Fog Creek Software
Discussion Board




Certified Software Development Professional

The topic of certification of software developers has come up here before as a way to "professionalize" the SD profession.  IEEE has created a certification for Software Development.  What do people think of this certification?  Has anyone here or that you know completed it?  It appears to reference a lot of books that are highly regarded such as The Pragmatic Programmer and Code Complete.  The main problem I see with it is it appears to adhere to some very rigid SD principles and methodologies while the industry is moving in the Agile development direction.

Here's the link:
http://www.computer.org/certification/

chris
Saturday, July 19, 2003

I disagree with some of the example questions that ask about a specific standard.  The questions, 'According to ISO/IEC Standard XXXXX or IEEE Standard XXXXXX what are... etc etc', must be written by people who study these standards daily and know them inside and out. 

Software development simply isn't like that.  How many software developers do you know that know the IEEE standard for software development?  I bet if you went out and asked 10 different developers what IEEE was they would say that they do not know.  Now ask an electrical engineer the same thing and he/she would probably be able to give you an answer.  Chances are that the EE is the one developing the software for the instrument they made or at least they are on the team that develops the software.

Software development shouldn't be rigid.  It simply isn't a rigid process.  There are general steps you follow, voila... SDLC, and this is what most developers know and follow.

To take a test that asks about specific standards, which were created by people with too much time on their hands, is ridiculous.  There is absolutely no need for this.  Did you pass your systems devleopment class?  Yes.  Ok then.

anon
Saturday, July 19, 2003

Ah my dear anon, you stole my response out from underneath me.

Here's the sample question:

>1.  ISO/IEC 12207:1995 describes which of the following:
>I.  A standard software life cycle process
>II.  A framework for software life cycle processes
>III.  A process for defining and improving software life cycle processes
>IV.  A process for the acquisition of systems and software products

The correct answer is I don't give a flying rat's ass. If someone I was interviewing knew the answer to this question, I would toss their resume in disgust. What a bunch of freaking wankers at the ieee to create such a worthless and idiotic certification. Those suckers paying $450 for the ahem privledge of answering these dumbass questions that don't have jackrabbit to do with real life development will get what they deserve.

Dennis Atkins
Saturday, July 19, 2003

Let me clarify a bit -- I could go through their study guide with my 13 year old niece and get her to where she could pass that cert with flying colors. And no, she doens't know squat about programming, development or architecture. These aren't questions that measure whether you know anything at all about development to call yourelf a professional, nor do the questions predict whether or not you can do anything at all. They simply test if you've memorized the study guide. The reason why I'd toss the resume is that anyone taking this cert must be so desperate that they think memorizing a bunch of multiple choice responses will compensate for their lack of ability. But it doesn't. I bet this certification will go over like a wet dream in the asian and arab countries where memorization of multiple choice responses are the way their unbelievably lame-ass education systems are run. That's my prediction -- you'll see plenty of folks from the 'listen, memorize and obey' countries sporting their freshly minted certifications.

Dennis Atkins
Saturday, July 19, 2003

I agree that question is particularly bad and it happens to be the first one on their list of examples.  Some of the other questions are ok and are general enough that they don't get into language or platform details.  Other than the details of particular questions, do people think this type of certification (general software development principles instead of vendor specific certs) is usefull?

chris
Saturday, July 19, 2003

I'd be much happier if it was a Certified Software Project Manager exam. IMHO that's where the focus needs to be - your project could be staffed with all certified software gurus and it's still gonna be a train wreck if the project manager doesn't know how design should work or what "refactor" means.

Philo

Philo
Saturday, July 19, 2003

I bet this certification will go over like a wet dream in the asian and arab countries where memorization of multiple choice responses are the way their unbelievably lame-ass education systems are run.

I don't know about arab or asian countries education, but I know rote memorization is the standard in US public schools.

Tom Vu
Saturday, July 19, 2003

I agree with both of you on those points.

--

This does prove one thing - it certifies that the ieee doesn't know anything about effective software development.

Check out the people who have passed so far -- surprised that so many of them live in cities where NASA has development centers, or where design of high end military systems is done? These processes are important for NASA and the Defense Department to certify their people in. Fortunately, the employees can pass these test while high on their crack. Unfortunately, the test doesn't check whether the developer knows that there is a difference between metric and imperial measurements when building a mars probe, or on what side to grind the mirror of a billion dollar space telescope. You think the guys who designed the O-ring on the space shuttle columbia or that signed off on an enormous frozen chunk of styrofoam hitting the wing at 1800 mph weren't following these processes? They were! They did everything correctly! THey followed the process with due diligence! No one can be blamed for the deaths of the users or the loss of hte space probes! It's not our fault. This certification is a money making scam, it's a way of avoiding personal responsibility in hiring ("But they were certified") and development ("I followed the process") decisions, and it's also a political statement about which flawor of the day is best.

Check this out:

23.  The best approach for the interchange of a large amount of structured data on the Web is to use:
[a] eXtensible Markup Language (XML)
[b] Hyperlink Markup Language (HTML)
[c] Component Object Model (COM)
[d] Java and Java-based technologies

Well? What's the answer?
Here's what I would respond:
"First, you misspelled hypertext. Second, I want precise definitions of 'interchange', 'large amount', 'data' and 'structured'.  Third, who is the meathead who decided these are the only choices and what is his justification for that. Fourth, where is the study empirically proving that one of these is the 'best' solution."

Check it out:
Gotcha #1: the data is 140 gigabytes of image data arranged in folders. One side of the connection has a dialup connection.
Answer: put it on a tape and overnight it.
Gotcha #2: the data is 1 gigabyte of image data arranged in folders and both sides of the connection have a T3 line.
Answer: ftp it.
Gotcha #3: the data is 100 terabytes of data stored in 8 gigabyte 'movies' and is being sold one file at a time to 75 million people with dial up lines for $15 each.
Answer: store on DVDs and charge for shipping.
Gotcha #4: the data is 12 megabytes of classified data on the new Penetrator Cyclonic Missle.
Answer: put encrypted data tape in a locked titanium suitcase, handcuffed to a trusted courier who is trained in the deadly arts.

Dennis Atkins
Saturday, July 19, 2003

As if Tom Vu!

Dennis Atkins
Saturday, July 19, 2003

Dennis:

they probably look at questions by individuals and pay them a certain amount if it is selected for the exam, ..then again it being IEEE - they look at questions sent in by people who have double PhD's:-)

Philo:

How about the PMI Certification?

Prakash S
Saturday, July 19, 2003

http://www.pmi.org/info/PDC_CertificationsOverview.asp
PMI's Project Management Certification.

A Software Build Guy
Saturday, July 19, 2003

Wow. Those are incredibly stupid questions. I know the IEEE had a lot of problems trying to get the certification process running simply because their first efforts at developing the body of knowledge were prepared by people with nothing better to do.

It comprised irrelevant garbage pulled from old text books, and so on. They had to revise the whole thing, trying to get experienced software developers to contribute. I'm not sure they succeeded. Some of the contributors are just technical writers for example. Nothing against technical writers, but they shouldn't be setting guidelines for software development.

The ACM is formally opposed to the process because they think it doesn't work. Having seen the IEEE questions, I'm with the ACM.

http://www.acm.org/serving/se_policy/selep_main.html

To the original poster, the only reason some form of "certification" would be useful is if it could be used to protect the investments that software engineers make in their careers; i.e. to exclude the easy entry that undermines wages and jobs. That's the main reason for "certifications" in other fields, whatever the ostensible reasons about protecting the public.

.
Saturday, July 19, 2003

Thanks dot, I agree with the acm on this too, and that article increased my opinion of the acm tremendously - those are cats who have actually developed.

Prakash made a good point that some of the questions seem to be taken out of old textbooks that advocated practices that have never really panned out for many people.

One thing that bothered me was that if that's the best the ieee computer society (should note here at least once that it is a different membership than the ieee proper) could come up with to certify a developer, they must not get to many people in their roster who have experience with real world development. A score on that test is just not going to be strongly correlated with development ability or competance.

Dennis Atkins
Sunday, July 20, 2003

Oh! it was dot who mentioned the textbooks and Prakash who suggested it was PhDs selecting them, both good points.

Dennis Atkins
Sunday, July 20, 2003

I'd be much happier if it was a Certified Software Project Manager exam. IMHO that's where the focus needs to be - your project could be staffed with all certified software gurus and it's still gonna be a train wreck if the project manager doesn't know how design should work or what "refactor" means. - Philo

How about the PMI Certification? - Prakash S

PMI's PMP ("Project Management Professional") certification would be the right answer to Philo's question if it (a) were specific to software development and (b) had anything to do with the real world. With regard to (a), I think PMI used to have some sort of industry-specific add-on to the PMP called the CAQ ("Certificate of Added Qualification"), but they no longer do. Trust me, you'll never see the word "refactor" on that test.

And with regard to (b), trust me on this -- I've had to forget  what little I know about project management in order to study up for the test I'm taking next Saturday. Blech.

Zahid
Sunday, July 20, 2003

Okay, now I've had a chance to review the links to computer.org and ACM ... like previous posters, I agree with ACM's stand. How can you license someone when there's no commonly accepted way to describe what they do and how they do it?

But I think it's important to note that the two pages are not necessarily contradictory ... the IEEE site proposes to certify software engineers, not to license them, where the ACM position paper speaks out against licensing. In fact, the ACM page specifically endorses certification. They may, however, have a concern about the broad scope of the IEEE certification.

za
Sunday, July 20, 2003

Zahid,
Can you tell me how did you prepare for this exam? books? training classes? resources?
And which one are you taking PMP or CAPM?
For a person with 3 years of PM experience which one is better? Thanks

Anon
Sunday, July 20, 2003

Can you tell me how did you prepare for this exam? books? training classes? resources?
And which one are you taking PMP or CAPM?
For a person with 3 years of PM experience which one is better? - Anon

I've been using study materials from Rita Mulcahy Corporation ( http://www.rmcproject.com/ ). They consist of a book called PMP Exam Prep, flash cards called Hot Topics, and test simulation software called FastTrack. I believe that, altogether, these cost my firm about $400.  The book is excellent and I would call it a must-have. I'm less excited about the other items, particularly FastTrack, which has a few "issues" that make it harder to use. Other sample tests are apparently available from companies called ESI and Bosson.

In addition to the RMC book, definitely get a hardcopy of PMBOK from PMI.

I'm taking the PMP exam, primarily because a lot of California state government agencies are requiring it for high-level roles. I'd say that, if you meet the qualifications (and it sounds like you do), I'd take that exam. I haven't seen many projects that require a CAPM, so I don't know how much value that has.

The PMP qualifications, if you have a bachelor's degree, are: 4500 hours of PM experience within the last six years, and 35 classroom hours of PM training. (Both "experience" and "training" are broadly defined.) If you don't have a bachelor's, you need 7500 hours of PM experience within the last eight years, and the 35 hours of PM training.

Anyway, see http://www.pmi.org for more info, or drop me an email for more info.

Zahid
Sunday, July 20, 2003

I have checked out pmi.org and I do qualify for PMP. Eventually I do want to take PMP. But as I stepping stone I thought I would try CAMP first. Anyway thank you very much for the information, Zahid.

Anon
Monday, July 21, 2003

*  Recent Topics

*  Fog Creek Home