On Friday 16 June 2006 05:00, John H Terpstra wrote:
> On Thursday 15 June 2006 14:45, Alan McKinnon wrote:
> > These are very valid points and in real life a Samba admin must
> > know Windows networking well. Hopefully the market will
> > understand this and know to look for evidence of competency in
> > both fields. But as I said elsewhere, I can't see how testing
> > those Windows competencies falls within LPI's mandate.
>
> How can we demonstrate that the LPI-3-Samba certification will have
> any value to the person to whom it is granted?  Will we at best be
> able to say: "Well, good old George knows how to configure Samba -
> but he may not be able to do anything of value with his ability."
>
> I would not like us to be a part of a process that can not assure a
> prospective employer that George will be of value.

Let me undercut this and clearly state what is underlying my position 
on this:

LPI does not have a mandate to test Windows competencies. That belongs 
to Microsoft. LPI tests Linux competencies.

I agree with you that in an ideal world we should be able to issue a 
cert that validates George's Samba and Windows skills, as they are 
used together. I just don't see how we can do the Windows part with 
any credibility. I also suspect that the hiring market will view with 
suspicion any cert that's not from the Microsoft stable and yet 
claims to validate Windows skills.

As an analogy, lets say I want to apply for a job as a programmer that 
will use enterprise product X. The requirements are knowledge of 
accounting, coding and product X. So I present a B.Comm degree 
majoring in CS, and an advanced level cert from vendor X. There is no 
one cert for all three things as they are quite different.

Samba is a bridging product between mostly incompatible worlds. I 
believe the best way to certify competency in these products is with 
two certs, one from each vendor/world

> > As a solution we could work off the assumption that an LPIC3
> > candidate already knows Windows networking, and create items
> > based on that assumption. If the candidate doesn't know his
> > Windows, he is unlikely to be able to answer the item correctly.
> > This is somewhat similar to the assumption in LPIC2 that the
> > candidate has/can successfully pass LPIC1.
>
> What level of due diligence is enough?
>
> Who can quantify the real value of LPI certification when at it the
> heart of it we make gross assumptions that may not have any
> substance?
>
> Can you give me an example Samba exam question that demonstates the
> point made?
>
> Can you define what it means to "know Windows"?
>
> If we set that as a prerequirement for LPI certification, how will
> LPI validate that the delegate has met the requirement?

I have no idea how to do these things at a fine level of detail, I'm 
still talking at a conceptual level, and I am by no means a Samba 
expert. I do see that it would be unwise for LPI to make statements 
about what constitutes a knowledge Windows networking person. The 
closest we can come is to select the most appropriate official cert 
from Microsoft and work from that point. 

This presupposes that Microsoft know what they are doing in this 
regard. Some will say that this isn't the case, but this is 
irrelevant, as MS sets the goal posts for their own product and we 
have to work with what is out there.

Suppose we select MS cert A as pre-requisite for LPIC3/Samba, there 
are two possible routes to follow from that point:

1. Require that the candidate achieve cert A before issuing an 
LPIC3/Samba cert, or
2. State that the Samba exam assumes that the candidate has mastered 
the objectives and knowledge level in cert A, as defined by Microsoft

All will be fine as long as cert A is a stable product, not subject to 
rapid change in it's objectives, and MS provides a stable mechanism 
that LPI (or anyone else) can use to verify that the candidate really 
did pass cert A. Most of this is beyond our control or influence.

> > The LPI promotional materials/web site can explicitly state "we
> > recommend that this certificate be supplemented with a suitable
> > certification in Windows networking as provided by that vendor".
>
> Sure! That should always be the case - even when basic Windows
> networking knowledge has been established. Microsoft Windows
> networking certification does little to help you to understand how
> MS Windows networking protocols function.
>
> Where in the current Windows Server 2003 and Windows XP
> Professional certification is NT4 technology fully covered? 
> Samba-3 is NT4-based. It knows enough ADS to permit Samba-3 to be
> an ADS client, but does the Linux admin know anything about either
> of these (NT4 and Non-Windows ADS clients)? If not, where does he
> go to obtain that knowledge?  Who certifies that knowledge today?

You know the answer to that much better than I do, so I'm not going to 
tell you how to do your job. I think we can agree though that there 
is only one entity that could certify that, and still be credible in 
the market - Microsoft itself. If they do, then all is well. If they 
don't, well then they should, and there is nothing we can go about 
that except encourage Microsoft to produce a certification program 
for their own product.

> So if LPI will not take essential knowledge seriously - who will? 
> Then why have a Samba exam?

I'm not saying that at all. We should take Samba very seriously. But 
it's inappropriate for us to test Windows competencies, much the same 
way as it's inappropriate for Alan Inc. to claim to test and certify 
John Terpstra's knowledge of Samba - I might be a pleasant, sincere 
guy who is good at his own job but I simply don't have the altitude 
and credibility to make that claim.

> I fully support LPI, and thus I am a little protective of its
> value.  If there is to be a Samba related exam, it has to have real
> value - like the rest of LPI certification.  There is no room for
> unfounded assumptions, and there is certainly no room for
> short-cuts that will undermine what LPI stands for.

Well I feel the same way about LPI certs. I don't want to see this one 
diluted with inappropriate objectives that we are not qualified to 
test.

However, I do believe we can find a suitable solution for this point, 
and I don't claim to have the answer at this point.

> I am sure you agree fully with this.  What I do not see from the
> thread so far, is the value and assurance that should be built into
> the LPIC3-Samba certificate.

With luck, this thread will clarify what that value and assurance 
should be. 

-- 
If only me, you and dead people understand hex, 
how many people understand hex?

Alan McKinnon
alan at linuxholdings dot co dot za
+27 82, double three seven, one nine three five
_______________________________________________
lpi-examdev mailing list
[email protected]
http://list.lpi.org/mailman/listinfo/lpi-examdev

Reply via email to