Menu

Deconstructing Certification

March 16, 2005

Micah Dubinko

"It isn't where you came from, it's where you're going that counts." -- Ella Fitzgerald

XML and the labyrinth of related technologies have gotten complicated enough that a resume-writer can no longer sum up her XML experience with short bullet points. Or can she? Members of the XML community come from diverse backgrounds and educational experiences. If you had to distill your XML abilities down to a single line, what could you say? Maybe "certified?"

Certifications are topics of curiosity, especially in the field of XML. The most common reaction on mailing lists is usually skeptical dismissal of such tests, usually on the grounds of the questions themselves. Recently, Razvan Mihaiu posted one of these questions on xml-dev, expressing bafflement at the answer the testers were expecting.

G. Ken Holman observed:

I never liked the questions, and I could often come up with two vastly different answers to a given question, or find real-world situations in my experience that could match more than one answer of a multiple-guess type of question.

Michael Kay put it even more succinctly: "Yes, it's a stupid question, and a stupid answer. Sounds like it came from an XML accreditation test: such things are nearly always stupid."

Can an XML Certification be useful, or is it a waste of time? To pick apart that question, one first has to look deeper. I was recently able to do some field research on the subject: Oleg Tkachenko wrote in his weblog about the certification for "IBM Certified Solution Developer - XML and Related Technologies." (Note that the vouchers for free certifications that Oleg mentions have been exhausted.) Other XML certifications include those at Whizlabs and Altova for their XML Spy product.

What's it like? The IBM test has 55 questions--you need to get 58% right to pass. All the questions are computer-delivered multiple choice, some with lengthy "exhibits" that appear in a separate window. Every question states how many correct answers are expected, whether it's one or up to 5. A little clock in the upper-right corner of the screen ticks away your 90 minutes as you go. The physical testing location is a small, lightly supervised room, usually occupied by other test-takers, where testers can't take in any electronics, paper, or pens other than what's provided. When you're done, the receptionist hands you a printout with your results.

In short, a stressful environment. Staring at one of the more bizarre questions, surrounded by ominous posters, while a timer ticks away your remaining minutes is quite unpleasant. The questions themselves weren't as bad as some of the stories you can read, though a few of them seemed completely off the rails. The process seems to be testing your ability to do well on tests as much as any specific XML skills. Is that fair?

Well, in an employment situation, it's also common to have muddy or conflicting information, typos in specifications, and tight deadlines. An important part of any technology job is rationally handling irrational input. So in that respect, the tests themselves are reasonable facsimiles of actual conditions, if a bit exaggerated.

On the technical level, the IBM test has questions about XSLT, XPath, DTDs, DOM, SAX, SOAP, XSL:FO, and XML Schema. Lots of XML Schema. These are divided up into categories for "Architecture", "Information Modeling", "XML Processing", "XML Rendering", and "Testing & Tuning". Though I would have preferred to see XML Schema balanced out more with Relax NG, overall it was a reasonable cross-section of "XML and Related Technologies". In other words, someone who passes this test would have to actually know something about the subject matter.

Certification: Who Cares?

The discussion so far still hasn't answered the question of whether XML Certification makes any difference. The answer depends on context.

As a mental experiment, think of a colleague you trust and respect for XML work, or perhaps an XML Cup winner. Do you care whether they have a piece of paper with a certification on it? Surely not. Now think of the 16th résumé down in a pile you need to go through to hire an XML developer. Does certification make you feel any better about that candidate? Maybe a little, depending on experience. Finally, picture yourself just learning XML, sitting in a classroom. The teacher's biography slide says "IBM Certified Solution Developer for XML". Do you feel better after seeing that? Probably.

The weight carried by a certification varies by both the experience level of the observer and that of the candidate. The more experienced the observer is, the more they tend to discount certification. At best, for the candidate, it's an amplification of acquired experience; beyond a certain experience level, amplification doesn't add anything. Recent comments on xml-dev certainly bear this out.

Actual experience wins over paper credentials every time, hands down. Amplifying a zero still yields a zero! Yet, speaking generally and not just of XML, some try to pass off certification as a substitute for experience, resulting in greater dilution of the value behind accreditation. In extreme cases, having accreditation can even lead to increased suspicion, or at least extra background checks for real experience.

Another interesting parallel exists between certification and academic diplomas or degrees. The level of effort isn't comparable — there's several orders of magnitude difference — but some of the surrounding machinery and social context are related. For one, the mental experiment above can be repeated, substituting "degree" for "certification", leading to similar conclusions. Even a degree isn't a substitute for real experience. Another parallel is that in both cases, there's a certain air of formality, even theater, built into the process, designed to engender a basic trust in the system.

Finally, if the question is "who cares?", don't overlook yourself. Going after any kind of a certification, or degree for that matter, takes a measure of self-confidence, determination, and effort. It feels good to accomplish something tangible, and it can positively shape your attitude toward whatever you go on to accomplish next.

Should You Do It?

In summary, should you pursue any of the XML certifications? If you regularly come into contact will less-experienced users, say as an instructor, or if your personal marketing message would benefit from having a concise bullet item summary of your XML skills, certification might be a good option. If your employer is willing to pay for it, and you have the experience to back up the paper, certification probably won't hurt. In any other case, it's where you're going that counts — figure that out, and proceed directly to point B.

Births, Deaths, and Marriages

Also in XML-Deviant

The More Things Change

Agile XML

Composition

Apple Watch

Life After Ajax?

Reminder: Extreme Markup papers due

The deadline of April 15 rapidly approaches for the August conference.

W3C Workshop on Rule Languages for Interoperability

Position papers for this workshop are due March 18. The actual event is April 27-28 in Washington, D.C.

XRules & Dynamic DOM group on Yahoo!

A discussion group for XRules and related topics.

Documents and Data

Free practice test for Altova XMLSpy certification.

XForms Myths? Point. Counterpoint.

Breakthrough in compression for English text.