This is a really cool hack, and I wish I could pay for things with a dumb watch. It's just the right level of useful and silly to be up my alley. But the article, as others have mentioned, is a little off. The author did not "invent" the guess-and-check method for verifying resonance. That's been a staple of radio since the beginning, which is why original tuner dials were actually variable capacitors

> Therefore, an ideal antenna should consist of a 22.12 metre long wire, but by convention fractions of λ-lambda (λ/2, λ/4, λ/8, λ/16, etc.) are opportunely chosen.

This sentence is confused enough to be incorrect. λ/2 is generally preferred as an antenna length (standard dipole configuration) because it will resonate at the appropriate frequency with desirable standing wave characteristics (current maximum and zero voltage at input, voltage maximum and current minimum at ends). λ/4 can be used as a half-dipole, but it requires a ground plane to resonate properly. There are also arguments to be made for a 5λ/8 antenna, but none that I'm aware of for λ/8 or λ/16.

In practice for small antennas, physical length and electrical length are only tenuously related, so it's a matter of creating a circuit that acts like an antenna of the chosen length.