Should Novell, Microsoft, and IBM set qualifications for our staffs?

Vendor-controlled "Certification" Raises Concerns

by Conrad Weisert, CVA
(originally published in the IDI Newsletter, Summer, 1994; layout revised and footnotes added, December, 1999.)
It began with Novell's CNE (Certified Network Engineer). Now we see a growing roster of major vendors setting up their own certification programs, complete with worldwide examinations and elaborate curricula.

Is this a desirable trend that will help confused managers screen out unqualified professionals? Is it a sneaky plot to infiltrate companies with loyal disciples? Will it have an impact on long-established professional certification programs or academic degree programs? Vigorous opinions on all sides are being expressed by individual professionals, companies, and academics.

Professional Certification: The Real Thing

ICCP the recognized certification body for computer people

In the midst of the confusion over all the recent vendor certification programs, the Institute for Certification of Computer Professionals continues to conduct its established programs, administering examinations and awarding recognized certificates. We've expressed reservations from time to time about the relevancy of some of the examination contents, but we firmly endorse the principle of professional certification in our increasingly specialized field.

The original CDP (Certified Data Processor) program launched by the DPMA (Data Processing Management Association) in the early 1960's was highly accounting oriented. Most knowledgeable computer professionals either ignored it or condemned it. When the program was transferred to the newly-established ICCP, it was overhauled.

Now known as the CCP (Certified Computer Professional), the program enjoys the support of the major professional societies, including the prestigious ACM (Association for Computing Machinery).

For information call the ICCP at (708)847 299-4227.


What ever happened to personal computers?

The personal computer revolution was hailed in 1981 as a welcome liberation from bureaucratic data centers. "These machines are so simple and straightforward," enthusiasts claimed, "that we won't need systems programmers or other support specialists any more."

The reality a dozen years later differs sharply from that early expectation. The systems on our desktops depend on layers of monstrously complex software, far more complex in many ways than any mainframe system. Except in the smallest configurations we still need systems programmers -- now with fancier titles -- to set them up and keep them running. Do they have to be certified, too?

What ever happened to manuals?

We used to complain about the prodigious volume of reference material we needed to refer to in order to cope with mainframe software and diagnose problems. We might have complained more, however, if those manuals had not been available to us.

Suppose instead we'd been told that we had to hire a certified expert who claimed to have memorized most of that information? Or to send our own technical staff to a series of expensive courses covering that information?

Replacing the customer engineer

One argument in favor of vendor controlled certification is the vanished role of the on-site (or on-call) customer engineer. A mainframe installation paying $20,000 a month or more to the mainframe vendor could call on such vendor consultants to get answers to questions, for help in diagnosing problems, and for general guidance in the use of the vendor's products.

If a vendor can't provide people at an affordable rate, proponents argue, then isn't the next best choice to use our own staff (or third-party consultants) who meet the qualifications the vendor would require of its own people?

Loyalty in question

Critics reply that that's exactly what worries them: Whose people are they? We may hire them and give them their paychecks. But suppose our network administrator has invested several weeks in taking courses and preparing for examinations and now has a coveted credential and a detailed knowledge of that one vendor's products -- knowledge that would be useless if we switched to another vendor. How objective can he or she be when we're considering alternatives for our next network operating system?

What ever happened to standards?

Vendor certification isn't limited to products like operating systems that are inherently unique. I got a call recently from a corporate training director interested in my C++ courses. Her second question was "Are you certified?" At first I wasn't sure what she meant1. "We wouldn't consider a C++ instructor, " she explained, " who isn't Microsoft certified!"

I tried to elicit her requirements in more detail and to explain to her that 98% of the first OOP/C++ course content should be entirely independent of any compiler, but she wasn't listening. You can't win them all.

In the case of a language or other product that's supposed to be based on vendor-independent standards, we might ask why we need certified support people, certified instructors, and yes, even certified programmers. And in the case of non-standard products, perhaps we ought to be telling our vendors to get their acts together and work toward greater industry standardization.

Concepts versus details

Undoubtedly, wide differences exist among the courses and examinations in verious certification programs. Nevertheless, some who have gone through the process report that the content was highly geared to memorizing facts rather than mastering concepts. Aren't these the very facts that ought to be instantly accessible to any of us in manuals or on-line help? Could holding such a certificate ever be a negative indicator? 2

Whither academic programs

As we see more and more recruiting ads that call for one or another vendor certificate, we also see fewer and fewer calls for degrees in Computer Science or Information Systems. A certificate from Novell, Microsoft, or IBM, according to some recruiters, is more relevant than an academic degree and thus more highly valued as a credential.

The impact is seen in concessions some universities are making to vendor orientation in their curricula. And the highly professional ACM has oriented this year's collegiate programming contest to one vendor's proprietary non-standard programming languages. Will we next see a Computer Science BA with a concentration in Windows programming?

What should an organization do in 1994?3

Tempting as it may be to meet a short-range skill need, I don't advise requiring or even giving much weight to a vendor certificate in recruiting outsiders. We can, however, send our staff to selected courses when they need skills that are not easily acquired from manuals and hands-on use, just as we've always done, and some of these courses may be given by the software vendor.

In any case, let's keep the pressure on our vendors to give us products that are:

  1. simple enough for experienced professionals to cope with after a reasonable orientation,
  2. well documented through manuals and on-line help, and
  3. consistent with ANSI or other industry standards.

Send us your experiences and opinions, positive or negative, and we'll report them in a future issue.


1 -- I do hold the ICCP certificate in C++ programming. As an independent and presumably unbiased consultant, I avoid entering into agreements with vendors, including, so far, product-oriented certification.

2 -- A job applicant holding one or more certifications from a single vendor would trigger my curiosity about his or her breadth of vision and objectivity. While I wouldn't discriminate against that applicant, I would certainly probe those issues carefully in the interview.

3 -- Although this article is over 5 years old, nothing in it is obsolete. If anything, the problem has gotten worse.

Return to IDI home page.