writes: I run the IT Department at a relatively small organization. Each year we go through an annual audit. This year the IT portion of the audit was much more extensive than in the past and we were provided with a written report that contained various recommendations. One of them was that staff be required to take a vacation each year that was at least a week in duration.
The reason behind the recommendation is that it would ensure that people are adequately cross trained and that no particular employee would become so critical that the organization couldn't function without them for a week.
I should add that the auditing company specializes on small non-profits and I could see where heavy reliance on one person could be a problem for some places though I don't think it's a issue for us at the moment. Anyway, I'm in a position to implement that policy and am not really opposed to it but I am wondering what Slashdot's take on it is. Seems to me we could accomplish the same thing (and more) by making sure each staff person leaves the office for a week long training session each year.
writes: I work with data loggers of various types and I use perl to parse the information. I rely heavily on the str2time function to parse the timestamps. It works pretty well except that I was getting strange errors with a new logger file format I was processing.
It turns out that the logger was outputting dates like "9/24/11 10:27:30 AM". On my OSX development machine, perl and str2time (via timelocal) interpreted that date as September 24, 2011. On our linux production server, it was interpreted as September 24th, 1911.
Have we already forgotten that 2 digit years are no-no? Anybody else running into these kinds of problems?