Mark Clarke <[email protected]> wrote: > I have read most of the responses to this and think it worth adding my > opinion, your view on this may vary :) > It is definitely the case that the employers thinks hands-on is better > and therefore prospective exam takers think its better too. It's very > hard to argue against the superiority of hands-on experience when > intuitively it sounds better than multiple choice questions. To be > honest. Testing via multiple choice is seen as inferior way to test by a > lot of people as compared to a written test etc. I don't wish to start a > debate about the merits of this but it is a perception that is wide > spread. >
I've been purposely avoiding the 'perception' debate in this, trying to focus on the 'how LPI could?' discussion. But I'll go ahead and throw my very opinionated (apologies in advance) view into the ring, especially being a lead and even manager in charge of hiring over 100 people in the last decade in Red Hat heavy shops. I look for a LPIC-1 with a RHCSA at a minimum.. I said this when I was at Red Hat for nearly a decade, and both Russell and Hacker (if you know who they are in Red Hat) know my view.** Hands-on: It's like a "hands-on interview," objectives with tasks on a real-system. Doesn't mean everything a company wants to cover would be covered, but it means if Red Hat interviewed a candidate for you, to those objectives. It says a lot, that you can survive 'hands on.' The problem? Hands-on exams are very, very 'narrow' in coverage. Example: I run into this constantly -- LPI Objective 102.3 Manage shared libraries. I have run into many, many, very experienced Linux sysadmins that have never messed with the system-wide ldconfig and cache. It'll bite you in all sorts of ways too, without even suggesting it's related too. When I have taught to the LPI objectives, done mentoring or even done Red Hat partner training (especially when partners' -- the one I'm training -- partners' own libraries become 'an issue'!), I point out several LPI objectives that don't get coverage in RHCSA or even RHCE, that one should know. And that's just one example. There are about a dozen I harp on when doing Linux training or mentoring. > I would suggest that its not an either or approach. We could have a part > that is multiple choice and a practical part. The practical part doesn't > have to be under exam conditions. It could be a task like write a bash > script that does x or some other assignemtn. The student is given 2 days > to do the task and submit the script/assignement and the testing can be > automated. > Now that's a good suggestion. Doesn't address the 'security' aspect, but it's a 'practical' one. Now comes the 'cost.' What would be the 'cost' of the review? I don't see LPI doing it at the normal rate. It's going to require manual review, and possibly 2nd or even 3rd review, especially if 'contested.' Now we're looking at typical 'licensure' level fees. But it could be worth it in the case of LPIC-3, even maybe a LPIC-2, after testing with LPIC-3. But more on that below ... Who cares if the person googles or gets help. Isn't this what people do > in real life? It is mainly to address the markets perception and it will > force the student to at least be a little familiar with hands-on work. I > know it may be a little tougher to make some small variation in the > script requirements to prevent copy/paste type responses but it might be > worth the effort. > Depends. Red Hat has issues in some locales in many different ways on many different things (not just Training). And they do their best to secure the integrity of their offerings. I don't want to go into depth other than to say that even 'performance-based' is not always 'fool-proof,' especially when trusting external parties. LPI also has the 'Membership' aspect that has been announced. Would members be willing to peer review as part of their duties in membership, possibly if it means free membership? Maybe a level of 'senior member' or 'proctor member' where it's, again, 'peer enforced'? This too is kinda how 'licensure' works as well. Maybe that's where LPI should head? I don't know, but I'm open to a good idea when I see one. Again, I have 'no power' but I've been unofficially around LPI since '04 (as a general PITA and overly opinionated SoB to the point you can see the origins of my scary, American Libertarianism too), but if there are good ideas to be had, we can try them. In fact, if and when 'Membership' becomes a reality, how many would be interested in forming 'boards' for such 'peer enforcement' such as reviewing 300 level or maybe even 200 level, 'script review'? Again, I'm speaking 100% as a peer here, and have 'no power.' I often play 'devil's advocate' on this list, annoyingly so (please don't take anything personally), and have been around a dozen different programs, including seeing several 'from the inside.' BTW, if I ever get 'too annoying,' feel free to ask Matt Rice to kick my butt of the list for a bit. ;) - bjs P.S. Again, to re-hash my prior (which was largely 'voice-to-text' on my phone), Red Hat can subsidize a great amount of expense, and did so to avoid the 'paper MCSE' from the get-go, to build a vendor-centric program. They bundled training in the preceding 4 days, before the Friday exams, to off-set costs ... which are still in excess of $1K/day in examination. But even this hasn't been viable for Red Hat in many aspects over time, and people complain about the lack of availability (which is a long debate). They've gone with on-line training 'subscriptions' and 'kiosks' as well, to off-set costs and access too. -- Bryan J Smith - http://www.linkedin.com/in/bjsmith E-mail: b.j.smith at ieee.org or me at bjsmith.me
_______________________________________________ lpi-examdev mailing list [email protected] https://list.lpi.org/mailman/listinfo/lpi-examdev
