It's a service framework for subscriber based circuit switched connections. The modulation format is separate and I doubt they expected to keep such a meager encoding for the long predicted life of the network.
The error was in not seeing packed switched networks as the clear winner not in choosing a modulation scheme that was designed to bring maximum bandwidth over existing copper infrastructure.
In any case, the strength was that you could bond up to 24 channels in a North American PRI package, which is what a lot of businesses considered their "T1" service, or used as a mixed group of DID and LOOP channels for calling.
Interestingly, in the end, they used bonded ADSL to provide an encapsulated 24 port PRI "circuit" to the customer using just 4 wires.
Actually, it was envisioned that you'd use the D channel for packet switched data traffic (not IP, more like X.25). Why would a private person use more than 16 kbit/s for a home computer? Much more important to have two high-quality voice lines! And we are busily working on ATM-based Broadband-ISDN to even make high quality video calls and TV available! Just you wait!
The error was in not seeing packed switched networks as the clear winner not in choosing a modulation scheme that was designed to bring maximum bandwidth over existing copper infrastructure.
In any case, the strength was that you could bond up to 24 channels in a North American PRI package, which is what a lot of businesses considered their "T1" service, or used as a mixed group of DID and LOOP channels for calling.
Interestingly, in the end, they used bonded ADSL to provide an encapsulated 24 port PRI "circuit" to the customer using just 4 wires.