We've discussed this before. A figure-of-eight is arguably theoretically 'safer' than a single ring, and I don't think you'll find any regulation which explicitly prohibits it. However, it can make testing a nightmare, particularly before one has worked out what one is dealing with!You have two things a lollipop and conductors in parallel. If either of the two 2.5mm² feeding the lollipop circuit have even a single socket then it's a figure of 8 and not permitted although I can't recall the rule it breaks only remember it was something we should test for.
IME, it's not usually done quite like that, although the distinction may be a little pedantic If there are several grid switches the two arms of the ring usually go to the two end switches, with all the switches being 'daisy-chained' - hence a true ring, with spurs, not 'parallel cables'. Indeed, there are often other sockets in the ring, as well as the spurs supplied by the grid switches - so, again, in that case, definitely not 'parallel cables'.Using two 2.5mm² in parallel is common for feed to grid switch in a kitchen which in turn as multi spurs ...
I'm with you on that one. For your son to say that something compliant can be 'wrong', simply because it is not common, is, IMO not appropriate.I had a heated argument with my son also an electrical engineer about standard methods. I said it's down to the electrician to read the paperwork and to adapt his testing procedure to comply with the way wired however odd that system may be. He said to wire in a way any other electrician would not expect to find is wrong even if it complies with the regulations as future electricians may as a result make mistakes.
Kind Regards, John