I believe there it's a bit more than a proposal down that way now, and it's more of a recognised trial? I think the documents regarding it are on the RGS site if you are interested in it.
It is a strange step as you say. I suppose if it saves on delay minutes, and gets trains on the move, then that will prick a lot of ears up!! Is the SWT area exceptionally plagued with signalling problems, so much so that they NEED to implement TBW a lot of the time? I would say no, and there is a few places where they regularly have the panel light up like a Christmas tree with faults!
At the moment a train will have to stop at every (defective) signal, contact the signaller and get talked past them in turn, look at it from the signallers point of view where he has 2 or 4 lines to deal with, this would be a huge workload with far to much margin for error and, as long as the driver can get through on the radio, this will take about 20 minimum.
The new proposal is for the signaller to tell the driver to pass SG1, SG3, SG5, SG7 at danger, stop at SG9 and report back (same as TBW now).
Train sets off at 30mph and gets througnh the section in 10 minutes.
This also has less margin for error as the signaller can treat the whole thing as one long section from SG1 to SG9 and their workload is substantially reduced. This would also be the same for trains on other lines as well.
Overall I think the principal is good but I dont want to see it used any more than absolutely necessary and definitely not replacing TBW and hand signalmen etc as a long term solution.