OK, so I got curious about how the length of daylight changes over time. I tried getting Wolfram Alpha to figure it out for me, and I couldn't get it to do anything more complex than "sunset - sunrise" for the current day. So, I went to NOAA's site http://www.srrb.noaa.gov/highlights/sunrise/sunrise.html, followed some links, found some equations, and wrote a quick program that would dump the data so I could plot it.

For those of you who are interested, here is the approximate duration of daylight at about 45 degrees north latitude. The first graph is the length of  daylight, and the second is a graph of how much (absolute value) the day lengths are changing. As one would expect, the day length doesn't change at all right around each solstice, and changes the most near the equinoxes. Here at 45 degrees north, the days change by almost 3.5 minutes per day near the equinoxes.

In case you want to play with it yourself, here is the perl script lod.pl. You run it using a Perl interpreter, and pass it your latitude (in degrees). It generates two files names lengths.dat and diffs.dat, which are gnuplot compatible data files, one data point per line (e.g. plot 'diffs.dat').