Friday, February 20, 2009

Executive Brief Interview with Mike Konrad and Jeff Dalton on CMMI and Agile

CMMI and Agile: Opposites Attract

Nice article on the subject in "Executive Brief."

Sunday, February 15, 2009

Does SAM apply if our customer mandates the supplier?

A company has a customer who supplies software modules that are integrated into the final product. The customer has mandated a 3rd party to supply another software module that is also integrated into the final product.

Is SAM applicable in this case? The company has no contract with either in respect of the supplied modules (other than the contract to deliver the final product). The company has no ability to manage the customer as they would a 3rd party supplier.

Wow, this reminds me of the time I met a judge and asked him "I have this friend . . . !"

It's a really interesting question. We often debate the merit of "alternative practices" amongst the SEI community (read: process nerds) with MOST people insisting that their "alternative implementaion" of SPs are "alternative practices." Usually they are not.  Alternative practices are an accepted, albeit rare, way to satisfy the Goals in CMMI, which I'm sure you know are required.

In this case, I'd say "we've got a live one here maw!"

Most of the practices in SAM SG1 are about selecting and acquiring a service/product from a supplier. Even though you're achieving SG1 your customer situation may very well serve as an alternative for the practices within it. Without knowing all of the details, I'd say there is a good chance you'd be OK here.

It's not so clean for SG2. You do, in fact, accept the product, and I'm sure you make sure that it works correctly, and that it can be integrated into your environment. So you're not completely off the hook for SAM.

Whatever you decide, I suggest you document your situation and when you're ready to engage a Lead Appraiser (hey wait a minute . . . isn't THAT SAM???) you discuss it with him/her to make sure you see eye-to-eye on this. If you don't, get a second opinion.

Friday, February 13, 2009

If we don't use all of the practices on smaller projects, can we still achieve ML2?

We're trying to reach ML2. But we have to deal with both large projects (from 700h to 10,000h) and small projects (<>700h)
When we conduct a SCAMPI A Appraisal we "sample" the projects that an organization is executing.  Even though we may only look at a subset of projects, the expectation is that ALL projects within an organizational unit are performing at ML2 (or 3, 4, or 5).  So that imprecise answer is no, you must be performing all of the practices on all projects.
The more precise answer is that there is room for "alternative" practices.  If you have, say, a one-person project, there should be alternatives to the standard process which is probably designed for teams of programmers.  An example may be practices that require progress reviews.  You may require larger teams to have weekly team meetings that are just for the project team.  In a one-person team, that person is having that same meeting 24x7!  So a practice like that may not be relevant as an event and may not be represented by evidence. 
Now, the GOAL must be achieved - even if an alternative to the practice is performed.  So you just can't do NOTHING, but there is some flexibility in the performing alternates to the practices.
Whether or not your Lead Appraiser understands these alternatives is another story though!

Wednesday, February 4, 2009

We're a ML3 company trying for ML4. But we don't understand GP3.2

I’m EPG practitioner of ML3 company and we would like to achieve ML 4. But, I’m not sure for interpreting GP 3.2 in the process. We are so confused whether every process has a measurement metric included in MA.

Because GP 3.2 specifies “collect work product, measures, measurement results, and improvement information derived from planning and performing the process …”.

Could you give me some advice for this?

The more cynical of my Lead Appraiser colleagues might ask “how did you get to ML3 without knowing this?” But I won’t do that. I can think of MANY reasons why this may have occurred!

GP3.2 exists in the model to codify the need to proactively gather information from the project’s use of your process that could be used to improve performance for future projects. This could be “reuse” data, such as useful templates or examples. It could be useful measures that provide the NEXT project with information that would help them. It also could be “lessons learned” gathered from performing retrospectives, post-mortems, and reviews of what worked and what didn’t.

So GP3.2 doesn’t require a metric for every process area, and the Process Area “MA” certainly doesn’t require this.

However, if you’re planning on moving to ML4 you will need to be measuring the performance of selected sub-processes. And to do this you will want to have some granular metrics about the performance of specific sub-processes. One practice that supports this is GP2.8, my favorite, because it requires you to consider these types of measures WAY back in ML2. These measures should help you identify “process levers” that you can manipulate to change the results of a process.

Sunday, February 1, 2009

Follow AskTheCMMIAppraiser on Twitter

To receive Tweets set your Twitter account to follow "CMMIAppraiser."  You'll get daily tips and commentary about the CMMI - all for free!

If we don't use Process Action Teams (PATs) will we fail an appraisal?

It may seem like a silly question, but it is important for me to understand. I’m not opposed to PATs, but I just need to know if without PATs we will fail the appraisal. We are going for CMMI Appraisal Level 2.

Thanks for your question.  This business just gets curiouser and curiouser (to borrow a word from Winnie the Pooh).

For the benefit of our readers, a Process Action Team, or PAT, is a group of individuals who focus on developing, deploying, and maintaining one or more processes within the organization's set of standard processes.  Here at the Broadsword labs we call them "SIGs" or Special Interest Groups because, well, we do most things differently around here!

Let's separate out your process improvement program and your Appraisal.  An appraisal is nothing more than a user acceptance test to determine if your processes and it's performance have reached ML2 performance.  These are two completely different events.  Your process improvement program is something that you execute in order to develop valuable and useful processes that give you that extra competitive edge, or if your like a lot of my clients, just helps reduce the chaos.  You might use PATs (or SIGs) to help you do that.  You often see SIGs as part of a ML3 effort as a result of executing the OPF and OPD processes.

But you're not doing that.  You're still going for ML2.  While I might recommend SIGs as part of your program at ML2, I would not be appraising their existence or their performance as part of any ML2 appraisal.  There is NOTHING in the model that mandates that PATs be used.

So the short answer is "NO" you don't need PATs for your ML2 appraisal.

Hope that helps!

How do we apply TS and PI to a pure testing environment?

We provide a pure testing service for one of our clients and we are wondering how to apply TS and PI to these activities.

We are going for CMMI Re-Appraisal (Level 3) during the last part of this month and our Lead Appraiser has clearly said that all process areas, except SAM and those at the organizational level are applicable to even these projects. He wants us to map the direct & indirect evidence against the SP's for TS and PI in the PIID sheet for these two projects.

How does one do that for a projects that is just Testing projects?

Excellent question!  The CMMI, especially the PA's Technical Solutions and Product Integration, are steeped in the language of software development, and it's not always obvious how these might apply to another discipline, such as testing (or paper clip distribution for that matter!)

But first I must digress.  SHAME ON YOUR LEAD APPRAISER for not providing you with some guidance!  Your original message (which I edited because it's my blog not yours) indicated he was insisting on certain things without providing you with some guidance and interpetation of the model.  He should be ashamed.  We're not auditors - the SEI clearly states that a SCAMPI Appraisal is a COLLABORATIVE event, not an audit.  While he can't do the work for you, he should be helping to guide you.  'nuff said.

The good news is it's entirely possible and simple to apply to these two PAs to your situation.

TS: What testing alternatives did you explore?  Did you use a purchased tool, build your own, or reuse something from another project?  Did you write test cases (code)?  Did you reuse them from other projects?  Did you create a test plan with instructions?  I bet you did a lot of this, and these are directly related to practices in TS.

PI: Did you set up and environment for testing?  If you do load, integration, and performance testing you must have determined sequence, developed interfaces. and so on right?  Well, these are good examples of tasks in the PI process area.  By the way, it's some of the SAME evidence you might provide for VAL and VER.

Best of luck on your appraisal!