How much energy was consumed when we turn on/off light? My parents told me to turn off the light when I am not using it. But I remember my physics teacher told me that the action of turning on/off a light can cause huge energy. I am wondering how much is that? Someone said the energy caused by this action are enough to keep the light on for 15 minutes.
If we require a configuration here, let's assume we use a T8 light tube, with 220V voltage from the house and 30W of the tube.
 A: From an answer by  Steve Selkowitz of Lawrence Berkeley National Labs
(I'm quoting it in full here because the site has changed and I had to pull this from an old mailing list archive)

Fluorescent Lighting - Should I turn the lights off?
There have been two very resilient energy myths that have dissuaded
  people from turning off fluorescent lamps. These myths are hard to
  dispel because inside each myth is a kernel of fact.
Inrush myth: I don't turn off fluorescent lights because the inrush
  energy when I turn the lights on is more than the energy I saved by
  turning them off.
Inrush fact: When fluorescent fixtures are turned on there is a
  MOMENTARY inrush of current. This inrush is 5 times greater than
  normal operating current for magnetic ballasts and can be as high as
  40 times greater than normal operating current for electronic
  ballasts. This inrush lasts for 1/10th of a second or less. Thus the
  energy content of the inrush event is comparable 5 seconds or less of
  normal operation. Thus, turning off the lights for more than 5 seconds
  will save more energy than leaving them on even when including the
  inrush current.

A: First thing, it depends on the type of light used.
For an incandescent bulb, a bit of energy is used up while turning it on; but not much.
On the other hand, tubelights have inductors (choke coils) in them. These come into play when switching the tubelight on and off only. This is because inductors oppose change in current, when there is a steady current, they act like wires (or more accurately, like resistors). As they oppose change in current, they do absorb and emit energy. 
When a tubelight is turned on, the choke absorbs some energy, whereas when you turn it off, the choke releases the same energy. There is almost no net energy transfer here (Energy "held" by an inductor's magnetic field is $\frac{1}{2}Li^2$). I say almost, as the heat lost through the inductor while the tubelight starts is a little bit different, and there are other small energy losses.
If we consider an "energy loss" as any energy used up during a period when no useful output (light) is given out by the tubelight, then at the time when the tubelight is starting, there is a small heat loss through the choke resistance. A smaller loss is there when the tubelight is turned off.
So really, I don't see where your value of 15 minutes came from. 15 seconds or 15 microseconds is more appropriate (the exact value would depend upon the parameters of the light, and what you interpret as an "energy loss").
A: It seems from what I have read on a couple of sites that the 15 minute recommendation refers mainly to CFLs. Incandescent light bulbs are very inefficient and you will always do best to turn them off. Same with halogen. However, with CFLs the bulbs life is shortened by the amount of times it is turned on and off. Therefore if you figure in the cost of replacing the bulb, that is where they come up with the recommendation of 15 minutes. LED's however are not affected by the amount of times it is turned on and off, so you can turn them off as much as you like.
