[BSDCert] about a buck fify this time.
jpb at sixshooter.v6.thrupoint.net
Wed Apr 13 23:48:08 EDT 2005
A welcome presentation! Comments in-line below...
* Matt Edlund <matt.edlund at gmail.com> [2005-04-12 22:12]:
> * Matt Edlund <matt.edlund at gmail.com> [2005-04-08 17:26]:
> > Well I saw a post about this certification on Slashdot and I thought I
> > would join the mailing list and see if it is something to take
> > seriously or not. To be honest, that is the largest problem that any
> > kind of BSD certification is going to face; if it will be taken
> > seriously or not. Reading through this I see lots of comments on
> I view 'seriousness' as having a) a solid point of view (not trolling,
> or time-wasting drivel), b) determination, and c) committment.
> I'd like to think we have all three. Time will certainly tell.
> I have no doubt that this is a committed and serious group. I think
> the idea is a very good one. The issue of "seriousness" is very
> complex more so than just than just being a result of the commitment
> of the design group. There has to be a general view that this
> certification is not only necessary, but provides highly skilled
> people, backed by industry recognized standards. To give an example,
My personal point of view is that the quality of the certification
will be in direct proportion to the skill level needed to attain it.
However, there are several considerations here-
1. Employers want verification of skills in the hiring process.
This is a nod to making the certification tough to get- only the
skilled pass the certification test(s), enabling employers to have
more confidence that the certification is valuable.
2. Candidates just want to pass the certification test(s). For the
less skilled who do not pass, this presents a difficult choice:
give up or keep studying/learning and try again. Some will give
up, some will go on. If they continue to fail, they may
eventually give up. Finding the balance point for the level
of difficulty of the exam(s) is quite difficult. We do not want
to discourage anyone, but we don't want to certify lamers either.
3. Providing multiple levels of certification is an option, but
will require additional work for everyone- BSDCG and candidates.
This issue has not yet been decided, but has been discussed at
length many times. Having multiple levels will definitely
help those less skilled to 'get in the game' with a basic cert.
This is the Cisco model, and it has proven effective, so there
is some precedent for moving in this direction. However,
it is not certain that generating a basic certification has
value for employers, who want more skills not less skills.
> I will use how people view the Cisco Certification program and the
> Microsoft Certification Program. Now, first off I am not holding
> Cisco certs themselves up to the spotlight, but instead how people in
> the industry view those certifications, especially in the early years.
> One of the things that Cisco did was increase the price of the
> certification so that only people who had the backing of the company
> they work for could really afford them. They also crafted a series of
Only partially true. I know many who paid their own way, but there
is no denying that the higher costs made it more difficult to get.
I paid my own way for Novell Instructor certification years ago.
> tests and practical examinations which were difficult with little to
> no "flex" room for people to "learn the system". Basically they
> created a way for people who already knew the material, to prove it.
True mostly in the early years, when there wasn't a large secondary
training market. Now, you can drive down the street in almost any
big city and see 'Cisco certification courses here' signs. Even Cisco
themselves publish books, magazines, on-line tutorials, self-contained
CD-ROM interactive modules, and on-line classes. I have no first-
hand knowledge, but I would say that the total production costs of all
their educational material is not recovered through course fees;
it's simply too big.
> Microsoft took a wholly different approach. Instead of creating a
> certification to display a professional's knowledge, they created a
> series of courses that were more geared at teaching people new to the
> industry the Microsoft client/server/enterprise setup. Their idea was
> that the more people you can get who "know" the system, the greater
> your market share. This was a direct result with Microsoft's
> (revolutionary at the time) experiments with "student pricing". They
> found that students who used Word at school, demanded it at the
> office, and managers, responded by making Office the #1 business
> application supplanting Word perfect for dominance in the word
> processing market. Instead of trying to restrict their base of
> applicants, Microsoft did the exact opposite. They contacted
> governments and lobbied to get their course covered under things like
> the Canadian Government's Employment insurance program. Even today,
> these schools continue to churn out thousands of Microsoft Certified
> Systems Engineers and Developers, flooding an already saturated market
> and further reducing the value of the certification. Obviously this
> is all done to tremendous profits for both Microsoft and the training
> industry, but give very little of lasting value to the consumer who
> purchased the training.
> Cisco made their mistake in that they went for too elite of a group.
> They priced their certifications at such a level that even companies
> who wanted to get their staff certified had second thoughts. However,
> that certification is extremely valuable to those who hold it.
This was (and still is) essentially market driven. The high-dollar
cost of the equipment, and the critical markets they serve played directly
into the hands of customers who *demanded* they provide high-quality training.
I wouldn't view it as a mistake. Their reputation was on the line
at a time when other vendors still had significant market share for
network equipment. By providing a certification to highly skilled
network engineers, they solved two problems at the same time- customer
demand for skilled, knowledgeable engineers and they filled a big gap in
> It is with this in mind that I came up with these principles that I
> believe that a BSD certification should have;
> Core Values that a BSD Certification should embody
> 1) RTFM
> 2) Value that increases over time
> 3) Respect for accomplishment and knowledge gained
> 4) Community
> 5) Money should never be a bar to skill
> This, at first, may seem like a joke. But think about how you came
> about your knowledge of BSD. Certainly the majority of us didn't
> learn the intricacies of pf.conf in a class room lecture hall. We
> read man pages, searched the web and the forums and we found the
> answers to our questions. If you ask a question only fit for a n00b,
> RTFM, in some form or another, is likely the answer you will receive.
> Reading the manual is the first step in every case.
I totally agree here. And fortunately, the BSD manual pages are
generally of very high quality, making learning much easier.
> 2) Value that increases over time
> I have a list of certifications including ones from Microsoft and
> Novel. I don't even put them on my resume anymore. The reasons for
> this are various. For one, the underling value of these
> certifications is questionable to begin with and like many experienced
> professionals I obtained them early on in my career a long time ago.
> We all have certifications made obsolete by newer versions of the
> software we trained on. While I know Windows servers intimately,
> certainly more so than when I got those certs, claiming I am an
> certified anything in Windows NT 4.0 is pointless. BSD is an
> operating system that has evolved and grown over time and I believe
> that the value of a BSD certification should do so also. To this
> effect I propose a structure of courses offered as levels, with only
> the most supremely skilled holding the highest levels. In this
> fashion we can insure that as a professional's knowledge grows, so
> will the value of the certification he holds. This leads me right
Only time will tell if the certification grows in value, but I agree
in principle here. As above, my view is that the value of the
certification will track very closely with the skill levels of
those who achieve certification. However, if all we focus on
is the 'supremely skilled', we will lose ground with those less
skilled. Ensuring that there is a clear path from lower skills
to the highest skills will encourage those less skilled to proceed.
> into my next value that I believe that a BSD certification should
> 3) Respect for accomplishment and value gained
> Some people jump off cliffs for the rush, some people drive
> dangerously fast down the highway, I solve computer problems. The
> rush that gets you when you solve a difficult problem and the beauty
> of a truly elegant solution is as addictive as any drug. When after
> days of research and hours of coding and testing all the errors go
> away and the problem is solved, the elation that we feel is why so
> many of us live and breath computers.
> If Linux has sharp edges, then BSD is made of broken glass. It
> requires careful handling and a sure knowledgeable hand. Respect for
My analogy is that of a sharp sword, but we are in agreement here too.
> this knowledge should be among our core principals. Certification
> should be based on skill and knowledge and where applicable should be
> granted upon recognition of such. Those who are the core developers
> of BSD and their peers in BSD should be those with the highest
I don't think that only developers deserve the highest levels of
recognition. There are certainly very knowledgeable system administrators
that use BSD and know it intimately, but who are not kernel coders.
There may be a need for a developer certification, but I think that
is a different certification altogether. We are focusing primarily
on the user community and sysadmins in particular.
> certifications. If you believe that Theo de Raadt needs testing to be
> considered certified at the very highest level you are welcome to try
> and tell him that. I am sure you will get the answer they deserve
Well said, but it's not our focus to tell anyone they need to be
certified. Our focus is to provide a certification that others
will *want to get* because it proves they have demonstrable skills.
> 4) Community
> I believe that an effective certification should come between the
> examples set by Microsoft and Cisco. While the certification should
> be a example of a professional's knowledge and skill, it shouldn't be
> limited to only those who work in the networking departments of
> fortune 500 companies and while it should be valuable to those who
> hold it, it must also strength the community from which it grows.
> Graduation shouldn't be the end of the involvement in the community
> for those who choose to earn their certification.
> I believe that this is more than contributing skill and knowledge but
> also in helping to choose and grade those who will be coming next.
> Multiple choice answers have a right and a wrong answer clearly
> spelled out, but do we really want people who can simply regurgitate
> knowledge or do we want people who can come up with elegant solutions
> based on skill? The tests should be both practical and essay written
> and I believe that a method of community involvement should be
> implemented to let those who already hold high levels of BSD
> certification be the ones who mark the tests. Web solutions are the
> obvious solution and a certification holder can choose how many
> questions they wish to mark or if they mark any answers at all. Those
> who's accomplishments advance our community and the BSD operating
> system should be lauded and rewarded publicly and with respect.
As much as I'd like to agree here, this approach presents problems-
- 'Certification shouldn't be the end' Nice idea, and has been
tried in other places such as SANS. It turns out to be an
enormous burden on those certified to be the gatekeepers of
certification for others. I'm not saying it's impossible; I'm
just saying that the effort required to be graders or reviewers
for others is substantial, and may not be possible for many. Consider
that the higher skilled often find themselves in the most critical
jobs at work (usually putting in extra hours already), and that
with family, social life, etc. there are just too many other
- People roll in and out of certifications faster than skills.
If a cert expires, does that mean I can't be a reviewer anymore?
If I screw up reviewing or grading does it cancel my certification?
Lots of issues here.
> 5) Money should be no bar to skill
> BSD is free, the information to gain the knowledge to earn this
> certification is free and therefore, so should the certification at
> all levels.
I doubt if this will be possible. There have already been costs
associated with the certification effort. More will come. And
as much as I'd like to fund it all myself, my credit card would
burst into flames.
There will be some cost, but as has already been stated, no one
in the BSDCG is out to make a fast buck. The Group already has
a sense of belonging to the BSD community as described above,
and that is what is driving much of the effort.
> By keeping these core values in mind we can insure that we have a
> certification that is in demand, provides value to it's holders, and
> strengthens the community that created it. We read the manual, we
> respect ours and others accomplishments and knowledge. We contribute
> to the larger BSD community and we commit to keeping the
> certifications free and accessible to everyone with the only bar being
> the knowledge of those who choose to so test their skills.
> But that is just my opinion
> Mathew Edlund
> Edmonton, Alberta
Very thoughtful presentation. Thanks for pulling it together.
More information about the BSDCert