|LATEST DEVELOPMENTS AT SITARA|
|Q 'n A from Ask SITARA on CMMISM|
SITARA has been asked several pertinent questions and some of them find a place here. Questions posed to SITARA & SITARA's responses to them will be posted from time to time. If you want to submit a question, please use the Ask SITARA Form.
QUESTION: With ref. to the Intro to CMMI - Staged Representation Training that I attended in 2002, kindly confirm if the same training will suffice for the requirements for the ATM for a CMMI Level 5 for Continuous Representation Model.
ANSWER: Technically speaking - yes. You would however need to have an idea of profiling, capability levels and related concepts which you can pick up by reading the continuous rep.
QUESTION: We need clarification on GP2.8. Some of our PAs like DAR do not have any metrics associated with them (as of now). Is it necessary that some metrics should be monitored to satisfy GP 2.8?
ANSWER: GP 2.8 is very specific about incident and problem or issue management. What I mean by incident is, ‘a special cause of variation’ attributed to a source that is ‘outside the scope of the process’ which nevertheless influences or affects the execution of the process (and the process could be a procedure or activity from any of the 25 process areas). Problem or Issue is ‘a common cause variation’ which is attributed to a procedure that is being used ‘routinely’ to perform the defined activities. A common cause variation may be influenced by tailoring. In other words, if tailoring is used on a project then its effects stand a good chance for GP2.8. Otherwise, from organizational process audits we know full well the effectiveness and efficiency of the procedures that form the process – so GP 2.8 may be applicable to a lesser extent and would be an indirect artifact that gets reviewed as a configuration item audit report or as a project monitoring and control process of status reports.
I reproduce below what GP 2.8 says …
The purpose of this generic practice is to perform the direct day-to-day monitoring and controlling of the process. Appropriate visibility into the process is maintained so that appropriate corrective action can be taken when necessary. Monitoring and controlling the process involves measuring appropriate attributes of the process or work products produced by the process.
Keeping the above definition in mind, I would think that a balance between a combination of both metrics oriented decisions (controlling the critical-to-X parameters), and non-metrics based decisions (controlling customer satisfaction, service level agreements, analysis of QA audit reports and the like) has to be established. It is not always necessary to demonstrate use of ‘metrics’ to control something that has produced variation in results – it is in my opinion not always possible to use metrics in software in this way. That is because measurement and metrics collection in software is always ‘intrusive’ and almost always impacts performance. Another way to look at it is, it is difficult to stop a developmental activity just because we see a variation (even though it may be gross). Many times there is no correlation between the control you exercise and make for an out-of-bound point and the consequent behavior of the process. (The ‘next point’ in a control chart is most definitely not influenced by what one did to bring an out-of-bound ‘point’ back into control).
Today, I have a totally different view on control charts for software! I think, there is more meaning in understand the cause of variation of ‘in-bound data’ rather than triggering a causal analysis for any out-of-bound value. Reason? Out-of-bound process value is most definitely a ‘special cause’ and could be a false alarm, or could be something that is just out of the control of the process (e.g., customer change request). So what is the point in analyzing it? I would rather want to know the reasons for why execution outcomes vary in-bound. For this I also have another thought. You need to use distributed control limits across the lifecycle of the project. What I mean is this:
If effort variation is a parameter that is of interest in Project Monitoring and Control, then I would like to see a control limit of let us say +/-40% in Requirements Capture phase, +/-30% in Analysis, +/-20% in Design and +/- 10% in code and least of all if any of variation in testing. I would think that it is more realistic this way because, clarity of the scope of work gets better and better over time. And, in my opinion, having a 40% variation upfront is much desired because we can allow for a better resolution of requirements since it is known to always be hazy in the initial stages … rather than saying it has to be 20% across the lifecycle. And, just to conform to the 20%, individuals will rush through into Analysis and design even before requirements stability can be established.
So, the answer to your question is we need not be unduly ‘metrics’ emphatic. We can take decisions based on ‘professional judgement’ and also exercising the necessary incident management procedures through DAR.
QUESTION: I need your advice on one of the things related to
In CMMi there is a concept of a SCAMPI-B assessment which is like a Pre Assessment / Mock Assessment as called in
Sw-CMM which is done prior to the final CBA-IPI. What I heard from the market is that there could be cases where we do a "formal" SCAMPI-B Assessment, which means the results of this assessment would also be flied with
SEI. Though not written anywhere but SEI recommends a formal SCAMPI-B Assessment in case a client goes in for a higher maturity level assessment.
ANSWER: Class B SCAMPI is required as part of a more rigorous Class A assessment. Yes, it is true.
QUESTION: I wish to know the specific differences in coverage of the CMMI Staged Model and the Continuous model. While I see that you have recommended the staged models for individuals already exposed to s/w CMM, I would still like to know the specific coverage details in the two sessions
QUESTION: After seeing the course content, I realize that the CMMI (continuous) has topics common to CMMI (staged)? Would you recommend exposure to both?
ANSWER: While the process areas are the same on both the Staged and the Continuous representations, the approach that has to be adopted and the rationale for process area classification is different. While the Staged representation focuses on "organizational maturity", the Continuous representation focuses on "process capability" of individual process areas; a very different concept.
Exposure to both representations is most definitely desirable since the continuous representation is NOT an incremental treatment of the subject matter of the CMMI (Staged representation). To initiate further research on this topic, you are encouraged to see the differences between them after downloading the models from www.sei.cmu.edu. In the continuous representation, there are 3 more generic goals to consider and therefore 6 more generic practices that would need institutionalization. Then there is an entire new body of knowledge surrounding process area profiles such as target profile, achievement profile, target staging and equivalent staging to understand.
Our personal opinion: Both models have distinctions that are worth noting.
QUESTION: Should I take the Staged or the Continuous representation?
ANSWER: If you have no prior experience with the CMM models, such as Software CMM, the staged representation of the CMMI is a good model to choose. It is a natural extension of how we think and therefore a lot more easier to comprehend. Besides there is the Software CMM legacy - which is a staged model, to consider!
QUESTION: Does the Introduction to CMMI-SE/SW/IPPD course also cover SCAMPI assessment methodology or if it does not, then do you propose to offer a course on SCAMPI?
ANSWER: The3-Day Introduction to CMMI is a very heavy course and is packed end to end. And the course may extend up to 6:00 PM each day. This is because 24 process areas need to be covered to the necessary degree of detail and depth. So, the SCAMPI methodology itself is out of scope of the course. However you are free to discuss questions regarding the SCAMPI assessment method with the course instructor (Raghav Nandyal) during the breaks.
A separate course on the SCAMPI Method "AS A PUBLIC OFFERING" is not being planned at the moment. It is normally taught as a prerequisite for the "SCAMPI assessment team". And, this requires a site to go IN for a formal SCAMPI assessment. SITARA normally provides a half-day overview at customer location, of the SCAMPI assessment process for Companies pursuing the SCAMPI assessment. It is important to know the differences between the 3 options of Class-A, Class-B and Class-C types of SCAMPI assessments. Suffice it to say here that a Class A is the most rigorous form of a SCAMPI assessment and is not very different from the classic CBA-IPI assessment as done by SITARA - wherein ALL phases of the CBA-IPI are engaged ONLY by an authorized lead assessor. And, this includes the mandatory Introduction to CMM, Assessment Team training and the document review.
QUESTION: Should I be trained on both the representations of the CMMI-SE/SW/IPPD?
ANSWER: It is NOT necessary to be trained on both the representations. The primary process area description remains the same in both representations. However, there are architectural differences that one needs to consider. Knowledge of both representations is most definitely an advantage to engineer robust process solutions. Even within a process area, certain practices are better enabled by "whole process areas". A classic example would be the Process Areas in the Support Category. There are other dependencies which become highlighted in the "continuous representation" that is not so very obvious in the staged representation.
QUESTION: Is there a specific sequence in addressing process areas in the continuous representation?
ANSWER: If you are not used to process improvement initiatives using the CMM, then a good 'first choice' would be to adopt the Staged Representation. By building layers of maturity on the institutionalized process infrastructure one level at a time, it makes for an efficient and an effective process framework in improvement initiative. However, if there is sufficient exposure to the staged concepts, using a continuous representation will enhance effectiveness of process areas by improving their capability levels.
QUESTION: Can you demystify the two representations for a layman?
ANSWER: The continuous representation is like having a "buffet lunch or dinner". You exercise choice over what you drink or eat from the buffet spread. You may choose to have the complete soup to nuts experience! Or, you might decide not to have the soup and go for the main course. And, if you are conscious of your diet you may skip the deserts and the nuts. Exercising this choice requires "expert" knowledge on what the buffet spread is and how to live with the decisions you make. In order for these choices to be effective one needs to have a priori information to an extent on what makes the best choice under the circumstances. Normally, such an a priori information is available when there is significant experience in dealing with situations which have been experienced first hand. When such experience is not available, say when you have to dine at a Thai restaurant serving Crocodile Soup, then before attempting to try the Crocodile Soup, it might be useful to taste "salt water"! A lower degree of risk and therefore a lower target of accomplishment!! In the CMMI context, a lower process area to target. And, then move up the value chain as we begin to "de-risk" the process improvement initiative by building self-sustaining process infrastructure at lower levels of maturity.
QUESTION: Can you describe the two representations of the CMMI-SE/SW/IPPD and point out the salient difference between them?
ANSWER: The continuous representation is composed of process areas belonging to categories such as Process Management, Project Management, Engineering and Support. Whereas, in the staged representation, process areas are clustered into Maturity Levels.
WE HAVE BEEN ASKED THE FOLLOWING ...
QUESTION: Who is a Transition Partner? Is this someone who transitions you from CMM to CMMI?
ANSWER: A transition partner is an organization "selected" by the SEI to represent in SEI's Technology Transition Mission. Qualified organizations are engaged by the SEI to provide authorized "training" such as the Introduction to CMMI-SE/SW/IPPD/SS course and also assessment services such as the SCAMPI method for the CMMI framework. SITARA Technologies is a transition partner with the SEI on BOTH the Introduction to CMMI and SCAMPI assessment services.
The objective of the SEI to use the services of transition partner organizations is to provide additional channels of distribution for CMMI training and assessment services that will expand and increase market penetration while providing the agreement partner with an "in-house" capability to offer services that are both "RECOGNIZED" and "AUTHORIZED" by the Software Engineering Institute.
Transition partner is NOT someone who transitions you from CMM to CMMI!
QUESTION: How are the services provided by a transition partner different from other agencies or organizations who offer similar services?
ANSWER: First and foremost, a transition partner organization has been selected by the SEI after an evaluation of a proposal to represent SEI in its technology transition mission. To this effect, transition partner organizations are "qualified" by the SEI to represent its interests. Besides, the process for becoming authorized by the SEI to either teach the Introduction to CMMI-SE/SW/IPPD/SS or offer SCAMPI assessment services involves a sequential set of approvals and successful completion of pre-requisites. Example: For SITARA Technologies to offer the Introduction to CMMI course on behalf of the SEI, its instructors would have to go through the following steps.
INSTRUCTOR AUTHORIZATION PROCESS:
1. Agreement Partner (such as SITARA Technologies) recommends candidate instructors based on the instructor selection criteria as specified by the SEI.
2. Candidate instructors attend at Agreement Partner's expense an SEI-authorized offering of the "Introduction to CMMI-SE/SW/IPPD course" (Continuous or Staged)
3. Candidate instructors attend at Agreement Partner's expense an SEI-authorized offering of the SEI's Intermediate Concepts of the CMMI course and successfully complete the mandatory requirements and pass a closed book examination.
4. SEI confirms candidate instructors (reviews candidate applications, and notifies selected candidates of instructor training dates).
5. Candidate instructors attend at Agreement Partner's expense the Instructor Training for the Introduction to CMMI-SE/SW/IPPD course and successfully complete the mandatory requirements and pass a closed book examination.
6. Each candidate instructor is then observed by an SEI authorized observer while teaching a Agreement Partner-delivered offering of the Introduction to CMMI-SE/SW/IPPD course. Candidate instructors must be observed within 180 days of completing Instructor Training for Introduction to CMMI-SE/SW/IPPD course.
Besides the above, a transition partner organization is bound by "royalty engagements and Copyright License Fees" as established and determined in the "LICENSE AGREEMENT " between SITARA Technologies and the SEI for the CMMI product suite.
Other agencies who are NOT 'Transition Partners' with the SEI will need to use the services of other transition partner organizations such as SITARA Technologies for their "service offerings" to be recognized by the SEI.
If these services such as the Introduction to CMMI-SE/SW/IPPD or the SCAMPI assessment services by agencies or organizations are offered WITHOUT using the services of a transition partner, then such courses and services are as good as NOT recognized by the SEI and will NOT be considered VALID.
QUESTION: How would a professional who is blindly led to believe that the two forms of training (one taught by a transition partner and the other by an agency or organization who is not a transition partner) are equivalent, going to reconcile with the difference?
ANSWER: If you have any doubts or questions about the "validity" of a public offering, please contact the SEI's licensing and contracts department.
Best rule to use: When in doubt, ask the SEI! Otherwise, there is no guarantee you are spending your money in the right place.
(Please visit us back ... More will be added soon!)
Note on Copyrights and Trademarks:
All trademarks and copyrights are acknowledged to belong to their respective owners.
For ease of navigation, use the selection box below ...