Re: [SystemSafety] Licensing software engineers - care to take a test?

From: Steve Tockey < >
Date: Sat, 15 Mar 2014 16:20:40 +0000

Oh. I should have added…

The example exam question is even wrong from the perspective that you don't "verify" requirements. You "verify" that the design and the code are consistent with the requirements (by testing and/or peer review). You "validate" that the requirements represent actual user/customer need. So the question really should have read "107. Requirements validation techniques include:"


From: Steve Tockey <Steve.Tockey_at_xxxxxx Date: Saturday, March 15, 2014 8:12 AM
Subject: Re: [SystemSafety] Licensing software engineers - care to take a test?

A couple of comments in reply:

*) Canada, UK, and Australia have been licensing software engineers for many years. In Canada it's called "P Eng". In the UK it's called a "Chartered Engineer". I don't recall the label in Australia. In each case, I don't know how much any kind of exam figures into the entire process but I'm pretty sure there's some kind of critical exam somewhere in the process. And in all cases, I'm not aware of software being treated as any kind of special case. In other words, they all parallel the process already established for Civil Engineering, Chemical Engineering, …

*) The difference between licensing and certification is that licensing is mandatory to practice in some arenas (e.g., in the US it is where the health, safety, and the welfare of the general public is at risk) while certification is purely voluntary. I run the Certified Software Development Professional (CSDP) and Certified Software Development Associate (CSDA) certification programs for IEEE Computer Society (see CSDP has been available for about 13 years, CSDA for about 6. CSDA can be obtained based on exam score alone. CSDP requires a 4-year degree (or CSDA) and about 5 years of on-the-job experience before one can even take the exam.

*) I am/was a member of SELC. The "Principles & Practices" exam (aka "P&P") that you are referring to is somewhat modeled on CSDP. There are many similarities but I can't speak for the precise details because I wasn't involved in the actual Software P&P exam. At an absolute minimum, CSDA/CSDP and the Software P&P exam are all based on the "Guide to the Software Engineering Body Of Knowledge" (aka "SWEBOK Guide", see According to the National Society of Professional Engineers (NSPE) one needs to be a professional engineer (PE) to work on any P&P exam. The Software P&P exam was created by people who are licensed PEs but in (mostly) different disciplines.

*) There have been licensed Software PEs for at least 10 years or more. The state of Texas had a waiver process where if you had roughly 15-20 years of on-the-job experience (less with a relevant advanced degree), you could get 2-3 already licensed PEs to write recommendation letters. The nationwide NSPE-level process was largely motivated by Texas getting inundated by waiver-process requests and not being able to handle them.

*) The requirements for getting a Software PE vary by state. Some states require a relevant technical undergraduate degree, some don't. Some states require the candidate to have already passed the "Fundamentals of Engineering" (aka "FE") exam. Some require more or less on-the-job experience. Some give more or less credit for having an advanced degree.

*) I would go for the Software PE myself, but my home state of Washington requires passing the FE exam first and it's quite the chore from looking at the FE study guide. An entire chapter of the study guide is on physical constants alone…

So understand that one is not able—under any circumstances that I am aware of—to become a licensed engineer of any kind (at least in the US) on the basis of one exam alone. Some degree of education and some degree of experience are required as well.

As well, as I say so often, certification and licensing do not prove competency—they only serve to reduce the risk of incompetency. Licensing a pilot, a doctor, a lawyer, or a hairdresser(!) does not guarantee that said individual is competent, either. Again, it just reduces the risk of them being incompetent.

Finally, the answer to the sample exam question is obviously (A). In (B), a project management review has little if anything to do with requirements verification. The same can be said for debugging in (C) and critical design reviews in (D). The question is really way too easy in my view. I would hope that the exam questions on the real P&P exam are much more thoughtful than that one. Given how difficult it is to write a good exam question, it's completely understandable that the poor questions make it to the "sample questions" pool and not on the real exam…

From: Les Chambers <les_at_xxxxxx Date: Friday, March 14, 2014 11:29 PM
Subject: [SystemSafety] Licensing software engineers - care to take a test?

There is a developing trend towards licensing software engineers, especially where they perform work on systems that affect the property, health and safety of the public at large. In the US, software engineers can be certified by State licensing boards if they have an accredited bachelor's degree in software engineering or equivalent, four to six years practical experience and can pass a multiple-choice test. I have included a summary of what's happening in the US below. Is anyone aware of similar testing initiatives elsewhere in the world?

Intrigued as to how you could possibly judge someone's competence to build a life-critical system with a multiple-choice test, I downloaded some sample questions. Given the conversation we recently had on V&V - would anyone care to test their arm on the right answer (as per the IEEE)? Cheers

Les Chambers
Chambers & Associates Pty Ltd<><> Twitter: _at_xxxxxx M: 0412 648 992
Intl M: +61 412 648 992
Ph: +61 7 3870 4199
Fax: +61 7 3870 4220

The System Safety Mailing List
systemsafety_at_xxxxxx Received on Sat Mar 15 2014 - 17:20:53 CET

This archive was generated by hypermail 2.3.0 : Tue Jun 04 2019 - 21:17:06 CEST