Thursday, December 20, 2007

I have two Appraisal Team questions

1. Exactly what can & cannot be kept following an appraisal in the way of written notes, excel files etc that where generated by the ATMs during the appraisal?

2. Can ATM team members use laptops during the final appraisal in document review, during interviews etc.?

Take only memories, leave only footprints . . . . oh, that's the National Forest!

Any SCAMPI Appraisal generates a large amount of written documentation including interview questions, answers, practice characterizations, findings, and observations. The Appraisal Plan, and subsequent entries in the SEI's SAS system will identify the formal appraisal outputs that, at a minimum, include Final Findings, Appraisal Plan, and the Appraisal Disclosure Statement. Any notes or written documents that identify the source of the information, for example an answer to an interview question, should be destroyed in order to maintain the "non-attribution" policy that exists within the method. NA is a critical component, as a violation cast doubt on the reliability of the responses to the questions. This includes spreadsheets and files that identify which project team information was derived from. I usually recommend destroying everything that is not part of the formal appraisal record.

I also usually ask the Appraisal Team Members to not use laptops during interviews. I have found that this is odistracting to the interviewees, and often makes them more nervous than they already are. I have also seen ATM's surfing the web and playing with email when they should be listening and taking notes.

Call me old-fashioned, but a yellow pad and pen is the best tool for this job.

Is their a set of standard questions for CMMI SCAMPI Appraisals?

Is there an “official” list of questions that appraisers use when they are examining an org’s processes to determine their alignment with a given level of the CMMI model? Or, do the appraisers build their own questions based on process characteristics associated with each of the five levels?

There is no “official” question bank for CMMI SCAMPI Appraisals. Client’s often ask this and reason that such a question bank would help to standardize the appraisal method, however the specific context of each appraisal (organization, industry, product types, culture, et al) make it virtually impossible to make standardization effective. It’s far more informative to design the questions around the organization than to try to get them to answer unfamiliar questions.

Here at Broadsword we have a database of several hundred potential questions worded in many different ways that are mapped to the practices in the CMMI. Some questions map to multiple practices or even multiple Process Areas. We select the appropriate question set after spending some time getting to know our client and then modify the high-context words to better suit the language of their culture. This seems to produce the best results.

Tuesday, December 11, 2007

Are these work products good evidence for CM and TS?

Could you please answer the following questions about CMMI:

1- Are configuration action proposal and configuration improvement proposal direct artifacts if I try to implement the Specific Practice SP1.1 of the Process Area CM (Configuration Management)?

2- Are examples of software process architecture include: object-oriented model, unified model, and waterfall model?

CM.SP1.1 asks us to "Identify Configuration Items" and is part of Generic Goal 1 "Establish Baselines." I'm not sure what a "Configuration Action Proposal" or a "Configuration Improvement Proposal" are, but they sound like "Organizational" work products, and not project-level work products. SP1.1 is looking, literally, for a list of work products that we think are going to be part of our project. Examples include: product specs, requirements, process diagrams, drawings, test cases, test results . . . ."

Software Process Architecture, if I understand your question, defines the process framework that will be used to develop software. OOD, RUP, and Waterfall are all examples of PART of that (Life cycle models) but probably don't include a lot of items that should be part of the process (customer commitment to requirements, PPQA-type audits, peer reviews . . . ). So RUP is a subset of the required architecture, but not the entire process architecture.

Why are GG4 and GG5 not applicable for the Staged Representation and Maturity Levels?

I have a question regarding Generic Goals and Practices in the CMMI. In the Staged Representation GG 1 to 3 are Applicable even though the Maturity Level is 5. But in Continuous Representation GG1 to GG5 are applicable to Capability Levels 1 to 5. Why is it like this?

Fortunately, your not the first person to express confusion about this one. The answer is simpler than you might think.

First, a small correction. GG1 is not required for the Staged Representation. Staged appraisals start with GG2. You are correct that GG4/5 do NOT apply to the Staged Representation.

When you approach the CMMI using the continuous representation (one process area at a time is appraised) you need all of the Generic Goals (GG1-GG5) because you are not being appraised on the interdependent aspects of the other Process Areas. In other words, if you being appraised just on Configuration Management (CM), we want to know if you planning it, assigning resources, etc for for Level 2, if your Institutionalizing a defined process (for Level 3), if your managing it quantitatively (for Level 4) and if your optimizing it continuously (Level 5). If you take the Staged approach (a set of PA's at each appraisals) those things in GG4 and GG5 are ALREADY present as process areas (QPM, OPP, CAR, and OID). Also, in ML3, OPD, OPF, and OT help support what is in GG3.

So, given all this, you PROBABLY won't go to Capability Level 4 in OPP and QPM, or Capability Level 5 in OID and CAR because it would be redundant to GG4 and GG5.

Hope this helps!