Thanks all
The original question was more of a theoretical one, and an engineering one. Other than testing (and OR mandating it's existance) is there a technical need for the E side if the phone line isn't needed.
My 'guess' at the location of the fault is purely based on probability due to line length. The E side is much longer than the D side, so higer probability. And yes, thats completely ignoring that junctions/connections are more likely to go wrong than a solid run of copper.
The disconnects are pretty intermittent - more frequent in cold and wet weather, but it can be fine for days or weeks, and then will start misbehaving for a few days.
As I think I said originally, there's no phone connected, and no internal wiring. Which is a pain to test, but it means it's hard to blame anything other than the line or the router.
If I was a proper friend, I'd get a wired phone and go listen to the line. But as she's more of a
fiend....