Re: Solaris experts wanted for SAGE project

[Pat Wilson <paw@northstar.dartmouth.edu>]

 Heather Stern <star@starshine.org> writes:
>  > [ This is my original ]
>  > > Folks -
>  >  
>  > > SAGE has been approached by a company that makes employment tests
>  > > (skills tests that hiring departments can give to applicants) for
>  > > help in finding authors and reviewers.  At the moment, they're
>  > > looking for someone to author a Solaris sysadmin test (or rather,
>  > > to write good Solaris questions; they worry about the actual
>  > > "constructing the test" parts).  The job does pay, and there are
>  > > contracts to sign.
>  
>  Personally since I think some of the most important traits of sysadmin'ing
>  are not OS based, I would consider it fair if the questions were fairly
>  generic, but the best answers were Solaris (or even nits about Solaris
>  versions) specific.

You're missing the point.  The tests which are under
consideration are tests that employers give to potential
employees.  Sort of like programming tests.  The current vision
is that SAGE would "seal" only *very specific* (e.g. Solaris
knowledge) tests.  No one's arguing that one should pick a
sysadmin based (solely) on the result of the test, but if HR
departments are looking for some indication that someone with "5 
years of Solaris experience" has some experience with Solaris,
then this should be able to indicate that.

Why is SAGE getting involved?  Bascially, because (a) these
things are going on anyway, and better to have _decent_ knowledge-based 
tests out there than ones which are terrible (and remember, HR
departments can't be counted on to know the difference), and (b)
it's one way of raising our profile a bit in the corporate world.

>  I don't favor multiple-choice only tests, unless all offer a "none of the
>  above" and a an "explaining in detail" pick.  I've often been annoyed at
>  a multiple choice quiz because the 'best answer' was still wrong in a
>  subtle way... and in sysadmin work, subtly wrong can easily lead to horrible
>  consequences.

And that's what the seal is for - it indicates that the questions
(which probably _will_ be multiple guess) are good.

>  Also I'd rather like to see several such Seals appear within a short
>  block of time of one another, so that no-one gets much mileage from trying
>  to claim they're the ONLY approved test (smudging us by implying their Seal
>  is exclusive).

I don't see how that's a problem, since it's _not_ exclusive,

Pat Wilson
paw@dartmouth.edu