SAN JOSE, Calif. – Only 12 people took the first exam given in the U.S. to certify software engineers working on safety-critical systems, and only six passed it. The National Council of Examiners for Engineering and Surveying is driving ahead with plans to offer the exam again next April while educating engineers and employers about its significance.
“This was the first offering and efforts are underway by the IEEE and others to promote the exam to future candidates,” said Tim Miller, director of examination services at NCEES. “I suspect there was a lot of ‘wait-and-see’ from potential candidates,” he added.
The exam gives a license that could open doors and provide job security for engineers working in utilities, traffic control, automotive, wastewater management and other critical infrastructure areas, backers say.
“I think there is going to be increasing demand from American public officials and the public at large that these types of systems be built under the supervision of and by U.S. licensed professional software engineers,” said Phillip Laplante, a professor of software engineering at Penn State who chaired the committee that developed the test. “The types of systems that are regulated under the licensing provisions cannot be [effectively] offshored,” he added.
Meanwhile LaPlante continues to work on educating engineers and employers about the importance of certification. “I have been pretty busy giving about two talks a month at various conferences and meetings and also spreading the word through publications, radio appearances and webinars,” he said.
The NCEES offers a study guide for the exam. About 30 states offered the exam last April.
I've been chronicling the exam saga in my blog for over a year. I've covered the history of this up to the current exam time and how we got to this point.
There is no free study guide, what they are selling is a list of books, for $40 from IEEE, the sponsor of this mess, that you must read. I I summarize the exam training material here:
covers NCEES refusal to disclose how many took the test when I asked. If this is about safety why hide the data? Glad someone finally got some numbers. The prerequisites for the test exclude a wide group of experienced people from taking the test, which is why they numbers are low, and will remain low.
In my personal view this is being pushed by people that are selling training material. I've covered that as well.
has links to all the past articles. Best to start with the oldest one first.
I have my PE in Electrical Engineering, but I have worked in firmware and hardware design and debug for the last 30 years. I became registered in late '87. Back then there was no such thing for software.
But also back then, it was seen as a right of passage and even if your target employment didn't require it, it was seen as a mark of quality. These days it seems irrelevant. It basically permits me to design building electrical systems which I could probably do if I put my mind to it, but I have no experience, so it would probably not be the best design.
As @Synthetic eluded to, it seemed the value began to erode when the waves of H1B's began to arrive and outsourcing became common.
Having looked at the links provided by @BPaddock_2, most of what is there seems irrelevant to what I do, and none of this could guarantee that the code written by someone that passed this test would be worth a flip.
Then, you have to wonder if this applies to hardware design which is becoming much like software thru the use of HDL's which actually look like software , require some of the same best practices as software but in fact is not the same thing at all. So at what point does this cross over to Electrical Engineering and could there be equivalency between the 2?
Fact is, any PE can stamp any drawing (that's the old school saying) as long as the contents reflect the regular work you do. So there are some grey boundaries.
Lastly, the study guide way back then was 49.95 but was a couple of hundred pages and was hard bound. But still pretty useless except as a guide to the categories of questions that would
be asked. So I purchased one each of a 1000+ pages "Problem Solver" study guide and worked the problems every night for 3 months. I passed with an 82, but the 49.95 book in of itself was pretty useless.
I can't help but feel that this is another effort to harvest money based on a created need rather than an actual benefit to having passed the test. Does anyone who took the test and passed or took the test and failed care to comment on the applicability of the test to the real world challenges of safety critical programming? I remember with some disdain the various tests that I took as an undergrad, they relied on theory rather than reality. I do remember most clearly one course Advanced Electronics, that brought the practical into the course and the related tests/quizzes/homework. Pretty sad considering all the courses that I took that only one stands out. Again, does this new test help to teach or really verify that someone is capable of safety related software?
What is the evidence that the committee that designed the exam is the best qualified entity to preventing software glitches in critical systems? Will the examination detect the "out of the box" thinkers who will really solve these currently unsolved problems? The worst software development experience I ever had was by a CMM level 4 certified vendor. They made mistakes that few self taught programmers would have made (and blamed us for not paying extra for the CMM level 4 COMPLIANT code). We didn't know that actually using their capabilities cost extra.
I've become cynical in my old age. All of these certifcation programs, I've come to realize, are simply a way to get money from people. They are simply another business model. They are designed to create a revenue stream for studying for and then taking the test, then they create an infinite revenue stream for people to retain their cerified status.
In other words, it's a racket.
Proof positive of this is what happens with H1B visas. Any purported value of these certifications gets trumped by lower cost labor.
The AMA is carefully watching this new telemedicine stuff carefully, for that same reason. You don't see them hyping up the wonders of technology applied to medicine, especially when that tech can get in everyone's hands.
Bert, I don't think that is a cynical view! I would say that you are being realistic! Any certification / licensing that is brand new would and should be questioned as to motive. That said, I was wondering also about the correctness of the test and the certification of those who created the test?
Bert, for once I'm even worse than you are! I see this as just an IEEE -sponsored turf war, they're upset that the "gold standard" for safety-critical development is RTCA DO-178C, and they know historically the professional organization for FAA-related work is SAE not IEEE. They're also "striking while the iron is hot" because they're just as acutely aware as I am that there hasn't been any serious new avionics development going on since the '07 crash. As far as medical is concerned the AMA may have good intentions but the only FDA SW development I was involved in I was threatened with being fired if I didn't sign testing documents fraudulently, I of course refused and they carried out their threat. So you start out by realizing the FAA carries a "big stick" but the FDA is a paper tiger where the monetary clout of the players insures there will never be any credibility to that field, therefore IEEE had to set up elsewhere. And this is all when I was just starting to get over using H-1B visas to bring in minimum-wage engineers (and low pay levels are of course FAR more important than product quality), please excuse me for my horrendous level of cynicism!
Given there is actually no shortage of engineers in the US I would suggest that the H-1B visas should only be given if the holders are paid at the very least the average local salary of an engineer in that field and at that level, not 1 cent less. Then they would realise they are better off hiring locals who generally better understand what they're working on. I worked for Delco years ago and they ran development programs in Singapore to design engine management systems. It was disaster, where people that may well have been good engineers were developing vehicle systems SW and had never driven a car. It just didn't work.
There's some misinformation in the article, Rick. Regarding "The exam gives a license that could open doors and provide job security for engineers working in utilities, traffic control, automotive, wastewater management and other critical infrastructure areas, backers say." The exam does no such thing. A license to become a Professional Engineer is granted by a licensing board (there's one for each US state plus a few more for US territories and DC). The boards that recognize software engineering (and not all do yet) will typically use a passing grade on the exam as one of the qualifications for licensing, but it's not the only qualification.
One motive for this exam and the prospect of licensing that underlies it is that software engineering has been struggling to become accepted in the engineering community. One of the criteria that many engineers cite for legitimacy of an emerging field of engineering is the existence of a PE licensing exam. This is why you can now get such exams for aerospace engineering, for example.
The fact that almost half the software exam takers flunked is one indication that the exam is not something to scoff at.
I laugh at the many claims that some organization or other (notably the IEEE) is promoting this exam in order to make money. The IEEE Computer Society and IEEE USA got involved at the request of the licensing community and the PE community because they were viewed as the most authoritative source of information on what constitutes genuine software engineering (developing software using engineering principles and techniques). They responded because doing so is consistent with their mission.
OK, laugh away. You've probably been flying commercially for quite awhile, therefore at multiple occasions you've trusted your life to software that was developed to extremely rigorous specifications that the IEEE had absolutely nothing to do with. I wouldn't take the risk to ride a bicycle at night that had an IEEE-approved taillight because their first interest is making a buck and their second interest is in "locking in" the need for the approval of academicians most of whom have never had to make anything work in their lives (and have nothing but contempt for those who do - that's how it was my entire time in school anyway - isn't it interesting no one there ever tried to dispel that impression?). The notion that "here comes the IEEE, THEY'RE authoritative, now throw away everything else you thought you knew about software safety engineering" is just an arrogant power and money grab, pure and simple, whether YOU think so or not.
I can't even imagine a single test doing anything to accurately indicate competence in software development. It sounds like this test is more about the process than about the coding, but being disciplined at following a process doesn't really indicate competence or lack thereof.
I think it could possibly help understand someone's attitude toward software development, but so can a good interview. Even it it were a good test, there are so many divisions of software development, that understanding one methodology would not necessarily say anything about the particular methodology that I might apply in my company. It's also quite possible that a person could do a good job of passing a test about the process, but not really be able to code.
From the perspective of someone who might be required to take the test, I didn't go into civil engineering for a reason. I can't imagine how spending a lot of time studying other engineering disciplines could make me a better software developer. (Except, of course, that I thank knowledge in general is a good thing)
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.