Re: Oracle Certified Master - is it any good?

On Tue, 26 Mar 2002, Howard J. Rogers wrote:

> > - sqlplus /nolog (how could they leave that out?)

> It's in the training course notes, so I would expect it would appear as a
> question at some point (the exams being based heavily on the training
> material). Being random, of course, you can't guarantee any particular
> subject will be tested.

I don't see this anywhere in Couchman's 8i book, which is endorsed by Oracle (although I can't understand why).

Granted that his text has been argued to be incomplete for awhile, but this is what I used to get certified.

> > - any _options in init.ora

> Well, they're hidden for a reason, of course.

Kyte covers a few in his experts book. This is another reason for community involvement. I want to see things that Oracle doesn't want me to know in this exam (just to make metalink more interesting).

> > - autotrace/tkprof/explain plan coverage is very poor

> Agreed. The thrust is on getting the things to work, not on interpreting the
> output. Particular raw trace files... they could do with some decent
> explanation.

1ZO-024 is "performance and tuning" - Couchman devotes 7 pages out of 1150 (plus the 3 Net8 chapters on PDF) on autotrace/tkprof. Short shrift indeed, for such an important subject.

> > UNIX stuff, including:
> > - exp/imp with named pipes
> > - here documents
> > - ipcs, ipcrm

> Mmmm.... Now we're in to operating system specifics. If you're going to
> have this one, I demand that you learn the intracacies of the Windows
> registry and the mysteries of VMS. ;-0

My initial reaction to this is that the exam needs some extra-credit, platform specific questions. The distribution of the questions should be based upon the market share of the related sever version at the time of the exam.

I don't know what percentage of Oracle sales are on UNIX, but I remember reading in Oramag that they had a 68% share of the UNIX market.

I could understand why DBAs stuck on Windows would install Linux and an Oracle client to deal with the data in the UNIX way (some things you just can't get with cygwin) - ditto for VMS and others. Understanding of database-OS integration should count for something in the exam structure.

> Well, I disagree with that bit (see above).
> I think what we end up agreeing on is that OCP is the *start* of knowledge,
> not its end.

The start of what Oracle wants you to know, at any rate.

   / Charles J. Fisher                   |"How many who came into this world /
  /  cfisher_at_rhadmin.org                 | with me have already left it!"   /
 /   http://rhadmin.org                  |       -Marcus Aurelius          /
---------------------------------------------------------------------------