On Tuesday 13 June 2006 17:35, Felipe Salum wrote: > Hi People. > > Just to start the thread for LPIC-3. > > It will be based in multiple-choice and fill in the blank questions > until now. I would like to ask what do you think about all LPIC-3 > exam series be based in lab exams ? > > I ask that because I see LPIC-3 as the top for LPI, and nothing > better than the top certification close with a lab exam and no more > paper based.
This is a topic worthy of discussion, even if only to come up with the right answer. But, hands-on testing for LPI is fraught with difficulties: First, which distro? If we stick to generic styles of putting a distro together, there are 5: RH-like, Debian-like, Gentoo, Slackware, LFS. And they are all very different at the level of fine detail. Admittedly, at LPIC-3 level it's reasonable to expect someone to be familiar with most, if not all, of them. Perhaps we could require that candidates use original sources of the apps being tested, but this is not feasible due to time constraints. Compiling samba, ldap, pam and friends is going to take at least an hour. Cost. Someone somewhere needs to put dedicated machines on the table. Marking scripts must be written and maintained, bandwidth and connectivity must be provided, *and* said bandwidth must be guaranteed. This is quite critical - the first three people in South Africa to write Novell's CLP all failed due to bandwidth downtime. All three report losing at least 20 minutes, none finished the test and needed less than 20 minutes to do so. All three just failed by a narrow margin. I firmly believe that exam labs don't scale to the breadth and depth that LPI aims for. Red Hat can get away with it as they control the servers and the OS being tested is an exactly defined thing, LPI has none of these luxuries. Another possibility is a practical project instead of a lab, much the same thing as a project one does at university or college but on a smaller scale. I'm thinking in terms of the original Java Architect cert from Sun here. Again, this doesn't scale well as it requires multiple people to mark the results, or it must be scripted. Scripting has to take account of every edge case and test for them. I'm not pouring cold water on the idea, I really would like to know if this can be done, and if not, exactly why not. -- 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
