Mark Clarke <[email protected]> wrote: > If its a piece of code, bash script, config file etc perhaps we could > run it through a test suite and if its all green you good to go if it > fails then you should try again.
Unless there's a condition the test suite cannot accommodate and/or the candidate requests a manual grade. Sometimes things can be solved in more than one way. This is what happens with Red Hat's performance-based exams. The automated graders cannot account for many things. I know, I requested a manual grade back in 2003 and Russel himself found both a solution for one problem that their automated grader didn't account for, and a false negative which I should have received credit for, increasing my score to over 90%. And then I was introduced to the other side, and how much it can happen. As has been pointed out its rather a small subset of problems that can be > tested practically. > Or programmatically, and even then there are limits. Automatic graders just cut down on the requirement for manual review, but it doesn't remove it. In fact, the automatic grader should be written to suggest manual grading too. Those expert systems have matured, but they still require and even don't always identify manual grading. > The student could log into a web site register and get the assignment > and the clock start ticking. 24 hours to submit etc. On completion > he/she uploads the assignment/practical and the machines handle the > rest. Its not a 0 cost idea unfortunately but it is more scalable than > the vm approach. It's cheaper to lower the human involvement, but it doesn't eliminate it. It also requires development on its own. > Its would kind of be like your end-of-year mark being > 75% exam and 25% of assignment submitted during the year. > If the membership thing takes off there could be an required period of > practical experience signed off by your employer to acquire a 2nd tier > membership. This is how doctors, lawyers and accountants work with their > professional bodies. > And licensed, professional engineers in the Anglo-American systems. ;) E.g., the British ended the French-Prussian practice of engineering being a graduate degree and 2 year internship, squashing it into a 4-5 year Bachelor's with an underclass (first 2 years) of pre-engineering core, instead of general education. This allowed engineering internships to increase to 4-5 years, as the British (and greater Commonwealth, as well as American and other systems) felt 2 years was not enough of an internship. I.e., most traditional licensure is 7-8 years of education and 2 years of internship, but engineering is 4-5 years of education with 4 years of internship. The 4 years of experience has carried over to select licensure certification and even licensure as well. - bjs DISCLAIMER: I am a traditionally educated, American engineer (ABET accredited Electrical and Computer Engineering graduate). -- 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
