
implements functionality and interface. Since the code that is
contemplated to be created is not assembly code, it just seems that
the original code is not terribly instructive. So just curious on
this matter.
..
Sorry, the two statements are not causally linked, they are just on
related topics. My discussions with Morris focused on the what is needed to get
SDS moved to a Windows program that doesn't have the memory constraints of the
current program. That requires working from the assembly code since the C
editor was not intended to be a functional clone of Medit, and the only real
documentation is the code.
..
You are correct that developing requirements for SDS uses different
information. To tackle it for real would require bringing up SDS so that all
of its features are visible and can be investigated. The macro code would
likely be helpful as well as the only sure way of finding all the features
would be to make sure that all of the macros have been covered.
At the same time, this would complete the documentation of the macros, much
of which I know is in the record so that it might not be necessary to go
through the code for them.
..
[
Rod Welch
wrote on May 22, 2002 2051... ]
Your feedback on FAR is helpful. You might note that none of the
links to the SDS review of your report on 020315 are active, nor is
there access to your report pursuant to the request for
confidentiality. On 020217 you noted that we need to give people an
example of problems that SDS solves. Similarly, Stuart Harrow at
DCMA, the agency for implementing FAR on projects like ISS wrote on
010608 that SDS is needed to improve management....
..
http://www.welchco.com/sd/08/00101/02/01/06/08/111431.HTM#RS4J
..
I have to agree with Stuart that technology alone cannot produce good
management. Given a desire to produce good management, however, technology
can most certainly help.
I am currently reading some work of Stafford Beer who developed a cybernetic
model of Viable Systems, and what sorts of effort are required by management
actually to manage the enterprise actually to manage the enterprise http://www.staffordbeer.com/ . I don't
recommend you dig into it as it is hairy stuff. I am just starting, but it
looks as though it may shed some light on how we are able to manage so
poorly from a perspective of the way systems actually work.
..
[
Rod Welch
wrote on May 22, 2002 2051... ]
Later Stuart advised that we need intellectual bridges to help people
grasp the need for SDS....
..
http://www.welchco.com/sd/08/00101/02/01/07/30/054920.HTM#LC5O
..
Bridges are certainly necessary. I think that an organization has to be in
pretty good shape to be able to make good use of tools. The more
comprehensive the tool, the more difficulty an organization in trouble has
with it. As I have said, until the management and the culture achieve the
determination to deliver the best product possible and to continue to align
with its own stated goals, a tool such as SDS which could help them do that
is simply not anything they can deal with.
..
[
Rod Welch
wrote on May 22, 2002 2051... ]
The case study on Boeing seems to fill this requirement, although the
ability of people to look at the face of disaster and shrug is
amazingly limitless, so long as other people are paying the bill.
..
This continues to amaze me. The deeper I dig in terms of analyzing the
mismatch between the documents that are supposed to describe what is to be
built and what they actually built (before writing the requirements), the
worse the situation looks. There are so many broken connections in the small
portion that I am able to analyze, that I shudder to think what the code
must be like.
..
[
Rod Welch
wrote on May 22, 2002 2051... ]
In any case, not pushing this too hard, but would like to explore with
Stuart whether there might be a constructive way to use the ISS
example to help him get some traction with SDS at his end, in a way
that does not have negative repercussions on anyone. So far none of
your remarks on 020315 nor today are in the public record. I think
the treatment in SDS of this information illustrates a general problem
without criticizing Boeing, since Boeing is in the mainstream of what
everybody else is doing. However, subject to your release this
remains confidential.
..
I don't have any problem wit the facts of what I have said being made
available, particularly to someone who might be interested and able to do
something about improving the situation, but there are cases where the
wording may be more harsh than I would have allowed myself in a public
report. The facts, however, are as I have stated them.
..
I don't think it is
possible to overstate the magnitude of the mess that the ISS requirements
are in. I think that the requirements and then the code need to be redone
from the beginning; that this is the prototype that should be thrown away. I
realize that this is unacceptable for a product that has cost the millions
of dollars that this one has, but that doesn't alter my opinion.
..
[
Rod Welch
wrote on May 22, 2002 2051... ]
Please let me know what you think when time permits.
..
Thanks for thinking about this.
Rod
Sincerely,
Garold L. Johnson
dynalt@dynalt.com