Addendum #2 - Tejas Software Consulting Newsletter
August/September 2002

-Danny Faught

My objections are as follows:

• Certification programs are premature, and will have a chilling effect on the improvement of the craft. We don't yet know what practices work, why they work, and what skills make them work. We don't even have a unifying vision of the craft. I'm worried that certification programs will reinforce an orthodoxy that has no business being an orthodoxy. This has already happened with the CMM, which is, in my opinion, an anti-methodology-- something that actively discourages the development of sound software development methodology by pressing the claim that we already know how to do it, and thus should stop "reinventing the wheel."
• Certification programs represent the values of a community, but there is no single community of testing. The CSQE process was dominated by NorthEastern technical culture, particularly the telecom interests. The Silicon Valley culture had almost no voice in the process. When a certification that is applied to all the craft is created by a few old boys and girls from only one little niche, that's a poorly founded program. Of course, they couldn't have gotten all the communities together, because the CSQE organizers had no concept that there was more than one community, and even if they knew that, they wouldn't have been able to handle the disagreement.
• Current certification programs are not based on any systematic research into what makes a tester successful; what skills they have. They are based on surveys, at best. Bad surveys, poorly executed.
• Certification programs aren't sufficiently tested. Is there some idea in the framers' minds that an incompetent tester would probably not pass their test, and that a competent tester would not fail the test? On what grounds do they conclude that? How is that even a testable hypothesis?
• Certification is the single most important prerequisite to licensing, and licensing will be the end of innovation, because it will drive up the costs and liabilities of innovating for refuseniks like me.
• Certification is motivated not by any altruistic notions of craft betterment, but rather is driven by the expectation of profits from scaring people into getting certified. Look at the Certified Microsft Systems Engineer program. It's an industry unto itself.
• I'm afraid certification is also motivated by the need that some extroverted mediocre practitioners have to feel like they are experts. They can't find the summit of the craft, so they make a summit wherever they happen to be standing, and grandfather themselves in.
• Real certification would have to be skill-based, like the way pilots and doctors are trained and certified. No one is even trying to do that.

I plan to continue developing an alternative vision of the craft that works better, and ultimately will be embraced by the world at large (I predict). This view happens to directly challenge the content of most certification orthodoxies. People who adopt my view of testing would probably not feel inclined to be certified, since to obtain the certification they would have to profess views that they did not actually believe.

This is a fight, in the sense that I think there will be real consequences for people on either side of the debate. I think, as time goes on, I will not be able to take a colleague seriously if he or she feels the need to be "certified." It will be as if a delegate to a meeting of top astronomers announces that they use Tarot cards to discover the age of the universe. I expect to see more attacks on my point of view, and I expect to write more attacks against the other side. Thus, there are people I am cordial with today whom I expect will be my enemies tomorrow.

The only certification I seek is peer certification. I seek out the finest minds in the field and test myself against them. I can imagine a skill-based certification program, in twenty years or so, that I would respect, but probably not any time soon.

-- James