|
Responding to your questions:
1. What are the names of the organizations/individuals within Digital
who have received the SEI assessment training?
Send a note to Meg Dumont at SMURF::MEG. She just hired someone
who is currently the only Digital employee who is certified to
do SEI assessments.
2. Are we using this capability to assess our own software process
maturity, or just selling services?
Yes, we are using it. I work in the SQM/TQM group which is chartered
by David Stone to provide consulting support to software engineering
in the area of quality planning and process assessments. Just last
week we decided to invest our current resources into the SEI model
for assessments. Look for more to come later in the summer and the
fall.
3. What are the results of our internal assessment, for the various
organizations within Digital which produce software?
Again, talk with Meg Dumont. I think her group is the only one which
has been formally assessed. Lots of groups, however, within the
Six Sigma for Software class have reviewed the assessment questions
from the SEI and no group so far has rated above 1. The SEI itself
did a number of assessments and 86% of the companies assessed rated
at 1. Very few companies in the world would even come close to rating
5 which would mean a well-defined, documented, measured, and
continually improving software process.
4. Which organizations within Digital are consciously moving toward
maturity levels 2 or 3? (Yourdon implies that whole organizations
have to make the move or it will not work; however, it seems that
Digital is fragmented enough at this point in time that it might
be possible for some subset of Digital to make progress without
being dragged down by the rest of the company).
The SEI model is definitely aimed at the organizational level. I
don't know about goals for 2 or 3. Perhaps Meg does.
5. I have heard that SEI's relationship with CMU has changed this year.
What impact will this have on the nature and/or quality of SEI's
research?
Again, try Meg.
In general Digital is barely at the point of scratching the surface
on this subject. We have a long way to go. If you are interested,
however, you may also contact Tony Hutchings at TPSYS::HUTCHINGS and
Susan Stefanec at SQM::STEFANEC.
Hope this helps,
Steve
|
|
A professor in an Advanced Design Methodologies class once
said to me that the Japanese Software Factory model
was, in his opinion, capable of achieving higher levels
of process maturity than was the "typical" process employed
in large companies in this country.
I have no idea whether he was correct, nor do I recall specifically
whether he was talking about the SEI model, but I thought his
comments were provocative. I suspect, although I don't know
a lot about the Factory model as used in Japan, that it might
be tough to implement here. My knowledge about it is restricted
to Business-Weekish reading about it.
Comments?
Glenn
|
|
Peter,
Reply .2 mentioned our notesfile, vaxwrk::us_case_program. There is a
group of us in Digital Services who are using the SEI model within our
organizations to bring about Continuous Software Process Improvement.
1. What are the names of the organizations/individuals within Digital
who have received the SEI assessment training?
Organizations: U.S. Digital Services, in particular, the various EICs
such as CIS-EIC, MAG EIC, SEIC, and TN&N.
Individuals: Quite a few -- we are predominently "self-trained." Ron
Madsen, U.S. Digital Services Software Process Improvement Program
Manager, is at the head of the effort. Closer to me is Walson Lee.
Walson works for Ron and manages the EIC's implementation effort.
My focus is the implementation effort with my EIC.
Based upon the SEI's CMM model, Ron designed a workshop, called the
CASE Implementation Workshop, to train people, like myself, to do
SEI-based self-assessments. ["CASE Program" was the program's
original name.]
Note that this is not formal training by SEI. We are not licensed or
approved assessors. Why? It is not our goal to simply 'do' assessments
and hand out formal ratings. Rather, it is to take the self-assessment
results, choose initiatives, and then implement them within the organiza-
tion. The goal is to improve our processes and products. We do our own
follow-up assessments. This determines the level and rate of improvement
and helps in the initiative selection process.
This is more then you'ld get by going through a formal assessment. A
formal assessment provides you with a rating (or score) and a suggested
list of action items or initiatives. From there, it's all up to you until
you have a follow-up assessment.
2. Are we using this capability to assess our own software process
maturity, or just selling services?
Within Digital Services, we are using it to assess our own software
process maturity. Since we are not licensed assessors, we cannot legally
"sell services." We probably have the information, etc., which would
enable us to do if we were licensed.
As your original note suggested, there is a group within DEC licensed
by SEI to perform assessments - but that is not us. From what I've heard,
they are using ability/capability to "sell services."
3. What are the results of our internal assessment, for the various
organizations within Digital which produce software?
Sorry, that infomation is confidential, in my opinion. When assessments
are done, even our own self-assessments, the participants are told that
the results will be kept confidential. To tell you the truth, I'm not
sure if this relates to responses to specific questions or the group's
overall 'score.' Until told otherwise, I'd prefer to keep things the
way they are.
Note that it is dangerous to focus on individual results or scores. If
your only focus is "The Score," then you loose sight of the true goal --
continuous Software Process improvement.
Within our organization, the score is strictly a 'gross' baseline.
4. Which organizations within Digital are consciously moving toward
maturity levels 2 or 3? (Yourdon implies that whole organizations
have to make the move or it will not work; however, it seems that
Digital is fragmented enough at this point in time that it might
be possible for some subset of Digital to make progress without
being dragged down by the rest of the company).
I would like to believe that we are part of that group making that
effort.
Though progress can indeed be made independent of what happens in the
rest of the company, our success is tempered by what the company does
or does not do to improve its processes and, ultimately, its products.
We are "customers." We use the products and tools produced by Digital
to deliver our solutions. The quality of our finished product is
dependent upon the quality of the "material" we use. If its 'bad,'
then it doesn't matter how 'good' our process(es) may be.
5. I have heard that SEI's relationship with CMU has changed this year.
What impact will this have on the nature and/or quality of SEI's
research?
Sorry, I'm of no help to you on this one.
If you want or need any additional information, feel free to contact
me or get into the VAXWRK::US_CASE_PROGRAM notesfile and drop us a
line.
Regards,
Ron Kosmatka, CIS-EIC
|
| Dave SFCPMO::GREENE has requested that I enter the following to clear up some
misconceptions. Please contact Dave directly or see either of the conferences
mentioned in .1 and .2.
Thanks,
Bob - co-moderator DIGITAL
.0>> In chapter 4 of _Decline_&_Fall_of_the_American_Programmer_
A good book. More people should read it just to get a reality check
if nothing else.
.0>> 1. What are the names of the organizations/individuals within Digital
.0>> who have received the SEI assessment training?
There are currently seven (7) individuals in Digital have been
trained by SEI to do SEI Software Process Assessments: myself and six
others. Four are in the US, two in Europe, and one in Japan:
* David Greene USA @CXN (myself)
* Cindy LaFrance USA @STO
* Mark Rabideau USA @DVO
* Kevin Smith USA @STO (moving to ZKO)
* David Christensen UK @RKG
* Sami Zahran UK @UCG
* Shuichi Kuskabe Japan @???
All of these people belong to different organizations within Digital
and have other responsibilities in addition to doing SEI
Assessments. As the assessment business grows (a relatively new
service), these people will devote more of their time to SEI
Assessments. Anyone interested in either selling SEI Assessments to
customers or performing a formal assessment internally should
contact Mark Rabideau @ dtn 553-3484 or via VAXmail COWPOK::RABIDEAU
or Mark Rabideau @ DVO. Mark is chartered with selling assessments
and coordinates scheduling of SEI-trained assessors.
.0>> 2. Are we using this capability to assess our own software process
.0>> maturity, or just selling services?
We do both. Naturally we'd like to sell as many assessments as
possible to outside customers and help make the stock go up. The
cost is reasonably cheap, about $50,000.
.0>> 3. What are the results of our internal assessment, for the various
.0>> organizations within Digital which produce software?
The results of any assessment are highly confidential. If any
assessor violates that confidentiality, the SEI will revoke that
individual's certification. If Digital as a company violates the
confidentiality, Digital will loose its license to perform
assessment. Even the fact that an organization has been assessed is
confidential. If the assessed organization wishes to publicize the
results, or the fact that they have been assessed--that is up to
them.
Note: companies are LICENSED, people are CERTIFIED. People are
transferable. licenses are not. For example, a certified person can
do an SEI Assessment under any of the 8 companies licensed to sell
SEI Assessments. But a certified person can perform an assessment
only through a licensed company. If a company ever looses its
license--SEI states it cannot be reinstated!
.0>> 4. Which organizations within Digital are consciously moving toward
.0>> maturity levels 2 or 3? (Yourdon implies that whole organizations
.0>> have to make the move or it will not work; ...
SEI Assessments are targeted at the organizational level (generally
60+ people). A typical assessment looks at 4-6 software projects
that provide a good cross section of the organization. Assessments
can be done for single projects if they are large enough. But a
single team of 8 people just won't work: partly because of the
techniques employed to do the assessment, partly because the
capability maturity model is designed around large organizations.
That's not to say that the CMM is valueless for small groups. Just
that in small projects, the economies of scale are different.
In my opinion, the Roger Pressman style of assessment if more
applicable to smaller groups.
.0>> 5. I have heard that SEI's relationship with CMU has changed this year.
I don't know about this. The Software Engineering Institute (SEI) is
a Federally Funded Research and Development (FFRD) organization. It
was established in 1984 and is funded by the US Department of Defense
(DARPA) and affiliated with Carnegie Mellon University and is located
Pittsburgh, Pennsylvania.
The SEI was established to address software quality concerns rising
within the US military establishment. These concerns involved
establishing mechanisms to assure that only the highest quality
software is used in all their procured systems. Its charter
includes the directive to expedite software engineering technology
transfer leading to "...rapid improvement of the quality of
operational software..." Though funded by the DoD, its research is
public domain (as its funded by taxpayer dollars) and it gaining
increasing visibility in the commercial/private sector.
.1>> Send a note to Meg Dumont at SMURF::MEG. She just hired someone
who is currently the only Digital employee who is certified to
do SEI assessments.
This is incorrect. See above.
.1>> Yes, we are using it. I work in the SQM/TQM group which is chartered
.1>> by David Stone to provide consulting support to software engineering
.1>> in the area of quality planning and process assessments.
As mentioned above, their research is public domain. You can
call/write to the SEI and they will mail out the requested info
(often, but not always, for free). You do not have to be formally
assessed to use the SEI Capability Maturity Model (CMM). However, it
has been SEI's experience that people often interpret (and a
corresponding questionnaire) the model incorrectly. This is a real
danger. People doing a "self-assessment" need to understand fully
the context/intent by which the CMM was developed. There are a lot
of organizations at the Initial Level (1) that believe themselves to
be at the Defined Level (3) based on their "self-assessment." For
this reason the Government will not accept "self-assessment" scores
[more and more Government procurement agencies are specifying a
formal SEI maturity level when evaluating proposals].
.1>> The SEI itself
.1>> did a number of assessments and 86% of the companies assessed rated
.1>> at 1. Very few companies in the world would even come close to rating
.1>> 5 which would mean a well-defined, documented, measured, and
.1>> continually improving software process.
According to SEI's January 1992 data the breakdown of organizations
is as follows:
1. Initial Level - 81%
2. Repeatable Level - 12%
3. Defined Level - 7%
4. Managed Level - 0%
5. Optimizing Level - 0%
The above data is based on about 300 software projects, and the
data is skewed toward the defense marketplace. Some individual
projects/teams have been found performing at the Managed/Optimizing
Levels--but NO organizations as a whole.
.4>> Reply .2 mentioned our notesfile, vaxwrk::us_case_program. There is a
.4>> group of us in Digital Services who are using the SEI model within our
.4>> organizations to bring about Continuous Software Process Improvement.
.4>> Individuals: Quite a few -- we are predominantly "self-trained." Ron
.4>> Madsen, U.S. Digital Services Software Process Improvement Program
.4>> Manager, is at the head of the effort.
.4>>
.4>> Based upon the SEI's CMM model, Ron designed a workshop, called the
.4>> CASE Implementation Workshop, to train people, like myself, to do
.4>> SEI-based self-assessments.
From what I have seen of this program. Ron has done some really good
work. However, the approach and degree of rigor associated with the
assessment and findings are different than those used for formal SEI
Assessments. I have some concerns about the implementation of this
program (at least for EIS). Due to Digital's fragmented nature, way
EIS does business, and the tendency for huge percentages of a
project's staff to vary from project to project. Given this, it will
be very difficult to achieve higher levels of maturity. Reaching
higher levels of maturity requires a common way of viewing the
development lifecyle at all levels of management and individual
contributors. Empirical data suggests this is extremely difficult to
achieve unless you have a stable core group (80%+) from project to
project. Reaching higher levels of maturity requires more than
having a tool chest and the DPM. Typically the people and management
issues (not technological issues) are the hardest to overcome.
As I am not associated with this program (nor are any of the other
SEI-trained assessors to my knowledge), I am unaware what steps
are being taken to resolve these issues.
.4>> ... It is not our goal to simply 'do' assessments
.4>> and hand out formal ratings. ...
Nor is this the goal of a formal SEI Assessment--the rating is a
by-product of the assessment! The goal of the assessment is to
initiate continuous improvement. The Assessment team uses the CMM as
the yardstick to guide the organization in identifying and
prioritizing findings. These findings are then used to plan an
improvement strategy for the organization.
Though the CMM is not rigid, deviations in prioritization of improvement
initiatives should be thoroughly understood. Historically, the
tendency for most groups that do self-assessements is to throw away
(or delay implementing) the pieces that are not convenient to
implement. These pieces tend to be the most critical, and also the
most difficult to implement. After all, if it was easy, it would
have been implemented already.
Dave
|