Coming Soon: Agile Certification

Scott tackles the topic of certification for agilists.


June 08, 2007
URL:http://www.drdobbs.com/architecture-and-design/coming-soon-agile-certification/199902742

Scott is a DDJ Senior Contributing Editor and author of numerous IT books. He can be contacted at www.ambysoft.com/ scottAmbler.html.


Lately, there has been a significant amount of discussion about certification on agile mailing lists and at various agile events. Certification is a divisive topic—many people are clearly against it, some are ambivalent, and a small minority is clearly for it. Given the choice, I would prefer that we didn't certify agile practitioners, but in the end, I'm a pragmatist: Like it or not, certification of agile practitioners is not only coming it's arguably already here.

Done right, good certification programs do seem to help raise the overall quality of certified professionals. Certification also provides professionals with a body of knowledge to fall back on, putting us in a position to say, "This isn't an acceptable approach," when others try to force us to do questionable things (for example, to commit to a fixed scope, fixed price, and fixed schedule project). In many fields, in particular both physical and landscape architecture, the only way you can land a really interesting project is to be certified. Without certification, you're only qualified to hold a supporting role on a major project. Finally, a rigorous certification program is a significant step for IT to actually become a respected profession—do you know of any other profession without a decent certification program in place?

Many agilists are reticent about certification (see Table 1), but I think that many of us still suffer from a "black or white" mindset on this issue. There is no such thing as a perfect certification process—some people are going to become certified that probably shouldn't have been, some people are going to abuse the certification process, and some organizations will not understand the limits of certification. Then again, some people will use certification to learn new skills, some people will use certification to communicate their real-world experience effectively and thereby gain employment commensurate to their actual ability, and some organizations will improve how they manage their staff. The potential benefits of agile certification certainly outweigh the costs.

Argument Response
We don't know enough to base a certification on This sounds sneakily like big requirements up front (BRUF) to me. Surely, we could have the courage to certify people on things that we do know something about now and have the courage to evolve the certification approach tomorrow when we know more?
Certification doesn't prove anything This is clearly false: Certification proves that you passed the requisite certification process. The bottom line is that the certification process can be as rigorous as we choose to make it. Certification will never be perfect, so just like other professions, we will need to trust but verify when appropriate.
Doctors are certified, but some are still bad Yes, there are no guarantees. But, would you go to a doctor who wasn't certified? Certification can help to weed out some of the people who really aren't qualified to do the job.
Companies will hire someone based solely on the fact that they have certification Yes, this is a really questionable approach. However, no matter what we do some HR departments will manage to find really stupid ways to work. The rest of us can still benefit from a good certification program.
I know a lot of really good people who aren't certified And I know a lot of really good people who are. So what?
You learn skills a little at a time, not as quantum strata measured by different levels of certification That's absolutely right, but varying levels of certification give people goals to work towards and help to provide a general way to distinguish between people's abilities.
There are a lot of really bad people out there with their XYZ certification And there are a lot of really good ones with it. All things being equal, aren't they better off with the knowledge gained by certification?

Table 1: Addressing the arguments against certification.

The Agile Alliance (AA) and Agile Project Leadership Network (APLN) are both considering certification (www.agilealliance.org/show/1796):

It is the position of the board of the Agile Alliance that employers should have confidence only in certifications that are skill-based and difficult to achieve. We also believe that employers should not require certification of employees.

To explore certification, the APLN recently started a learning and recognition program; its discussion forum is the apln-lr@ yahoogroups.com mailing list.

Towards Certification

I believe that much of the pushback against agile certification is the result of existing certification programs that are questionable at best. One clearly insufficient approach to certification is to just take a course. This is the strategy to becoming a Certified Scrum Master (see the sidebar on page 68) and the Open Group's TOGAF Architect Certification. With these two certification programs, you respectively take a two-day or a four-day course; and at the end, so long as you paid reasonable attention to the instructor, you become certified. The only thing that certification strategies such as this certify is your ability to pay for a course and then show up for it.

The second clearly insufficient approach is to simply take an exam, an approach that is common for many technology certifications, such as becoming a Sun Certified Java Programmer (SCJP). A similar but more rigorous certification is the Microsoft Certified Solution Developer (MCSD) for .NET, which requires you to pass five exams. The only thing that these certification strategies prove is your ability to study and memorize a bunch of facts.

To make product-specific certification viable, I believe that vendors should certify people on specific versions of their products and require people to recertify either when a major version is released or a specific period of time, two to three years seems fair, has elapsed. An examination-based approach is arguably sufficient for product certification for two reasons. First, our expectations of someone being DB2 v9 Certified is much lower than someone being a Certified Data Warehouse Architect. Second, vendors are motivated to ensure that the examinations are rigorous because if it's too easy to earn the certification, then it becomes meaningless—potentially harming the perception of their products in the marketplace.

An example of a reasonably good certification within the agile community is the Dynamic System Development Method (DSDM) Practitioner Certification (na.dsdm.org/na/training/accreditation.asp). This can be earned via two ways:

The Project Management Institute (PMI), with its Project Management Professional (PMP) certification, (www.pmi.org/info/PDC_PMP.asp) takes a better approach. While I don't agree with all of the material it teaches people, I do respect the process. To earn your PMP, you must either have 7500 hours in a position of responsibility leading and directing people and 60 months of project management experience, or you must have a Bachelor's degree, 4500 hours of experience, and 36 months of project experience. Additionally, you must agree and adhere to the PMI's code of professional conduct and pass a rigorous multiple-choice exam. Furthermore, every three years, you must renew your PMP status by earning 60 professional development units (PDUs) and by reaffirming the code of ethics and professional conduct. PDUs are earned by taking accredited training, providing training, or by reading professional books.

Doing Certification Right

To be truly meaningful, a certification must have some sort of experiential component to it (typically involving internship or apprenticeship) and the applications must be verified by other certified professionals. I recently spent time with Paul Priess, the President of the International Association of Software Architects (www.iasahome.org), who is currently leading the effort to develop a robust certification program for IT architects. His vision is to have three levels of certification:

The IASA program is a work in progress and the organization looking both for volunteers to help be involved and for funding. Although this effort isn't specifically geared towards agile development, I believe that there are several important lessons that we can learn from it:

Unfortunately, the people who are the best candidates for running an effective certification effort are the most reluctant to do so. Within IT, the IASA is one of the few organizations striving to do it right. I believe that both the Agile Alliance and the APLN could easily emulate IASA's example and, more importantly, have the moral imperative to lead the agile certification effort. I also believe IASA should consider partnering with an existing, reputable organization such as the Association for Computing Machinery (ACM) or the Institute of Electrical and Electronics Engineer's Computer Society (IEEE-CS) to accomplish this. Time will tell.

Bringing Ethics to Scrum Certification

The Scrum Alliance continues to embarrass itself, and to a lesser extent the agile community as a whole, with its continued operation of the Certified Scrum Master (CSM) program. To "earn" this designation you need to take a two-day course, at the end of which, the instructor decides whether to award you with it. There is no test and there appears to be a 99 percent plus pass rate—the Scrum Alliance won't provide any figures, but several existing and former Scrum instructors have privately told me that at most a few dozen people have been refused the CSM designation. Some Scrum instructors brag about how they've failed a few people who just sat there and did e-mail during the course instead of listening to them. Talk about not getting it.

It is clearly deceptive to claim that you're a "certified master" of something after taking a two-day course. Although an argument currently rages within the Scrum community over whether the problem is the use of the word "Certified" or "Master," this serves merely to distract people from the real issue. From discussions on several mailing lists, it is very clear that people who aren't a CSM recognize the deception, people who are CSMs recognize the deception, and even the CSM trainers recognize the deception. Yet the CSM program carries on.

The Scrum Alliance is currently working on a new certification effort with a bit of meat behind it, although in my opinion, they lost the moral imperative to do so a long time ago. The people involved with the Scrum Alliance chose years ago to operate the existing certification program and have had ample time to address the ethical issues surrounding it. Ethics is a reflection of the conscious choices that you make, and the Scrum Alliance has clearly made their choices.

The challenge that we face is that many good people have chosen to become CSMs, but their reputations are at risk of being tarnished because of the actions of others. All is not lost, there are several ways that we can take back the ethical high ground:

  • You can simply choose to stop claiming to be a CSM.
  • When you indicate on your CV or in your e-mail signature that you are a CSM, you could follow with the text "earned by taking a two-day course".
  • You could include a URL to a page that overviews the CSM designation.

Organizations can also choose to act ethically. They can require their staff to follow one or more of the aforementioned recommendations for all types of certification, not just for CSMs. They can also choose to promote a code of ethics within IT—I highly suggest the one developed by the ACM/IEEE-CS (www.acm.org/constitution/code.html).

—S.A.

Terms of Service | Privacy Statement | Copyright © 2024 UBM Tech, All rights reserved.