LPI exams - Ubuntu Certification

[Alan McKinnon]

On Wed, 2006-07-05 at 14:05 +0100, Daniel Carrera wrote:
> On Wed, 2006-05-07 at 14:41 +0200, Alan McKinnon wrote:
> > You can relax, every "sample exam" out there (bar one) is seriously
> > broken and has very little relevance to the real exam.
> 
> Ok, but then why are they linked to from the LPI website? And how is one
> supposed to know when one is ready if the samples are broken? That
> actually makes me nervous in a different way.

The website lists some community-maintained links as a service to
candidates like yourself, and LPI takes no responsibility for the
content of the linked sites - much like Canonical doesn't take
responsibility for things people put on the wiki. It's a thankless task
deciding if a link is still relevant, and any choice is bound to upset
someone somewhere, so webmasters the world over don't do it.

The definitive reference, and the only valid one, is the list of
objectives at www.lpi.org. All other sites are only the opinion of their
respective authors and should be viewed as such.

> > None of them are psychometrically valid,
> 
> Could you explain what psychometrically valid means, and how you decide
> that a question is valid?

I don't the psychometrician does that. There's a standard practice for
this, first a question is reviewed by subject matter experts, then
piloted on real exams as a non-scoring question. It has to behave in a
statistically meaningful way before it gets moved into the permanent
pool of exam questions

> > The exception is the samples at praxis.de which are actually quite good
> 
> Could you give me a link to those? Do they have any in English?

http://www.linux-praxis.de/lpisim/

> > ISA has been dropped from the LPI Objectives, and replaced with 
> > PCI/SATA/USB in more detail than it was before.
> 
> Good... but here's an example where the practise exam makes you study
> for the wrong thing.

Granted, but there's not much LPI can do about outdated web pages on the
web at large. I also learned this one the hard way - trust only the most
recent list at www.lpi.org

> > You've just made a classic flaw :-) Such hardware is indeed obsolete if you want
> > to buy a machine today. But they hang around for years after the vendors stop 
> > selling them, and admins need to know how the hardware works.
> 
> No, I didn't. You just missed the point. I was comparing the age of the
> hardware with the software. If we're going to test for knowledge of
> hardware that is old enough to have an ISA card then we should test for
> the software that such hardware may be running which includes at least
> one distribution using the 2.2 kernel. If we're going to test for 7
> year-old hardware, then we should consider the possibility of a 1.5
> year-old Linux distribution (Debian Woody, using the 2.2 kernel). The
> end result is that there is no answer for this question because it
> depends on what kernel you are running.
> 
> I think I've shown that I understand the issue with ISA cards better
> than my answer to that sample question is likely to indicate.

I'll have to double-check that and get back to you, but I think kernel
2.2 was dropped from the objectives at the same time as ISA.
> 
> 
> > So two, maybe three such questions are a good thing. 10 of them would be bad.
> 
> How many questions in the exam?

65 scoring questions. English on-line exams at a computer based vendor
contain 15 or 20 pilot questions that are being tested for validity. You
are not told which ones these are, and you are allocated an extra 30
minutes time.

Paper-based exams are 65 questions, 90 minutes

> > > In some questions the exam seems to be measuring the wrong things.
> > 
> > Seems more like the sample exams are deeply flawed :-)
> 
> That is bot comforting and unsettling at the same time.

:-)

The best third party guides I have yet seen are in book form - ExamCram
LPIC-1 by Ross Brunson, and LPIC-1 by Roderick Smith. The books are
packed away at home where I can't get to the ISBM numbers, but both are
available at Amazon.

Both authors fully groked what the exam is all about and the standard
required, and explain it in their books in detail, without giving
anything away.

> > If you find some error in the actual exam you write, there is a feedback 
> > section at the end where you can enter your comments and improve it for the next 
> > candidate.
> 
> If that happens, and I show that I know what I'm talking about in my
> comment, will I get the point for that question?

If you feel your results are not fair, then you are welcome to mail
info at lpi.org and take it up there. Obviously, you should not describe
your reasoning in an open mail before you get a public key for the
correct person on the other end.

> > Do keep in mind that there is no known method to measure what we techies would like
> > an exam to measure. We want to know if someone knows Linux well, but you can't measure
> > that - it's intangible and exists only in your mind. So instead all exams measure some 
> > measurable side-effect and rely on that to correlate to the real thing. 
> 
> A problem with this type of exam in general is that it's not going to
> measure my ability to solve a real-life problem. For example, it doesn't
> matter if I know the -name flag for 'find'; I can get a similar result
> piping through grep. Any given problem has multiple valid solutions.
> Sadly, only direct observation can show his sort of knowledge.

Bingo! you hit the nail on the head. No exam can test that, it has to be
done through detailed observation over an extended period of time.

Not even the RHCE does it, and I speak from experience - I wrote mine 10
days ago. RHCE and LPI are both excellent exams, I personally know most
of the people in this country that have one or both certificates, and I
can assure you there are very few false positives in the results. The
RHCE is a very good test, but it definitely doesn't do what most of it's
fan-boys out there claim it does. Both exams are very evenly matched
IMNSHO, but measure competency in very different ways.

alan