That's not quite what I meant - perhaps I should spell out how it works a bit more.
Online journey planners work off a huge heap of data on fares, the timetable, network topology, and routeing. None of this is published - nor is it readable by people (even if it's mostly text files). The restrictions are also set out as text, and these are published on the
NRE‡ site, for example in
nre.co.uk/lc.
BRFares takes the fares data files and presents them in a form we can read. It also takes the restriction code from NRE, and shows that almost unaltered (it now adds a bit of fancy formatting). The "unpublished restrictions" are only unpublished in the sense that they are part of the data feeds, intended for internal railway industry use. But they should match the text - and that's probably a legal requirement! So if they don't agree, it's not a sneaky trick it's a cock-up.
I think (from observation and guesswork) the restriction data is produced with a lot of computer help but far from automatically; the text is produced by humans. I guess that editing the old text is the easiest way to do that, but is error-prone. There's a lot of scope for silliness, some of which has been reported on the forum. This is partly due to the crude way
GWR▸ define restrictions at each station, blocking all trains within narrow time intervals. They can (and do) sometimes disallow travel by accident, where the times are based on one line's timetable and happen to catch trains on connecting line. All told, restrictions are a very messy bit of the system.