Re: I'm seeking for OCP certification

From: DA Morgan <damorgan_at_x.washington.edu>
Date: Fri, 13 May 2005 07:54:21 -0700
Message-ID: <1115995809.639835_at_yasure>


[Quoted] Lewis C wrote:

> I wouldn't compare a DBA or developer to a surgeon or any other
> medical profession. Very few DBAs have a job that could kill someone.

I have heard this argument before and it is absolutely untrue. Lets examine the facts.

The biggest Oracle installations on the planet (from a US standpoint as I can't speak for other countries) are at Homeland Security, FBI, NSA, and CIA. A mistake can be fatal ... not just to one person ... but potentially to thousands or millions.

All hospitals have databases used for everything from scheduling surgical facilities to running their pharmacies. Want to bet whether an error can have fatal results?

How about the databases controlling maintenance of airplanes? Willing to bet your life on it?

And I could go on with thousands of other examples.

But even if fatal consequences are not associated with databases where do you think your credit card information is held? Your bank accounts? Your income tax records? How happy would your life be if your stock broker had a little problem in their system and transferred your retirement funds to my account?

But your argument also doesn't hold water in that the average family physician, dentist, pharmacist never stands between life and death in their day-to-day practice and yet all are licensed? How many attorneys deal with life-and-death issues? Not all bridges fell down before civil engineers were licensed? Not all airplanes crashed before aerospace engineering was properly santioned. Were the Wright Brothers licensed? Would you want the engineers at Boeing and Airbus to be winging it the same way? Sorry I just can't buy this argument. Oft repeated but never backed up with fact.

> Or that will make a building or a bridge fall down, for that matter.
> Other than that, I mostly agree.

Having ranted a bit above ... I feel slightly chagrined now reading this ... but I'll leave the above intact with only slight misgivings.

> On 1 and 2: For Oracle's test from Thompson, I found the questions to
> be fairly well written and technically correct. There were a few that
> could interpreted different ways. I really think they try to do a
> good job on this. The tests go through a beta period and they are
> updated over time.

I agree: They try. The problem is they don't succeed and as Jonathan Lewis and others have pointed out repeatedly here they don't succeed in far too many cases.

> On 3: I completely agree. At least a C level. What's that, 85%?
 >
> On 4: I'm not sure how that would be tested. Also, there should a
> progression. Drop OCA, OCP and OCM and switch to Junior level cert,
> senior cert, journeyman, etc. Have a baseline test and then make it
> harder and longer for higher levels.

Using the model of the medical boards as an example here is one way I would test a DBA for competence.

Take the DBA candidate into a room with 6 single-proc Oracle servers. Each server is running Oracle and has a different problem. One might be missing a control file, another have a corrupt log file, another someone to a hex editor to the System tablespace, etc. Instruct the DBA candidate to choose any 3 machines and:

1. Properly diagnose the problem
2. Fix it
3. Bring it back on line with as much recovery as is possible
4. If recovery is not complete explain why

Would it be difficult? You bet.
Would it be profitable for someone giving the test? Not the point. Would those that passed be technically competent? Provably so.

> On 5: I didn't know that occurred. Does Oracle sell the answers?

No but others do and Oracle should sue.

> And on 6: Prohibitively expensive to grade.

My point exactly. Certification should not be about profitability and will never have any meaning so long as that is part of the criteria. But look at my example, above. Cost of grading = $0.

> You come from academia so I think your viewpoint on expense will
> probably be different than mine. I don't think a certification should
> be unattainable for most people, and that includes the cost.
> Not just
> the cost of the test (and I think $125 is very fair) but also the
> costs of training material. I do think someone should have a few
> years before they can be designated senior. I don't think they should
> have to pay $2500.
>
> Thanks,
>
> Lewis

I think a reasonable cost is in the $2500 range. That is what medical students currently pay and that is after incurring many years of expensive education that no developer or DBA has done. Further, first year physicians make less money than DBAs so, if anything, the DBA is in a position to pay far more.

But you miss the point. This isn't about the DBA or their money. Quite frankly the personal comfort of the DBA is irrelevant to certification. It is about employers having a level of confidence in the competence of those they hire just as you go to a licensed physician or surgeon or attorney because you want to know you too are in good hands. If you had a problem with your taxes and the IRS would you want me handling them for you, however well meaning, or would you prefer a CPA or tax attorney?

The point of certification for engineers, dentists, pharmacists, any other profession is proof of competence. $125, about the cost of what? a few tickets to a baseball game? Is an insult to the concept. Certification, if it exists, must have both meaning and value. And I would expect that if real certification existed those with it would be able to command substantially higher salaries thus repaying the cost of the required education and testing.

-- 
Daniel A. Morgan
University of Washington
damorgan_at_x.washington.edu
(replace 'x' with 'u' to respond)
Received on Fri May 13 2005 - 16:54:21 CEST

Original text of this message