[Talk-GB] Prow_ref format

Robert Whittaker (OSM lists) robert.whittaker+osm at gmail.com
Sun Nov 5 10:42:59 UTC 2017

On 4 November 2017 at 17:49, Dave F <davefoxfac63 at btinternet.com> wrote:
> I've started adding Prow_ref=* to the paths within my Local Authority. I've
> been using the format as decided by them.
> I noticed another mapper has already added a few, but using the format by
> Barry Cornelius at rowmaps.com. I think this shouldn't be used as it's
> Barry's own concoction.
> As the LA is the organisation someone would most likely converse with about
> PROWs, it seemed sensible to use the format issued by them. It makes
> verification of any updates *much* easier.

I'd agree with that. However, one should be careful about drawing
conclusions about what the LA's official referencing system actually
is. The legal record of Rights of Way is held in the Definitive Map
and Statement, whereas we may well be using an electronic
representation of the definitive map for our mapping. In translating
the data to their computer systems, the LA may have altered the
reference format from that used on the Definitive Map. As others have
already noted, there are also inconsistencies in how an LA itself will
refer to their own paths.

> To check I looked at the wiki:
> https://wiki.openstreetmap.org/wiki/Key:prow_ref
> I wasn't really surprised to find another format recommended. A couple
> things appear wrong with this:
> * including the parish name in any format other than as issued by the LA
> will lead to confusion if their boundaries are amended
> * path abbreviations are unnecessary as their classifications are already
> defined in other OSM tags (highway & designation)
> Having a 'standard' within OSM seems counter productive as it would make it
> non-standard with the vast majority of LAs.

Through my work with my tool at
http://robert.mathmos.net/osm/prow/progress/ it's obvious that
different LAs use different reference styles and with different
degrees of consistency. What is clear though is that in OSM we should
adopt a single format *within* each LA to ensure that our reference
values can be uniformly interpreted and processed. But I would agree
that we need to allow different formats to be used for different LAs.
The suggestion that you note in the wiki I think came from me, based
on an attempt to standardise the formats used by some of my local LAs
(Norfolk and Suffolk). Looking at the Definitive Map and Statements,
they weren't entirely consistent, but the suggestion of "Parish FP 12"
was the closest thing to a common standard I could see. IIRC, both BY
and BOAT were used interchangeably; I recommended BY for consistency
with the other two-letter abbreviations (FP, BR, RB) that were more
universal. The "Parish FP 12" suggestion was only intended for use
where there was not another obviously different de facto standard in
use by an LA. If that was the case, then I would expect the LA's own
format to take precedence.

Including the parish name is not redundant in numbering schemes where
numbers are only unique to a given parish. While the current parish
boundaries often align with the historic boundaries that were in
existence when the Rights of Way were recorded, some have changed over
time. LAs typically retain the original numbers and parishes when
boundaries change, but will often add any new paths with the correct
parish number. The result is that the parent parish cannot be
determined by geography (or the history of the geography) alone. A
"Footpath no 12" with the boundary of a particular current parish
might be numbered as part of that parish, or it might "belong" to a
neighbouring parish that included that land in years gone by. The
parish name is important, as it typically tells you in which file the
Definitive Statement for the route will be found.

Since I've added more counties to
http://robert.mathmos.net/osm/prow/progress/ I've added support for
(almost) arbitrary prow_ref formats, but each county needs to be
assigned exactly one. With the counties I've set up so far, it's
usually been obvious whether there's a different de facto standard
from the LA or not. But there were some edge cases, where I've perhaps
erred more towards my standard format that I should have done --
although some of that was based on existing use of formats in OSM. If
there are any disagreements with what I've gone with in my tool, then
please let me know.


PS: I've just added Warwickshire Rights of Way to my tool at
http://robert.mathmos.net/osm/prow/progress/warks/ . The prow_ref
format used on their definitive map is XXnna where XX is a one or two
letter code for the historic borough/district, nn is a 1-3 digit
number, and a is an optional lower-case suffix. This is what my tool
will is currently detecingt for this county.

Robert Whittaker

More information about the Talk-GB mailing list