> > Well as a fairly independent person in this matter, i will volunteer
> > to coordinate this. Unless there are any objections - i already
>
> TIMTOWTDI kind of screws things up. Different people will code in
> different styles. How can you evaluate this?
I don't think it's a huge problem. For a start, certification can be
multiple choice, which eliminates the need to deal with correct answers in
esoteric style. (shamless plug for my amazing perltest project -
http://www.snowdrift.org/computers/perl/pt/)
Secondly, there is absolutely not reason why the certification can't agree
on a compulsory style. This is what happens in driving tests. There is more
than one way to turn the steering wheel, from suicide spinner to hand over
hand, but the driving certification declares that shuffling (or whatever
it's called) is the 'correct way'. No-one actually believes that you're a
dangerous driver if you use one of the other methods in some situations, and
it all works OK.
Thirdly, IMO certification is more about establishing that the candidate
doesn't do stupid things than that they are very clever. The driving test
only seeks to establish that you've read the highway code, and can get from
A to B without screwing up.
Likewise Perl certification should seek to show that a candidate has RTFM'd
and get get from A to B without screwing up, for various values of A and B .
>
> If I see a sensible plan for certification, this sounds sensible, but
> consider what most people think of eg. MCSEs.
Most _people_ consider MCSEs a useful way of gauge a minimum standard of
knowledge in prospective employees. Most professional system engineers and
programers may feel differently.
