On 25/09/2016 12:58, [email protected] wrote: Hi Kenneth,
Thanks for the info. See my additions below. > Hi all, > > 3 topics. Spain, biased results and PBT. > > > Spain: > ------ > > Most courses are paid by the state. Every year I teach 8-10 courses like > these, and maybe 1 private course (way cheaper than in Germany. 80 hours > LPIC1 costs 1400 €, including both tests). > > People gets free training + 1 exam voucher in regions with high > unenployment rates (even people with a job), and just a shorter course > in the rest of the country (40-60 hours in Barcelona/Madrid/Valencia, no > free vouchers, 64-80 hours in Andalucía/Aragón/Galicia + free voucher). > > So, many of our students are unemployed people trying to switch from > other IT branches. They are around 50%. The other half are IT people who > need to improve their Linux knowledge. > We are looking into accessing funds here in South Africa but we have a whole different set of challenges. Most business rely on the private sector. But I do think this is the way the world is moving - the people paying for the training are not the people receiving the training. > > > Biased test results: > -------------------------- > > I wanted to share something else about our tests. Also personal > experiences :) > > One of my worst students ever got a 100% score in LPIC1, 100%!!! Not > 50%+1 or 75%, he had no mistakes at all!!!! Believe me, you'd never hire > him if you knew his real knowledge (that's why more companies are doing > PBTs to candidates nowadays). > > This guy was unable to install stuff without breaking the system, wasn't > good understanding basic concepts, he had plenty of mistakes in the > commandline and he was usually unable to end exercises in a 80h LPIC1 > training, and yet he scored 100% on LPIC-101!!!! I know that there has always a debate about Universities versus Technical colleges at least here in SA. Technical colleges produce people who can hit the ground running, as far as business is concerned, but universities produce people who had a better theoretical background and after a few months would over take their technical college counterparts. After a few years the university grads would be advancing up the management chain while those with practical training where mostly still doing what they had been taught. This is the challenge of education -> teach people to think or teach them to do? I do think LPI produces people who are more aware of what can be done with Linux systems. But the world is changing. With the increasing salaries people are demanding and the increasing mobility in the work force businesses now want people to come with the skills already and are not prepared to carry the cost of training people up like in the old days. The social contract between entry level employees and employers has broken down - this is one of the reasons for high youth unemployment. I know we won't look at entry level candidates for this very reason. Not sure what people in other countries experiences are. > How can that be possible? he just memorized testking Q&A better than > other classmates. There was at least 4 good students (2 amazing) in the > same course, and none went above 90%. > > Are we really ensuring to his next employer that he knows about Linux? > Please, don't, I hope he's not working in anything related with Linux, > or that company will never trust LPIC again. > > Then, a geek working for the government, always showing me cool stuff > and making a lot of R&D, always one of the 1st in finish the practices, > etc.... took 201 exam and he failed to pass it by just 5 points, because > he didn't cheated with testing and he felt confident on his knowledge. > Tricky modifiers and collectd made the rest. > > No, our evaluation system isn't working at all. We can't feel happy and > proud with a system producing these biased results, no matter how wide > the scope of the test was, our results aren't reflecting knowledge. > > If you can cheat it and pass it without real knowledge, the evaluation > system is broken. > > If we want to keep using Q&A evaluations, we really need to increase our > pool to thousands of questions and renew them more often. Otherwise > anyone paying 35$ will be able to pass our tests. > > I.E. Microsoft requires partners to have an ammount of certified people > to become a platinum partner and get extra licensing discounts. SMB > certify their administrative & finnancial staff by buying que answers, > so they get that extra discount. > > A certification exists to ensure companies that their employees are > ready for their duties, and we can't hire people just because they have > LPIC certifications, due to testking/pass4sure and so on. > > PS: I saw LPIC303 Q&A from one of these providers. They had loads of > "OCR" mistakes, like -- converted in -. So it looks like some partner is > making extra cash by selling our tests. Contact me in private on how to > try to detect who's it. > > ----------------------- > > PBTs: > ----- > > 1st, I don't pretend to end tests anylonger. I agree they let us > evaluate a wider range of topics in less time than a PBT, but it isn't > enough to ensure that you're ready to perform these tasks, and that's > what companies expect from certified technicians. > > I don't understand how can a PBT be so expensive (well, I understand it, > but would never pay those ammounts for it!). You just need from 1 to 3 > small VMs for up to 3 hours and a script. You don't even need a desktop, > just SSH in a minimal installation, and let everything else to be done > by the student. In most practices you could even use docker, easing the > evaluation script no matter what they do with the poor SSH. The only challenge I hear raised may be poor Internet connectivity in some part of the world like in Africa. I don't think this is a serious problem anymore but there will always be excpetions. > > If properly automated it could be a cloud provider starting a bunch of > VMs and giving you login/pass/IP on the exam page, and a "finish" button > to trigger the evaluation script. So, close to zero IT maintenance cost, > not those crazy daily ammounts! > > Running 1x2GB VM for LPIC1, 2x2GB for LPIC2 and 2-3 for LPIC3. Most > cloud providers offer these kind of VMs for less than 0.07$/ hour, and > probably some provider would offer us them for free to publitize their > services among our students. That's less than 1$ for a 4 hour test with > 3 VMs. Delivering paper tests worldwide is more expensive. > > Then integration with LPIc tests can be coded in less than 2 weeks, no > matter how messy our website is. The rest goes for the scripting > required to evaluate students, which isn't very tricky because you > perform quite simple operaitons in these PBTs. Once you have a > framework, it would take less than 1 hour of coding for each topic. > > Yes, we need to do some work beforehand, and I don't know how many > people takes these tests yearly, but it would make sense to analyze what > tests will became profitable earlier. My bet, begin with test+PBT in > LPIC1 and 2, then keep improving with best selling LPIC3s. > > As a side effect, we could perform just 1 test for LPIC1 and 1 for > LPIC2, because we could evaluate all the contents in less than 4 hours > (2 hours Q&A + 2 PBT). > > A lot of people would love this, because passing 1 test will get you > certified, not like nowadays, where many of my students have only 101 or > 201. Now it costs around 2x200+VAT € to get LPIC1 or 2. I bet a lot > more people would pay 300€+VAT for a full-level test instead of current > 2x200€. > > Getting RHCSA costs around 450€+VAT, so, we could offer equivalent > certification degrees at a lower price, thus (I hope) increasing sales. > > You can build a system like this with less than 2 weeks of a web > developer and 2 weeks more for a sysadmin preparing all LPIC1 & 2 > topics, then all training partners testing it with free vouchers :-) > > Yes, it's never a 1 month project, but starting 2 VMs and running a > script isn't that tricky. > > > IMHO, a test like that would be way more attractive for self-paced > students or people trying to certify their current knowledge and not > willing to take a full course. PUE wondered about doing "launching > courses" of just 2-3 days for these people, so there's still room to > make business. > > > > Regards, > > Kenneth > > A 2016-09-25 08:59, Mark Clarke escrigué: >> Hi all, >> >> I am not responding to any specific contribution because there are many >> but here are some opinions I wish to contribute in no particular order >> of importance. (I think we need a brainstorming sessions to get all >> contributions and then ponder them for a while to see what we make of >> it >> all.) >> >> * The perception about practical vs mcq based exams will differ from >> region to region, >> >> * I understand the pros/cons of both. i.e scalability, cost, etc. As >> pointed out - ultimately its the perceived value that will win, In SA >> the perception is practical is better. >> >> * We could perhaps address a lot of this perception with a good >> marketing campaign. It will at least partially address the issue if not >> completely, >> >> * A lot of vendors can use training as a loss leader - they are making >> their money from selling their products and training needs to be >> provided. We are not in that position unless we come up with other >> services >> >> * Personally I think people perceive practical exams to be better >> because they can pass the mcq exam with a combination of knowledge and >> guess work but still feel they have no idea how to do it practically. >> i.e put them in front of a server and say set up postfix they won't >> know how to begin. (The practical based ones are very simple set ups - >> just apt-get install etc with some useradd type stuff.). >> >> * The perception may depends on the level of experience of the person >> who is doing the test. If you already have years of experience then the >> practical test is simple and the LPI objectives cover a wider, richer >> range of topics. In SA most people doing the test are trying to enter >> the job market, climb the ladder, just getting into sysadmin etc - so >> its all new. Its possible to pass a mcq exam without ever having >> touched >> a computer but you can never do that with a practical exam. For me this >> talks to who we think is doing the exam - someone who wants to prove >> their existing skills or people looking to gain new skills. Most people >> with skill don't feel the need to prove they have them with a cert >> >> * One of the big issues is the cost of doing the cert - others require >> 1 >> exam and you get a cert - we require 2 exams and it costs more than >> other providers, >> >> * A lot of the bigger training provides are just middle-men whose >> business goes something like this. Is there demand for a course? Is >> there a provider I can get certs and training material from? Are there >> trainers? Put bums on seats -> profit! They will not create their own >> material nor do a lot of marketing to get a cert going. I have been >> told >> they need training material so if anyone complains about the training >> and wants their money back - they can say its the official course >> material so they "off-the-hook". If the instructor has to create the >> material - they have no way of knowing if the client is right about the >> poor training content. >> >> --------- >> >> Some wild ideas and observations to add to the debate or be dismissed >> out of hand :) >> >> * We could require the candidate to verify their number of years of >> experience. As an employer I know the challenge in employing IT people >> as compared to other professions such as accountants, engineers etc is >> you kind of know what you getting once someone has had a practical >> period of training and their is a professional body to ensure standard. >> An engineer with 5 years experience etc. >> >> *All over the world students are protesting the high cost of education >> and universities look like they are doomed as an education model. >> Don't >> get me wrong I think they have a vital role to play but there does >> appear to be a crisis of education. This represents an opportunity and >> threat - we need to understand where things are going. Our business >> model, as certification providers and training providers may be doomed >> or destined for greatness :) >> >> * People are going the self-study route. Face-to-face training is going >> to trend downwards. This may not be relevant to LPI directly, being a >> certification body but its going to affect training partners. How do we >> deliver training? >> >> * In our market some vendors struggle to get their training off the >> ground due to lack of demand. The bigger training partners with >> campuses >> all over the country need classes of a minimum size to make training >> viable. If this cannot be attained the courses are never run. Its >> chicken-egg scenario for a vendors. Business doesn't adopt the product >> because there is no skill and people don't skill up because there is no >> demand from employers. Hence the vendors are finding ways to address >> this by changing their approach and taking it all in-house again. i.e >> not using the training partner model. The threat to LPI is we cannot >> afford to make training "free" - this is part of the global movement >> for >> people wanting "free" education. (Don't want to start a debate about >> free education please - just saying its a trend. Of course someone is >> paying for it somewhere.) >> >> * Certs are proliferating and bodies are creating more and more as the >> business model is simple for most certification providers: more certs = >> more money. This is leading to fragmentation, market confusion, and >> smaller and smaller piece of the pie. Those who will win are those with >> a strong brand. Luckily we got that covered and a programme is already >> in place. >> >> *Training material is becoming harder to provider. The technology >> changes are too rapid for O'reilly, Apress, etc to provide books and >> still make money. It has to become all epub based to lower costs of >> distributions and to make updating less costly. >> >> >> cheers >> >> -- >> Mark Clarke >> 📱 +2711-781 8014 >> 🌠www.JumpingBean.co.za >> >> _______________________________________________ >> lpi-examdev mailing list >> [email protected] >> http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev > _______________________________________________ > lpi-examdev mailing list > [email protected] > http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev -- Mark Clarke 📱 +2711-781 8014 🌍 www.JumpingBean.co.za _______________________________________________ lpi-examdev mailing list [email protected] http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev
