Hi wangi.
(Non OSM mapping / GIS people may want to skip this message completely...)
You make an excellent point regarding OSM tagging - basically that you don't tag for the rendering. This principle would suggest that we should record that a route exists whatever quality it exhibits, adding data about signs and other qualities in addition to the basic information.
But I propose that in reality things are a bit more complicated.
The basic definition of a cycle route in the OSM wiki for example says "Cycle routes... are named or numbered or otherwise signed routes." I'd suggest that in this phrase we're talking about something which exists beyond a set of plans, maps or data. The closest we get to a definition of a route which doesn't really exist on the ground is the idea of a proposed route.
But you could read things more flexibly I suppose... which brings me to my second and primary point. The key issue is that there is currently no way to tag cycle routes to describe the quality of the signage and more general infrastructure. I'd propose that there won't be one in the near future either. Not only are we rubbing up against very subjective measures, which perhaps don't belong on OSM at all, but we're also trapped a little by the basic 'relation' system used for cycle routes. Unless we start tagging cycle routes as super-relations then tags such as signage quality will need to refer to the whole route. There's no simple way to tag that one bit of route is well signed and another not signed at all.
So given these things I propose that we're back to the original question. What bits of the RR route actually exist (beyond lines on a map), and which bits don't? And given the general state of the route we have to be a bit less forgiving about what actually 'exists' than we would be if there were two signs on either end of an old railway path or equivalent.
I do get that the RR route has some signs and in parts definitely should be in the OSM data, but if a majority of them are missing, or all are missing for some areas, then I think we should consider either removing at least some pieces of route from the data, or at the least downgrading them to 'proposed' status.