understanding that the hole in question has to be less than 1/2 the λ.
Interesting things happen at λ/2: the holes act as slot antennae, and this size provides the optimum coupling from one side to the other. Many mobile phone antennae use a waveguide with λ/2 slots to deliberately couple the signal to the outside world. However, it doesn't form a 'brick wall' filter so that slightly higher frequencies are blocked, rather they are attenuated at a rate I assume from experience in the wired world to be about 6dB/octave, so the transmittance with λ/10 holes would be about 0.1%. Even then that's not the whole story - X-ray machines have a filter to filter out the 'soft' X-rays, and this is a continuous sheet (no holes) of aluminium - the harder (>~70kEv) X-rays pass through it, albeit with some attenuation (exposure times are increased to compensate).
A real-world example I had was for an Oxygen sensor on some equipment I was designing. I sent the circuit diagram of the company standard for this (plus mods to suit the new equipment) along with various notes to the PCB design house (I was too busy with other aspects of the design to do the PCB layout as well), and the design as returned suffered from extreme sensitivity to interference (would alarm if a mobile phone was in use within a few metres). I redid the layout of the sensor PCB to better match my notes (minimum loop areas, balanced input paths, maximum ground plane coverage with minimal holes) without changing the circuitry, and the new layout survived having a mobile phone resting on the sensor with a reading change of only about +-2 in the last digit (better than 100 times less change). With the mobile phone frequency of 1.8GHz, this gives λ/2 at about 8cm and the whole PCB was only 3cm across so should have been immune by your reasoning, and the old design longest ground-plane void was about 15mm, λ/2 for about 8GHz which is well above mobile phone frequencies. My re-layout had a longest ground-plane void of under 5mm which should have given an improvement of about 12-18dB, so some of the overall improvement will have been due to the other measures, but this illustrates the point.
At least you read the link though.
No, actually I opened it, glanced through the content, and said to myself, 'OK, that's what he's using'. The math-heavy link I opened, said to myself 'Looks familiar, I'd have to go back to the textbooks to check it, but I trust this website, and anyway the theory is not too relevant to the discussion'.
Quiz type question: does anyone know which physicist-authored SF book had aliens communicating with humans (initially) by thermal X-rays? It's such a strong concept, it stuck in my mind.