From: ptkwt@... (Phil Tomson) Date: 2005-11-04T03:52:09+09:00 Subject: Re: Ruby Certification In article , wrote: > >On Nov 3, 2005, at 11:41 AM, Obie Fernandez wrote: >> Assuming we're on the verge of a big boom in Ruby jobs, shouldn't >> there be a place where employers can verify a candidates reputation? >> I'm thinking a testimonial or trust-based system of some sort. > >Speaking as someone who has done hiring in a area with lots >of "certification" (Cisco Networking). I can say it means nothing >when I see it on a resume. If I want to know if someone understands >BGP in an interview, I can ask one or two questions and within five >minutes have a pretty good handle on their level of expertise. >I found that there was very little correlation between what I could >discern from someone in person to what was on their resume relative >to "certification". > > >> When the market for Java programmers exploded there were hug numbers >> of highly (un)qualified people getting hired to do Java and causing >> all sorts of grief to everyone involved. Lots of them had (or claimed >> to have) certification. > >Sounds like bad hiring practices to me. I think *technical* knowledge >is one of the *easiest* things to figure out in an interview. What I >find really hard to figure out is if the person is going to have a good >attitude, work well with others, be good with customers, and so on. > I would even maintain that a technical interview is not a very good way to guauge someone's technical expertise and talent. There are very talented people who lock-up in that situation. And it is an artificial situation: In real life how often do you find yourself locked in a room without any reference books and no internet access? It would be better to consider a candidate's community involvment including open source code the candidate has produced. Lacking that, I'd like to see more 'real-world' technical interviews where the interviewer brings in a laptop (with wifi access, a compiler or interpretter, program editors, IDEs, etc), some reference books and outlines a problem s/he would like to have the interviewee solve. The interiewer then leaves the room for 2 or 3 (maybe more) hours and checks back later to see the code written by the interviewee. Phil