[consulting] The Case Against Drupal Certification

[Evan Leibovitch]

Jeff Greenberg wrote:

> So with development, even though I give a nod to the fact that sooner
> or later there will be Drupal certification, I again question the value.

As you should. As should everyone. There are perfectly legitimate
reasons to conclude that a Drupal certification might simply be too
impractical to succeed. And even the most avid supporter of
certification needs to be skeptical and aware of the pitfalls.

It is a mistake to infer, based on what I have written, that a Drupal
certification is naturally advisable, let alone inevitable. There are
many factors working against it, but these factors have little to do
with many of the common objections raised in the current discussion.

To me, the biggest hurdles facing any Drupal certification are these:


*(1) How do you set a standard based on a moving target?*

Drupal's rapid pace of evolution is a certification body's nightmare, in
some ways writing an exam for Drupal is like taking a leisurely stroll
on Star Trek's Genesis Planet. By the time that reasonable skills exams
(and supporting training materials) are available based on Drupal
version {x}, it's likely that release {x+1} is in beta testing or
better. Then there is the fact that some of the most-used modules such
as CCK and Views don't necessarily have development schedules that are
completely in sync with core. And some modules eventually find their way
into core.

Quality exams take time to create, and if the changes between version
{x} and {x+1} are substantial, the core certification objectives need to
change as well as the exam questions themselves.
Book writers, course developers and others will have a hard time keeping
pace. Some already do fall behind.

It might be necessary for Drupal to slow its release cycle in order to
make a certification practical, and I'm certain that such a tradeoff
would be unacceptable to many -- including me.


*(2) You can't easily test creativity*

Doing certification exams for system administration because there is a
general set of software tools which are used to implement the sysadmin's
changes and policies. Even so, you can't easily test troubleshooting
skills which require quick thinking and creativity. This kind of talent
needs to be discovered in interviews and reference checking, not
certification.
For software development the situation is more acute. You can test one's
knowledge of the programming language concepts and syntax, of the APIs
and libraries. But you can't easily test the creative component of
software development any more easily than you can test troubleshooting.

Sure, you can do it with hand-on tests, but using such techniques
skyrockets the cost of the exams (the RHCE exam is $749), makes them
less accessible and slower to develop, and introduces a level of
subjectivity into the testing that could invite "technique bias" on the
part of the testers.


*(3) A question of value*

LPI was very fortunate to get sponsors before the dot-com bust. Getting
sponsors now will be harder to find and funding in lesser amounts. That
will make exam development slower (if the project has to depend on
part-time volunteers) and possibly of lower quality (if you can't hire a
psychometrician to oversee the program). This could be recouped through
exam fees, but there's a pretty low limit on what most people would be
prepared to pay. For most people, certification will all boil down to a
brutally cold assessment of value -- "will getting certified get me more
job prospects and/or more money per hour than not being certified?" If
the answer to this question is not a clear "yes" the certification is a
bust.

Sure, there are some people who collect IT certifications like I collect
hot sauces, but there aren't enough of those people around to keep a
certification viable. Most will only get certified only if they must, or
if they see compelling financial value. That means that buy-in --
respect -- from hiring bodies is critical, so their participation as
stakeholders in the certification process is vital. Having a
certification that doesn't concentrate on testing skills which are
important to hiring bodies is pointless; having the objectives only
driven by developers will only get the project so far.


I honestly don't think a Drupal certification can succeed without
tackling the above issues. Having said that, these obstacles are
challenges rather than deal-breakers -- none of them is insurmountable,
but they will require significant time and dedication to address.

Is anyone here -- developers, vendors, hiring agencies, and/or others --
up to the task?

(I'll be glad to advise and assist a certification initiative -- as a
non-developer, this would be a way for me to give back based on my own
skills -- but I'm in no position to lead or drive such an effort.)

- Evan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.drupal.org/pipermail/consulting/attachments/20090812/48261176/attachment.htm>