Re: Perl or Web Certifications?

["Jeff Jonas" <jeffj@panix.com>]

> Does anyone else think the whole certification scheme is a scam?

I'm of 2 minds on that.

    the optimist:

There's a need for showing accomplishment and knowledge learned on-the-job.
If there were a proper career path via apprenticeship
for engineers (EE, computer science, network, etc),
programmers, analysts, sysops, etc.
then there would be respect for certifications, licenses or whatever.
Particularly if administered by well-known "neutral" organizations
such as the IEEE or ACM.
I have my Bachelor's degree in engineering and am continuing for my Master's
since I see much merit in what I learn in college.
There's no substitute for going thru a formal class in semaphores,
synchronication, threads and multi-processing to really understand
the details and design considerations.  Certain fundamentals MUST be taught,
and that's classroom stuff since on-the-job training always
glosses over the details and design considerations
("that's what's in place, kid, work with it").

But I'm a practical guy and a lot of the college requirements
have proven mostly useless in my 20 year career.
I learned a lot "on the job" and that keeps increasing with every job I have.
The success and proliferation of the O'Reilly NutShell books
are a clue that learning things as-needed on-the-job ASAP is more the norm,
particularly as companies are cheap
and won't give the budget or time for training classes.

My high school had 2 tracts: academic and vocational.
Vocational ought to offer a parallel career path to college
since "high tech" is so volatile
(yet I've already contradicted myself: we're told to "learn on the job",
but on our own time, our own resources).

Colleges use standardized tests such as SAT, GRE and such for admission.
Industry feels a similar need.
Every time I interview, I'm given a different test
to prove I know C, C++, Unix, etc.
Many of the tests are faulty or ambiguous.
It would be nice if I had a certificate to show them and skip that.


    the pessimist:

I learned Perl entirely on the job (thanks, Seth!)
but none of my work environments were truly object-oriented
so regardless of the language or environment (C++ or Perl),
I'm not expressing myself as neatly as possible.
How am I to better myself if there's no clear path-to-enlightenment?


I have inferred negative feelings about M$ certifications from

a) a friend at HP whose entire department got M$ certifications
in many things, not to better their careers, but because the managers were
competing to see whose group would achieve the most in a year.
One must wonder what kickback///incentive the managers were receiving for that.
Oh, none of the employees got advancements or bonuses for doing well.

b) at one job, at every lunchtime, I saw a fellow
with his nose deep in a thick exam-cram book.
He was always studying for yet another M$ certification.
Apparently his job depended on his constantly gaining new certifications.

c) the certifications are book-knowledge.  No hands-on testing.
Many co-workers at several assignments were M$ certified
and still could not answer basic questions about using M$ products.
(such as what's the Visual Studio equivalent of "make clean"
so I can integrate source control system such as CVS,
what do these temporary files do?)

d) the M$ certifications expire by time and version.
Things like NT 4.0 certification means little anymore.

e) how is one NOT to feel that M$ certifications are yet another
source of revenue without the need to make any new products?


Many other companies are riding on M$'s coat-tails: BrainBench, Cisco,
even RedHat Linux.
I took the RCHE (Red Hat Certified Engineer) exam
and missed by just 1-2 questions.  $750 spent with nothing to show.
I could not afford the $2,000 Red Hat classes,
so I bought Michael Jang's RH302 RedHat 8.0 study guide since it claims
  "100% complete coverage - all official test objectives
   for exam RH302 are covered in detail".
I need help with ramming the book up Jang's ass.
Seriouly, I need some way to prove that the book was fraudulent in its claims.
I wrote over 5 pages of errata
since the book was positively WRONG with many answers,
and did NOT cover many things I needed to know to pass the exam.
Several fellows at the test commented that the book I had was useless.
The publisher listed no errata
and acted as if I was the only one to request corrections.
The book spent too much time on PC basics
and not enough on the Linux subsystems that were on the test.
It didn't even go into mail forwarding, mail aliases,
automounter and other things that were REQUIRED on the test.
I lost a LOT of my study time just proving that the book was wrong!

Only thru the kindness of a fellow I met during the test breaks
was I clued-in to the exam's vigor.
To quickly summarize: the test has 3 parts

1) your machine won't boot.  Using a rescue floppy and netboot,
repair it so the list of requirement is met (certain accounts are
properly set up, certain services auto start, etc).
You many NOT reinstall the OS.
You must leave the room and the system must reboot properly
and fulfill the requirements.  No second chances.
But you may reboot the system as many times
as time permits to assure you're ready.
I was surprised that most of us completed that with time to spare.

2) Web-based multiple choice quiz.

3) You're given a list of requirements.
You must install the OS on a blanked system.
But IT IS IMPOSSIBLE TO FULFILL ALL THE REQUIREMENTS!
There is NOT ENOUGH TIME!  I had *NO WARNING* of the strategy needed
to endure this, for the grading is on several sections!
Spend too much time on one section and the overall score may be insufficient!
This is where the book was outright WRONG!
Not only was there no clue to test-taking strategies, but
there was no time for "X" or web-based configuration tools:
editing the config files directly is the most time efficient
(but the book emphasized GUI based configuration tools,
many of which still spewed errors and warnings at me with RH 8.0).

Oh, and since the RHCE 302 is a superset of the technician's test,
there's some way to qualify for both, or just the lesser.
I had no clue how to aim for the "lesser one" during the exam.

And there are many contradictions:
Red Hat has a vested interest in folks taking their classes,
but won't publish a self-training book for those of us who
cannot afford that.  So they give trusted folks "early access"
for the training manuals, yet take no responsibility for
its content or usefulness.  And folks fear speaking up
since we're all under a non-disclosure agreement
which risks revoking ALL certifications present and future.

That's mostly moot now since RedHat is now to version 9
and all certifications are based on the latest version.

To bring this back to the original question:
I felt good about Red Hat until my RHCE experience.
So who can you trust for unbiased certification?

-- jeffj
-
**Majordomo list services provided by PANIX <URL:http://www.panix.com>**
**To Unsubscribe, send "unsubscribe phl" to majordomo@lists.pm.org**