Professional Software Engineer Exam coming in April 2013

Really? Just because its structured like waterfall doesn't mean its "how to do waterfall". I read the req spec and there isn't a thing on here that isn't valuable information for every engineer no matter what process they use.

Read what it actually covers and the topics they test and tell me there's something on there that is completely useless.

Yeah, not a big waterfall (or agile, or any other pyramid scheme) fan myself, but this kind of bullshit did get a man on the Moon and then a mixture of them got some kit on Mars, and generally they are doing OK.

Some of them can actually scan you or your friend's internal organs to see if they're alright too - that's not going to be php on a lamp stack. I'm guessing they are probably not moving fast and breaking things either. They may be taking it easy with your friends internal organs, just so they don't get sued by the growing legal community on HN.

There's also all that bizarre software that somehow delivers you to the airport of your choice. I wonder how that happens - do they have their Agile/Waterfall/Kanban/Nonsense boards on their office walls in order to get you there?

What would you rather see on an exam then? Is the entire notion of requirements/design/etc. not "agile" enough, or what?

There are writers who write novels, and there are writers who write textbooks. While the end results look a lot alike. Both have covers, both have words, both convey concepts, the purpose of the books is completely different. As such, the methodology for writing the book is different. There might even be some commonalities between the two. You might do fact checking between the two... but the facts matter a lot more in a textbook, then a novel. While you actually can use the same process a group of professors use to develop a textbook to write your novel, it doesn't make sense to use the same process a novel writer might use to develop a textbook. To do so would be potentially irresponsible to the students who consume the textbook.

And "Maintenance" is only 7.5%!

Yes, like requirements, prototyping, user interfaces, minimizing complexity, profiling and performance, unit-to-usability testing, metrics, refactoring, performance monitoring...

All that ugly stuff that's wrong with the industry.

While I agree, I think that a large portion of developers want some way to set themselves apart from the masses - to be recognized as Uber 10X code ninjas. Those people will never rest till they create their own credentials, whether it be number of contributions to opensource projects on Github or Coderwall achievements or a standardized test.

That's because in "high brow software engineering", coding is for "workers" (as in "construction workers").

I'll always remember the words of one my professors when practicing for the EIT/FE. He said, "The EIT doesn't test to see if you know Engineering, it tests to see if you know how to read." I have found that statement to be quite true, and the difference between those who pass and those who don't are more a function of how well you read and comprehend exactly what the question is asking for more than anything else (e.g., the "closest answer" is often a different Engineering requirement than the "best answer").

So, which language should they use? Java? C#? Knuth's MIX?

What if the job requires a different language -- does the Engineer's professional certification suddenly become invalid, thus denying him work because of a technicality, even though he otherwise has tons of experience in that language?

Also - that would really fragment the hiring scene. But a bonanza for the testing company.

It could be necessary to be licensed for government jobs or jobs in large companies building real time critical systems (ie power grid control software, avionics, etc). I doubt it'll ever catch on for programming jobs in mobile/web/etc

I wouldn't be surprised if state government and possibly federal government programming jobs required it for all but entry level jobs as they currently do with things like civil engineers.

> "Will it be necessary to study up and get licensed to remain competitive?"

Only if you intend to work for code sweatshops and "consultancies" that value credentialing above track record/experience.

In other words, if you're stuck at the bottom of the programming jobs barrel, it may behoove you to get this. For those of us lucky enough to avoid that fate, we'll spend more time laughing at it than studying it.

Probably nothing. But companies that either choose or are "forced" into this will probably start to differentiate "software engineers" from "programmers" (treating the latters as just a "technical labour").

The new exam just gives states a way to evaluate applicants for licenses in software engineering. I haven’t heard that any state—not even Texas, which sponsored development of the exam—has changed its engineering practice laws or regulations.

If and when such changes are made, I expect the result would be something like the situation for electrical engineers now: a nominal requirement for licensure coupled with broad exemptions that make it practically irrelevant for most people working in the field.

In most states, only a licensed (registered) engineer may use the title “electrical engineer,” practice electrical engineering, or offer electrical engineering services to the public, but there is usually a so-called “industrial exemption” from these requirements; this section of the California Business and Professions Code is typical:

  6747.  (a) This chapter, except for those provisions
  that apply to civil engineers and civil engineering,
  shall not apply to the performance of engineering work
  by a manufacturing, mining, public utility, research
  and development, or other industrial corporation, or
  by employees of that corporation, provided that work
  is in connection with, or incidental to, the products,
  systems, or services of that corporation or its
  affiliates.
   (b) For purposes of this section, “employees” also
  includes consultants, temporary employees, contract
  employees, and those persons hired pursuant to third-
  party contracts.

As I understand it, this exemption exists because employers convinced the Legislature that corporate liability, including strict liability for product defects, would be an adequate substitute for licensure as a means to protect the public from professional incompetence.

Practically speaking, the only electrical engineers who must have licenses are those designing electric power systems for buildings or working for public utilities, which often choose to employ only licensed engineers in positions of responsibility, even when this is not legally required.

As a licensed engineer myself, I’m not sure this state of affairs is best for the public or for the engineering profession, but I’m also not convinced that software engineering, as a discipline, is sufficiently mature that even a nominal requirement for licensure would provide enough real benefits to justify the burden on practitioners.

It is the last step. The other steps vary by state, but in general, it requires that you have at least four years of documented and progressive experience under the direction of another licensed Engineer. You also need to have taken the EIT, which is a general Engineering exam that covers all disciplines, not just computers (brush up on your thermodynamics and concrete mixing!).

If you do not have an ABET degree, then it is still possible to obtain a PE, but the requirements generally go from four years to sixteen years of documented experience.

I may be wrong, I don't think this will be a route for Computer Science. It is more for those individuals who have been programming under other Engineering hats ("controls", "electrical", "mechanical") but now have a way to take an exam that satisfies their actual Engineering work. For instance, I might have fit more under the hat of Software Engineer (since I did a lot of embedded programming), but had to take the Electrical Engineering exam instead due to lack of selection.

> I am guessing I could answer enough (how many is enough?) of these 80 multiple choice questions to pass the exam right now.

The cutoff scores vary by year, but roughly 70% is a good benchmark for passing. Unless you fail, your actual score is not revealed. You can see the pass rates here [1], but not the actual cutoffs.

[1] http://www.ncees.org/Exams/PE_exam.php

Generally, the PE exam itself is considered one of the least difficult portions of the process, though it's still stressful for certain disciplines. The PE exam topics are also generally known to lag reality, regardless of discipline, and often by a decade. It's also open book and some people bring entire carts of books, though I find that not to be very effective since you basically only have six minutes per question.

The major part of the application process is that you will need to work under the direct supervision of a PE, and have increasing levels of documented responsibility, which is often harder to satisfy. Depending on the state, three of the five (minimum) of your recommendations must come from licensed PEs, who need to be able to certify more about you than simply your character.

Usually when threads like this appear, too much emphasis is placed on the exam rather than the requirements that allow you to sit for the exam in the first place. The PE license process is not like a Microsoft or Cisco certification. It took me a good six months to get all the paperwork ready to apply for the exam. I minimally studied for the exam itself. Consequently, four years ends up being more like five or six, in my case, due to less than stellar record keeping.

And for an (admittedly dated) example of what can go horribly, horribly wrong in some fields: http://en.wikipedia.org/wiki/Therac-25

It's ok if your Web app occasionally spits a 500 error or your game crashes once in a while. It's not ok if your radiation therapy machine occasionally flips out and kills people with a massive overdose.

This exam is just a scam. Engineers working in those areas you described are fully aware and do not need an exam to proove it. Engineers new to those areas will not get a job because of having done this exam. With just a so low percentage in QA passers know jack about testing very-whatever scenarios. They just get dipped into QA sauce like a hairless chicken breast. Same for all other areas described in the exam IMHO. There is a 'great' handbook from some US military on how to develop their software. It is a massive block of everything - but after working on every single item described in there you have in depth know edge about processes, code standards and test coverage.