[BSDcert] Ideas on BSD Certifications

[Jacek Artymiak (devGuide.net)]

Mike M. Volokhov wrote:

> Thanks to Jacek there are some sort of digest for recently discussed
> topics. Let me pass trough it.
 > ...
> For the two development directions under BSD systems there are not so
> smooth (at least as for me). Those possible includes:
> ... 
> For system developers this also should cover:
> ...

So far, the best way to certify developers' skills was through their 
contributions to the favorite project.  If they can write something that 
meets' the respective project leader's criteria for quality.  If people 
use it, and the developers themselves keep on maintaining their code 
over long periods of time, then it's the best proof of their skills.  I 
don't think that any certification can replace public scrutiny of the 
code itself.  Just my .02 polish new zloty's (PLN) worth ;-)

>>1) Certificate duration.  I'd vote for 2 years.  That seems to be a
>>fairly reasonable period of time.
> 
> 
> IMHO this should depends on exam topics. For fundamental priciples this
> period may be prologued up to 3 (or even 4) years total.

That's true.  Perhaps the security cert ought to be shorter, say 2 
years, while the networking and system administration certs could be 
valid for 4 years.

>>3) Q & A vs. test-labs.  I'd vote against automated Q & A tests.  I'd go
>>for a series of test sessions that judge the ability to handle specific
>>tasks.  The candidate ought to be given a challenge, e.g. to configure a
>>certain type of a firewall setup and then monitor it and write a report.
> 
> 
> This is a question of ballance. While test-labs provides more
> comprehensive and adequate testing, Q&A sets is easy to implement and
> pass. We can combine both: first, theoreticall questions; next,
> practice. On the other hand, Q&A results may be verified by any people
> who have a master copy, but practical test must be checked by
> expirienced human on prepared evaluation model and thus cannot be passed
> at just any test-center.

BSD has always been about quality and technical excellence, which is why 
I'd like to see the same attributes associated with the BSD 
certification process.  True, test-labs will be expensive to create and 
maintain, and the tests will take longer, but at least we will be able 
to convince people that it is worth the effort and that we are going for 
quality rather than quantity.  I think that strategy will prove to be 
the right one in the long run.

People generally associate a lot more value with something that is rare 
or scarce, so if there are only a few high-quality centers that prepare 
and test people, then the candidates will be more likely to consider 
such certification as truly valuable, even if they have to wait for 
their turn a few weeks.

>>4) Target level.  I'd go after the Tillman's "low-to-middle" category of
>>people who want a certificate because they need it to get a job.  They
>>will be more motivated to go into the trouble of certifying their skills.
> 
> 
> Average BSD user is fairly knowledgable. So test must be directed to
> people who knows how to exit from vi(1).

Sure.  No question about that.

-- 
Jacek Artymiak
jacek at devguide.net