Sixteen Volts

I had already forgotten the site Hacknot and its essays. Coming back to it, I find "The A to Z of Programmer Predilections" to be quite funny, especially the parts "C++ Colin", "Open Source Oliver" and "X-Files Xavier". "Interview With The Sociopath" was also funny, especially the part

Companies intent upon creating the impression that they really care about the quality of their people will give potential candidates a hokey COMP101-level programming problem to solve prior to granting them an audience. The solution provided is then dissected carefully and assessed according to criteria that the candidate was not made aware of at the time the assignment was set. Ridiculous extrapolations and inferences about the author's general programming ability are then made based upon the given code sample.

The beauty of this technique is that because the problem has been offered context-free, the candidate has no idea what design forces should influence their solution. They don't know what importance to assign to non-functional criteria such as performance, extensibility, genericity and memory consumption. The weight of these factors might significantly influence the form of the solution. By withholding them, and because the factors are often in conflict with each other, it is impossible for the candidate to submit a solution that is correct. Simply change the criteria for evaluation to the opposite of whatever qualities their solution actually contains.

For example, if their solution is readily extensible, claim that it is too complex. If they have favored clarity over efficiency, criticize their solution for its verbosity and memory footprint. If they have provided you only with code, select documentation-level and handover-readiness as the critera-du-jour - question the absence of release notes.

Posted by Ilkka Kokkarinen on Wednesday, May 10, 2006 at 11:48 AM