[consulting] Drupal Certification

[Evan Leibovitch]

Hello all,

I'd like ad add a few comments on a number of points raised in this
discussion.

Some background: I was one of two people that took conversations about
certification from a couple of email lists, brought interested people
(and sponsors) together, and created the Linux Professional Institute.
LPI, a non-profit corporation, now operates in almost every country and
is generally respected as a high-quality certification that is
independent of any vendor. I was Board Chair of LPI for its first six
years of operation. Currently, I am advising a group within the
OpenOffice.org community that is looking to start its own certification
program and have also been involved in the design of the Ubuntu and SuSE
certifications.

To address some of the general points (and add a few more):

1) Certification isn't for everyone in the Drupal community, it's not
even for most people. But it can be extremely useful to those who do
believe it has value. If you don't believe that certification is useful,
or even if you think it's supremely dumb, that's fine -- don't get
involved in it. ACtively ignore it, add filters to your email reader if
it bothers you that much. But please, don't be a barrier to those who do
see a value. Please try to see a bigger picture, and don't extrapolate
that because *you* won't benefit from certification, nobody can. The
most gifted Drupal developers will never need certification, and peer
recommendation is extremely powerful. But as Drupal grows, those elites
at the top of the food chain become an ever-shrinking percentage of the
community as a whole, and those at the bottom need ways to identify and
promote their talents -- especially if they have little experience.
Submitting a new module to d.o that nobody uses and falls out of
maintenance within a year should not really be an indication of its
author's suitability to do a project. And not all Drupal projects
require software developers -- the skills of making really good use of
Views and CCK, or creating custom themes based on Zen, are now
marketable in themselves.

2) Certification is, at its heart, a standard -- a line drawn by someone
or some group that says "this is what you need to know about Drupal to
be recognized as being at a certain skill level". The rest is creation
and delivery of materials (books, podcasts, training courses, etc.)  and
the mechanics of testing, and there are many companies and institutions
that can offer those components. But the standards setting itself needs
to come from the community, not a single vendor (even if that vendor is
Acquia). That means (IMO) that the DA need to be involved, if not in the
creation of a cert, than the blessing of a trusted group willing to do
it. While it's fine to encourage diverse and competing implementations
(books, courses, even testing methods) there ought not to be
encouragement of multiple standards -- that's a guaranteed lose-lose
path that simply kills the effort as well as any active act of sabotage
possibly could. OpenOffice.org once had a policy of not endorsing any
certification and waiting for programs to rise up and compete -- but the
real end result was that nobody would make the effort without the
endorsement. So OOo lost about five years of wanting a certification but
never getting one. Now that policy appears to be changing and people
within the community are eagerly up to the challenge.

3) Certification is not, and never will be, a sole or even primary
criteria for hiring quality people. Anyone who hires an employee based
only bassed on a certification -- whether LPI, PMI, Cisco or any other
cert -- deserves the disappointment they're guaranteed to have later.
Certification is no substitute for checking references, good interviews,
or even examining code samples (as applicable). However, most people who
like certification know this, and understand the useful but secondary
role certification can play in helping to match people with employers or
contracts. Developer certifications are even trickier; you can tests
someone's mechanical skills at the language and libraries but you can't
easily test the creative aspect.

4) Certifications can be very expensive to produce well, and may require
significant financial resources if wanting a globally-accessible program
such as the ones done by LPI, Zend and MySQL. Each LPI level cost about
$200K to produce, even with the help of a significant volunteer
community. It is possible to to a cert right on little cash but it needs
many dedicated volunteers with skills in education, standards, and
marketing. One interesting low-budget model that may be worth examining
is that of the BSD certification (http://www.bsdcertification.org/).
Obviously, if there is a larger budget (sponsors, anyone?) the program
can be done faster and less expensively for exam-takers.

5) In certification, brands and trademarks do matter because trust in
the certification program is critical to its success. At one level trust
(in the legitimacy of the certificate) is indeed its only real product.
There will be expenses incurred in fighting cheating, brain-dumps and
memorizers. In some regions of the world, certifications are important
to employers but are just a game to candidates, who will go to extreme
lengths to get their certs with minimum effort (including paying an
expect to take the exam under their name).

6) There has to be an honest evaluation of why a certification is
wanted; there are many possible reasons for wanting one, and the reasons
and/or prioritization will determine the development (quality cost and
speed) of the program.
- Removing barriers of entry to Drupal implementors who are having a
hard time finding qualified people to make/run the site;
- Increasing opportunities for training organizations and book
publishers targeting Drupal students
- Revenue generation (and/or membership size) for the Drupal Association
- Increase publicity for Drupal
- Increase the skill levels of (most) Drupal practitioners
- Increase mainstream 'respect' for Drupal as an alternative to other CMSs

I've probably raised more questions than answers, but that was the
intent. This is a discussion that the community -- or at least the
subset of the community that thinks certification will he useful --
should have.

Evan Leibovitch
Open Source Architect
Faculty of Education and Centre for Refugee Studies
York University, Toronto