Hi there,

it seems that this discussion circles around two assumptions:

*  Interactive system administration in general implies intense vi knowledge

*  Interactive system administration is the opposite of automated system
administration

I'd like to question both assumptions, with some arguments already brought
up in this discussion.

First of all, fixing things almost always means using an editor. No matter
whether it was the almost-junior administrator or the Ansible playbook that
messed up my /etc/fstab. It doesn't matter what this editor is; although
there might be certain requirements: If my system is already in a broken
state, the editor should run on the system. It might be vi from an
initramfs, but it might as well be something else running from a rescue
image booted via PXE. If I use automation, I will likely need proper yaml
support and refactoring, at least, if I want to be able to persist my fix
for future deployments.

Likewise, fixing things in an automated environment still requires failure
analysis. This will likely be an interactive login; except that changes
will not be made on the target system but in the automation to then
redeploy the system (entirely, ideally, since someone touched it).

The Linux knowledge in both cases overlaps a lot, the only question is
which editor will be available by that time. It might be vi in some cases,
but these cases will become fewer over time. Even the secure minimal system
will likely not contain any editor soon as these systems seem to be
immutable (hence, automatically build).

What does that mean for our candidate's jobs and our exams? Shall we cover
both cases by adding VS Code, Sublime, Atom? Especially when working on
various files, being proficient in any of these editors can be as cool as
being proficient in vi. But is this proficiency a requirement in LPIC-1? Or
is it the ability to edit files? It seems to be the latter, and it seems
that there is a common understanding that basic usage of 'fancy' editors is
considered a pre-requirement for LPIC-1 while at least a vimtutor amount of
vi knowledge doesn't hurt anyone but might safe some of us a lot of
trouble. I do see good reasons, both in this list and in the previous JTA,
to keep at least this amount of vi knowledge in LPIC-1. This will likely be
one or two weights. This should even enough to cover Bryans cases. Can we
justify more? We shall check during the next revision of LPIC-1 (due in
2021). Do you want me to add a future change consideration to the
objectives?

Oh, and none of this answers the general concern of the general LPIC exam
structure, but I wouldn't discuss this entirely on the matter of vi.

Fabian

-- 
Fabian Thorns <[email protected]> GPG: F1426B12
Director of Certification Development, Linux Professional Institute

On Wed, Aug 7, 2019 at 7:50 PM Bryan J Smith <[email protected]> wrote:

> Bear Giles <[email protected]> wrote:
>
>> What about bastion hosts? I know some allow transparent SSH forwarding
>> but aren't many configured to prohibit forwarding? You have to drop into a
>> shell on the bastion and establish connections from it.
>>
>
> Well, I do think we're getting beyond LPIC-1 and even 'common' setups.
>
> [OT] Anecdotal:  Although nothing is more frustrating than a Windows, let
> alone Azure, jump host.  I usually find the solution is to force a VMware
> or Windows admin to use even just 'nano' over such a connection, versus SSH
> jump (or even socks proxy).  They quickly 'change their tune' on the
> solution.
>
> Anselm Lingnau <[email protected]> wrote:
>
>> Vi is like chemotherapy. It's a nasty business. There are legitimate
>> situations where there's really no alternative to chemotherapy but that
>> doesn't mean everyone should have it all the time.
>>
>
> So why not just yank it then?  If we're going to go from 2nd highest to
> absolute lowest, of which there are so _few_ 'Weight: 1' items already, why
> not just yank it altogether?
>
> That's why started this thread, hence ...
>
> G. Matthew Rice <[email protected]> wrote:
>
>> Looking at the comments from various people, though, there are no zeros
>> for vi...so, unless people want to follow up on Bryan's break-fix vs
>> rebuild-fix topic (which is what I think he was trying to do in this
>> through)...
>>
> If it's just a matter of how much vi to test, our individual opinions
>> don't matter here.  The next JTA will take care of that.  Past JTAs have
>> even eliminated entire objectives.
>>
>
> Definitely JTA this.
>
> Because if interactive/visual text editing is going to go from being ...
> - Weighted 2nd highest, in the top 10 objectives
> To being ...
> - Weighted the absolute lowest, in the bottom 5 objectives
> Then why bother with it?
>
> And possibly then question even everything 'interactive' as secondary?
> We're certainly at that point for 2020+.
>
> - bjs
>
>
> _______________________________________________
> lpi-examdev mailing list
> [email protected]
> https://list.lpi.org/mailman/listinfo/lpi-examdev



-- 
Fabian Thorns <[email protected]> GPG: F1426B12
Director of Certification Development, Linux Professional Institute
_______________________________________________
lpi-examdev mailing list
[email protected]
https://list.lpi.org/mailman/listinfo/lpi-examdev

Reply via email to