Re: Recertification Requirements Revisited

[Alan Mead]

Here are my thoughts:

(1) I agree with Tom and Chuck that recertification should be on the
"regular" exams.  I think it should also be full price (which should be as
cheap as we can make it).

(2) One thought that I haven't noticed (lately) is that recertification can
be simply to make sure the person hasn't forgotten everything (they may not
have touched Linux in the interim).   I cannot remember all the HP-UX or
VMS tricks I knew when I programmed in those environments in school.

(3) Of course, brand new stuff comes out all the time.  But, it seems like
many things are getting easier so maybe it's not the end of the world if an
Admin's exam-certified knowledge is a little bit out-of-date (they can
probably hack the easier stuff if they knew the earlier and harder version).  

(4) As a practical matter, I think that it is a real problem (for us) for
people to certify and then have their certifications invalidated by a new
technology.  This is a problem because I think the expense (in time and
money) of training and exams will be a burden for some which they will
undertake as an investment.  I think we need to provide a way for them to
be sure it is a good investment from which they will receive a return.  We
cannot give them a raise but we can reassure them that we won't "cancel"
their investment in a few days because a new exam is out.

Alternatively, we could close the exams for some period of time before the
new exams come out...   But I think there are many, many disadvantages to
this.

(5) I would say that some of the recertification ideas we've discussed have
been quite complex (e.g., evaluating releases to determine when content had
changed enough...) and would achieve an ideal that I can appreciate.  But I
think others have advocated a simple approach.  I definitely think simple
would be best.  I think that we could become more complex later if we can
get a simple recertification cycle up and running.

(7) I think I am agreeing with many when I argue that LPI should not force
recertification.

Thus, I vote that LPI maintain a database of publicly-available info
regarding who was certified and at what date.  I think the record should
also include a link to a web page describing that exam in detail (including
revision numbers if need be) so that if an employer could tell whether some
key technology was part of the exam.  And along with this database, I
propose that we adopt a written, advisory, policy that one should be
recertified every 18 to 24 months.  

That means that people who certify right before we change exams have up to
2 years to milk their certification before they would be seen as
out-of-date by an employer.  While this offends my perfect ideal of how
this thing should work, it makes a lot of practical sense.  If we wanted to
emphasize this policy, we could mark the records visually after 24 moths
and purge them after, say, 36 or 48.

Alternatively, we could require (or advise) people to be recertified
within, say, 12 months of a new exam being implemented.

I think we should shoot to roll out a new exam every 18 months or so
accumulating changes from the last year and a half.  Having a (relatively)
fixed period would lessen pressure to evaluate each release.  Of course, we
could be flexible if we felt like a new exam was needed by a big change.

-Alan


________________________________________________________________________
This message was sent by the linux-cert-program mailing list. To unsubscribe:
echo unsubscribe | mail -s '' [EMAIL PROTECTED]