I'm not a great fan of the numericals (largely because they take me too long!) but I did enjoy this one. As usual, I had to back-track once or twice because I had made too hasty an assumption -- the sort of thing that is quickly sorted out in verbal crosswords but tends to take longer to unravel in numericals. Perhaps I'm becoming a convert, though, as I love the odd corners of number theory that the setters exploit. It was a shame about the misprint, but to me at least 5 down came up so late in the logical sequence that I couldn't believe I could have got that far only to be proved wrong, and I looked at this site, The Site That Must Not Be Named, and the official Listener site for confirmation. So thanks, gwizardry.
One thing I would love to see, perhaps in a setter's blog, is how the setter ensures that there really is a unique solution. Given the intricacy of the procedures I can see that the probability of an alternative is minuscule, but is that what setters rely on? An exhaustive check would be a challenge for Deep Thought.