Linux certification

[Brian Densmore]

Been done already by LPI. Most grueling is of course Red Hat. It will
become the standard by which all others will be judged (or so Red Hat
hopes). Certification isn't necessarily anti-opensource. True the gurus
of Linux will never be asked if they have certification, except by total
incompetent personnel managers. But most of us aren't that well known in
the Linux world, and certification states that you have some "proven"
knowledge. I can say I am the greatest Linux wizard the world has ever
known. But if I don't have a portfolio of developed software to back
that up, how is a potential employer going to know that I really am? No
certification is necessary, but like all things man-made there are ways
of cheating and abusing the system. No we don't need another
certification process; what we need is a way to make one that already
exists so far above the rest in quality that there is no question as to
the quality of a certain certification. I say let's support the LPI
certification, the open-source certification created by the Linux
community.

Opensource your mind,
Brian

> -----Original Message-----
> From: DCT Jared Smith [mailto:jared at dctkc.com]
> Sent: Wednesday, October 17, 2001 10:52 AM
> To: kclug at kclug.org
> Subject: Linux certification
> 
> 
> The whole certification process is antithetical to the methodology
> of Open Source, or Free Software, with the following explanations:
> 
> 1. At its core, Free Software allows _anyone_ to participate, and
> by their merits alone (not through money, or fancy marketing
> techniques, or passing some exams), they are able to build software
> that people all over the world contribute to, and use. Thus the idea
> of 'certification' is already built in to the way Open Source works,
> yet there is no official list being kept. It stays organic, this way.
> For instance, does anyone question Linus's ability to administer
> a Linux installation? No. He wrote the thing, for God's sake. Same
> with any of the other Open Source wizards; their abilities certify
> them automatically.
> 
> 2. Some won't see it this way, but Certification is a process of
> excluding some people from participation... and it is in a manner
> which can be cheated more easily than raw competition between
> programmers, who are challenged by the fact that if their code
> doesn't work well, it gets forked into something that works better,
> not for malice or money, but for the benefit of all users. When you
> start Certifying people, you create whole new way to exclude
> people, and it's not entirely merit-based.
> 
> 3. Certifying people who know Linux is like certifying people
> who know how to write. Authors usually become popular because
> of their natural ability, not because they passed an exam 
> certifying them
> to be able to write. Whenever you begin certifying people, 
> you create a
> new layer of people who are able to pass certification exams, 
> but unable
> to effectively administer Linux, ie 'paperware' admins. This is true,
> and sad when it happens.
> 
> 4. The whole idear of 'Certification' is created by people who do not
> understand these points--at heart, they are interested in 
> marketing, and
> they
> hire geeks to help them: ie, they have found a way to make money
> by creating something out of thin air (Certification 
> Standards), and then
> demanding students pay money to meet those standards. This is
> a marketing technique which is not the way Open Source actually
> works.
> 
> H o w e v e r ,
> 
> To be a purist about certification, when the whole world is
> clamoring for certification, is one approach. Thus, you could say
> "Certification is Anti-Open-Source, therefore I will not participate."
> End of discussion.
> 
> Another approach is, "Certification is Anti-Open-Source, but
> employers cannot grasp that concept, and will soon be asking
> all new Unix admins to be certified... therefore, let's do what we can
> to make sure the certification process is grueling and 
> difficult, unable to
> be passed by people who have no clue about Linux, because it
> involves a deep array of real-world examples and situations which
> will challenge a 'paperware' candidate, but be fun for a person
> who is like the current Linux gurus."
> 
> I hope you followed me there. What I mean, is, as 'Certification'
> folks sweep into town, offering 'Certification courses', which
> employers will be impressed by while actual admins will find them
> to be annoyingly shallow,
> 
> why not put together a rock-solid Certification Preparation course,
> as part of KCLUG's activities, which ensures QUALITY of
> knowledge above QUANTITY of knowledge. This is a difficult
> challenge. Raise the standard; for if not us, who else will? 
> I personally
> believe
> that Linux is best learned, not from books, but from actual 
> use. I think one
> of
> the reasons people are attracted to Linux is that it is FREE 
> TO LEARN IT.
> So, charging lots of money to learn Linux creates a layer of 
> people who miss
> the point of Open Source. That is a waste of resources, if you ask me.
> 
> The Certification Prep course put together by KCLUG would have
> a reputation of being the most difficult, but also of producing the
> highest-quality graduates (no matter what Certification they finally
> pursued). And we would make sure the standard was so high,
> the 'paperware' courses would always be striving to meet that 
> standard.
> It's possible. Thus, the whole certification process becomes another
> Open Source project... :-)
> 
> To do this, the Certification Preparation course needs to 
> provide LOTS of
> hands-on Linux'ing. And students should not be 'babysat.' Let 
> em sweat,
> that's the best way to learn Linux; but not sweat so much 
> they give up...
> 
> It believe part of the course should involve a full-on hacker 
> Honeypot,
> administered by some serious Linux gurus, for the purpose of 
> teaching people
> security issues--give the students something to 'hack' into, yet do it
> legitimately,
> evaluating their progress all the way, and providing informal 
> mentoring
> (the way real hackers learn) and you will create a group of people who
> KNOW SECURITY, not just _know about_ security. Firewalls built
> by such graduates will be comprehensive, detailed, and flexible, not
> out-of-the-box click-here firewalls.
> 
> (What's a honeypot?) Here's a honeypot link:
> http://www.linuxsecurity.com/resource_files/intrusion_detectio
n/network-intr
usion-detection.html#11.

I believe part of the course should involve creating and managing a
real live mail server, creating and managing a real live webserver,
a real live network server, and all of this could be linked together in
a real-live setting which is running on a local area network that
doesn't
need to cost a million, that webserver needs to have  some serious
scripting happening on the server side, to keep graduates from knowing
administration without knowing at least PHP, Python, or Perl, definitely
configuring a whole new install of Apache from the makefile on upward
(not using precompiled exes).

These are examples from my own very limited knowledge of Linux,
we need a lot more examples of this kind of information in order
to build the kind of excitement which will make this whole idea come
to pass.

Kansas City can give something distinct to the world: A group of
Certified Linux folks who really know what they're doing. It will
require a collaboration of many, many people to make this happen
well, but that's what Linux is all about, is it not? Surely, others
are facing this same challenge, so let's build on what's out
there, but keep the Quality High, and the flavor distinctly
Barbecue.

If we do this, it needs to be done right. Otherwise, why begin.

It should be inexpensive, not driven by money, but then again,
what fifteen people have ten hours a week to spare for the next
few months, to make this happen? It will not be easy, but the
final result will be a blessing to us all.

Thanks for listening.

-Jared