There is nothing wrong with building experimental circuits, as these lead to innovation, and really should not be discouraged. One thing I would add to the circuit though is to include a 2 Ohm 10 or 15 Watt resistor in the supply line.
I see the plans as an excellent tool for educational purposes for exposing kids to many of life's lessons. I am glad it was posted.
I salvaged a number of "Emergency Ballasts," removed the charging aspects of them, as well as the boiled out nicads, and most worked well straight off 12 volts, but a few died quick deaths, but the replacement transistors had enough design difference from the originals, that a 2 volt resistor of significant wattage was needed. These are typically a double ended system. But are only rated at 70 Lumens on a 4 foot tube. But they work well on the 20 Watt tubes down to some undercabinet T12 types.
The emergency ballasts lighting circuit is essentially upsized from the Zetex circuit I have mentioned several times. Which BTW, is not much more than a freerunning inverter supply.
"Q" is in an oversimplified way- how sharp can you tune your tuned circuit?- which is what you actually have with florescent lighting. Which also explains in a nutshell most of the problems people have described so far.
BTW, Lead in CRT's is actually the clarifying agent for glass. If your monitor is a CRT, you are reading this through fine lead crystal. Between 2 and 6 per cent content. Anything higher leads to weak glass and an implossion hazard potential. Most of what you read on CRT's is circular hype by someone who meant well. Nipping the end seal where the pins are is a reasonably "safer" method of dealing with the static charge they may hold, as well as eliminating the implosion hazard, and there is no mercury in a CRT.
On matters of noise: In the audible spectrum, been covered well. In the RF range- most electronic ballast circuits will emit Amplitude Modulated noise. It's inescabale Put an AM pocket radio next to one, and you should hear a whine. It is a matter of where you locate your operating frequency for your lighting circuit that may or may not interfere with some modes of communications. GPS is not an AM signal, so it will be immune. Longer wavelength HF can be where you will hear your RF interference- when my batteries get low, tunning below 5 MHZ is tough. HF gear is in an oversimplified description an AM mode reciever.
For migrane sufferers- florescent lighting may not be the best choice, no matter how they are driven.