Stay cool, let it run. That is pretty much a conclusion after having studied MTBF-figures, theory about heat and transistors and hard drive manufacturers fancy broschures as well.
In short: As long as you keep all critical temperatures well below the limits the computer will feel very good running continuesly.
The worst stress, still except overheating, for the mobo and processor and every other component soldered to the mobo is major temperature variations. Major temp variances are only to be seen in machines turned on and off. Leave it off and nothing will happen...
Leave it on and it will be fine.
MTBF (Mean time between failures) are theoretical assumptions made taking all components in consideration, with the temperature over time never getting close to the limits, with normal humidity, the power turned on, a certain duty cycle and so on. 70.000 hours (equals to 2916 days or 8 years) is pretty much a nonsense figure, still the only one we have. If your CDRW is of average quality in every aspect, used according to the actual duty cycle (which I guess isn't told) there should be no problem to let the elecronics inside it idle for years. Then, do you really think you will use that component in five years? Or ten?
Equally the worst stress moments for a HDD is... yes, turning it on and off. Somehow we have to compromise between the factors power consumtion and component lifetime.
As long as your dad can afford to pay the electrcal bills you can let it all run. Check the processor, mobo and harddrive temperatures some day when you have some time left, make sure the computer has been running for hours doing it's normal work first. If the temperatures are okey there are no worries.
Finally: Things get hot in there. From time to time I would advice everybody to step inside the puter chassi, remove all dust and shit, unplug and replug (?) all cables, check cables for detoriation (or whatever the correct word is) and make sure air can flow as supposed thru the housing.
Hope this helps somewhat.