well, I have no RME stuff - but I find their description of the disadvantages of a traditional wordclock signal highly convincing.
Bottom line is that word clock frequency is relatively low which consequently makes jitter errors relatively large in time scale. To avoid this, they 'reclock' devices (via optical cables) by their own technology which includes a well-stabilized higher clock signal.
This may not be applicable (fully?) in large scale studio installations, but in the typical Scope environment it would make sense - given the fact that you actuall can connect all devices in the suggested way. I currently lack the proper link, tho - sorry. It was one of their white-papers.
If I were to install it in my system, I'd replace the Scope board on the sequencer machine with am RME card featuringv that technology, put the Scpoe board in it's own box and have it reclocked by the RME via Adat. Then I'd slave my (old) A16 to the Scope board.
This way errors should be reduced to what the lightpipe may introduce, but the basic clock signal should be highly accurate.
I'd prefer this solution because it would offer me an addional 'active' device for less than the external clock.
For the reasons explained in the RME papers I wouldn't expect a low cost dedicated clock to be as precise in comparison, let alone the BNC cabling, terminating and possible error introduction...
Did I mention Ethernet RG58 networks... anyone remembers them ? as a constant source of enjoyment...
cheers, Tom