This is a well-written article which I enjoyed reading.
However, I do think we should ask ourselves just how remarkable some of these things really are. There is nothing particularly unusual about any of the feats described in the section on the Iyonix, for example. Compared to a SARPC or something I suppose these might seem amazing, but compared to an equal amount of money spent on a modern PC or Mac there's nothing special about them at all. Equally, automatically detecting MDI / MDIX and transferring a file from one computer to another is hardly a ground-breaking development.
I do feel sometimes that we delude ourselves that these tasks can only be done this well on an Iyonix / RPC / whatever, when the reality is that any modern computer can do them very well. Often, we pay a premium for a system that is in reality no more capable that many others that cost much less.
We don't buy RISC OS systems because they're better, we buy them because we like them better, and this a crucial point. I think sometimes we feel we have something to prove and have to justify our choice by saying that it's better, when actually the fact that we happen to like it is justification enough.