What I did learn is that where keycodes assigns one code (e.g. PRSC) to a single key (e.g. 107), symbols can assign the same keysym to multiple keycodes (e.g. Print can be triggered by both PRSC and RWIN). xev is happy with this, but Cinnamon appears to expect a keysym to be bound to a single key and will only use the first one that's defined; though its idea of "first" is unclear, whether by the numeric key value, the first one that appears in geometry or just something entirely arbitrary. So if I want to use RWIN to function as print screen, I have to unassign Print from PRSC by assigning something else. And my original idea of assigning NoSymbol doesn't appear to have the desired effect of easily disabling a key: either because that's not what it's intended to do or because Cinnamon has other ideas. In the end I chose a keysym that I don't use, Help from Sun (and presumably DEC) keyboards which appears to have no function in Cinnamon, and I can assign that to as many keys as I want since it has no purpose: so I've used it to both remove Print from PRSC and to disable the typo-prone LWIN altogether. Ugh, all day to do that, though I did go down numerous blind alleys in the process.
I'm still not entirely happy with it: though it now functions exactly as I would like, what I do not like is having to modify the system files: okay, I could've avoided modifying symbols/gb by creating a stand-along matias file, but the sysadmin in me still doesn't like what is essentially a single-user custom config being in a system directory and I still have to modify two of the rules files so that the applications (e.g. Cinnamon's keyboard settings) can find it. I guess it may be possible to merge the system files with those in any given user's home directory (obviously just for that user) and I know it is possible to do either that or to override them using the command-line, but whether it's possible to get Cinnamon and the like to do it without major system-level reconfiguration remains to be seen; and if system-level configuration of Cinnamon is the answer, then the solution is worse than the problem! It's been an experience, but argh.
Anyway, again, that's of really questionable relevance to the Matias!
It's interesting what you say about the differences between the Click and the Quiet switches, though I'm not surprised: I think for me the audio feedback also provides part of that "tactility", and I think I also prefer the feel of switches that "clack" which obviously damped switches aren't going to do. I may actually try a Matias Noisy for my daytime keyboard: I'm taken enough with this one that I think it's worth trying out too. I'd probably stick with Matias' own keyboards, partly as it takes me a while to re-adapt from one layout to another, partly for the comments about respective reliability/QA/whatevs and partly because, however easy it was setting up Windows with my new key-bindings, I've no idea how straightforward it is to switch between keymaps: probably not difficult but that is one thing that's really simple with xkb. And which I still don't want to use anyway.

Oh, and I'm sure I've used Xfce at some point: sounds very familiar. I remember experimenting with lots of window managers probably in the late '90s when there were suddenly a rash of them appearing like KDE, Gnome and the aforementioned. But the reason I went with Cinnamon isn't because I especially like it or its features, it'd just got to that point where I couldn't be arsed, and even where customisation is easy, it's often one of those things where the more I meddle, the worse it gets.

Ah, just discovered one problem with my new print screen key: it occupies the position where my fingers still expect to find compose (RALT) so now I'm taking a screenshot every time I want to type an extended character. Sigh.