Though many countries may be candidates, let us choose South Africa for an implementation.Why ?
There is money there. There is a lot of business there.
This is needed to be adaptable, so that some nodes (such as those on trains) can move about, and someone who buys a Radio Telex can Plug in, without any administrative work. Some key Radio Telexes could act as address servers, that notice changes in the map, and change things around accordingly. The UUCP map scheme, and the PathAlias program that goes with it, would be a good starting point.
Certain nodes, such as those on trains, may need a large amount of storage. This would probably, but not necessarily, be a disk. This may also be needed for other strategic sites (stations at railway junctions, etc.)
Attention needs to be paid to vandalism / espionage - someone may try to re-route messages by changing the maps. However, encryption reduces the value of 'stolen' packets. The Radio Telexes themselves Have a variable power transmitter (like a Cellular phone) and a receiver.
Have some local, expandable, storage - memory. (Flash?) Have an internal battery? This may discourage people from switching it off, and (maybe) only be on receive until powerup again.
Fortunately, it takes very few people to make a system. Two offices 20 miles apart could get up and running right away. However, it would best be started through a Dealer network.
A Dealer must eventually be able to make money by just selling Radio Telexes. Initially, he/she could also have a messaging service, as a letter-writer for those who could not buy a Radio Telex themselves. A train-mounted Radio Telex would be a big asset - and would enable them to offer an inter-city service. This could come as a price of becoming a Dealership, or could be a Government sponsored Radio Telex. Its so cheap!!Andy Rabagliati WZI Consulting