On Thu, Aug 16, 2001 at 02:58:17PM -0400, Chris Nandor wrote:
> In article <[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] (John Adams) wrote:
> > I guess I'm not too picky about whether value is arbitrary or inherent
> > (especially since my philosophy is that all value is ultimately
> > arbitrary) as long as there's value there.
>
> . . . my point is that certifications have literally *no* value or meaning,
> and a reasonable person cannot even argue that they do. They do not say you
> are competent. They do not say you know anything in particular. All they
> say is that you took a test and got a certain percentage correct. They mean
> nothing at all to anyone who knows anything about the subject matter.
`And a reasonable person cannot even argue that they do?' Chris, you're
coming dangerously close to invoking Godwin's Law here. [ for the record, I
submit myself as a reasonable person who is arguing that he does ]
A well-administered certification will *correlate* to competence.
> > How does passing the bar exam damage
> > law? How does being a member of the appropriate medical society (and the
> > certification, including putting those little letters on his nameplate)
> > make my prosthodontist less competent?
Strawman. Those are licensed professions. You're forbidden by law
from doctoring or lawyering without appropriate certification -- this isn't
a question of separating the wheat from the chaff.
> I didn't say it says anything about competence. That is my point! It
> doesn't say you are competent, it doesn't say you are incompetent.
> While I will concede professional societies have some value, even those
> have extremely limited value. I would never trust myself to a doctor or
> lawyer based on his certification by a professional society.
>
> Would I ever consider certification of some technology thingamajig?
> Perhaps, but only in those areas where everyone is certified. For
> example, I wouldn't look too closely at a Novell engineer who wasn't a
> CNE. Why? Because *everyone* who knows Novell is a CNE. And therein
> lies the danger. We force people to become certified just because
> everyone else is, and people who may not want to be certified, but are
> far more qualified than many who are, get marginalized.
If I were paying a contractor to build a bridge, I would want the
chief engineer to be certified as understanding best common practices
in bridge-building. Moreover, I would want that certification to come
from an organization that I believed to have no conflict of interest
in handing out certifications (MCSE and CNE don't fit the no-conflict-
of-interest criterion). And I would appreciate professional guidance
in evaluating professionals from outside of my areas of expertise, like
bridge-builders.
> > What utility it has is statistical.
>
> And the statistics are clear: almost all good perl programmers are
> uncertified. *shrug*
I'm sure that you know the flaw in this argument, Chris.
> > There aren't any absolute guarantees that someone is competent
>
> Sure there are! A thorough interviewing and screening process can
> absolutely determine this.
If you can always know when you've thoroughly screened somebody,
then you've solved the Halting Problem. An interview is a type of
test. If it's theoretically possible that an interview can distinguish
programmers by skill level, then it's also possible that another sort
of test can. Are you claiming that certification programs necessarily
exclude whatever wise evaluation techniques _you_ would use in an
interview?
> > , including the informal, ad-hoc certification (for lack of a
> > better word) that goes on in the interview process. It does, however,
> > help sort out the obvious losers, and that's a utility.
>
> My contention is that it does no such thing. I have met many obvious
> losers who were certified in one technology or another. In fact, the
> MAJORITY of people I've met who are certified by Microsoft are pretty
> clueless.
Microsoft has a conflict of interest that skews MCSE scores. MCSE
could only be fixed by having someone other than Microsoft issue it.
I'm not suggesting that any certification process can (or should)
guarantee either 0 false positives or 0 false negatives. A useful
certification program, IMHO, is one that turns up more qualified
candidates than a dartboard method would. What are your criteria?
- Kurt
