The reason video is so effective is that while books merely explain information, video forces you to DEMONSTRATE why you know something.

However, books allow the recipient to digest information at her own speed, e.g. she can stop for some thinking or calculations or drawing.

I would also quibble your contrast between 'useless' library research and useful original research.

On one hand, library research can be useful for falsification (uncovering errors, misinterpretations of data by the original researchers), for digging auxiliary data for a new discovery (say, think of older photographs of a just discovered asteroid that allow a more precise orbit calculation), and sometimes it really can find overlooked patterns (this happens quite often in astronomy, where surveys amass heaps of data, which give work to astronomers for years to decades).

On the other hand, original research can be useless, too: when researchers go and measure stuff that was measured a thousand times, when they start with too little a sample, or a bad setup, or with a setup that makes it incomparable to other results, and so on.

_{*Lunatic*, n.
One whose delusions are out of fashion.}