[Talk-GB] Footpath Open Data is not always accurate.

[Philip Barnes]

On Mon, 2017-02-06 at 14:53 +0000, Andy Townsend wrote:
>     On 06/02/2017 11:18, Colin Smale wrote:
> 
>     
>     
> >       
> >       On 2017-02-06 09:57, Dave F wrote:
> >       
> > >          On 05/02/2017 11:33,
> > >         Colin Smale wrote:
> > > 
> > >         
> > > >           Any paths that no longer follow the official route
> > > > (as per
> > > >             the DM/DS) should not be tagged as PROW and
> > > > probably as
> > > >             access=permissive unless they go across otherwise
> > > > public
> > > >             land. The official route is still a public right of
> > > > way,
> > > >             it's just no longer usable as such.
> > > >         
> > > 
> > >         
> > > 
> > >         We should be mapping what's on the ground, as PROW signs
> > > &
> > >         stiles indicate, even if that doesn't correspond with the
> > >         definitive map. They should be tagged to correspond with
> > > the
> > >         signs status.
> > 
> >       Not sure I agree with this - the "on the ground" principle
> > can
> >         be taken too far. The real principle is "objective
> >         verifiability" - so two independent "mappers" would come to
> > the
> >         same conclusion. That doesn't always imply that things are
> >         actually visible on site, only that there is an agreed
> > "single
> >         point of truth". In my book that single point of truth
> > would be
> >         the Definitive Map and Definitive Statement, and NOT the
> > signs.
> >     
> 
>     
> 
>     To be honest, I think just applying a bit of common sense is the
>     thing to do here.  I normally "map what's on the ground" but it's
>     pretty common to find PRoW signs pointing in odd directions,
> often
>     where some local scally has decided to have a play with the
> sign. 
>     You can usually figure out where it's supposed to go though,
> usually
>     from signage along the way.  Similarly many people in a
> particular
>     area can point to "the footpath that officially goes through
>     someone's house" or "the footpath that officially goes through a
>     sewage farm".  Usually these are just an error (FSVO error) on
>     whatever map they occur on (for all the reasons already
> discussed).
> 
>     
> 
>     Adding an source explicit source for "designation" if it's not
>     on-the-ground signage does make sense to me though, if only to
> avoid
>     the problems that we had with people "helpfully" filling in names
>     from OS Locator (even when a split-second of thought would have
>     suggested that those names might not be corrent due to obvious
>     spelling errors etc.).  
> 
>     
> 
>     Of course, not all "obviously wrong" paths are wrong, though -
> like
>     http://www.openstreetmap.org/#map=19/53.02259/-1.45416 which is a
>     footpath through a (former) pub.
> 
>     
> 

As Andy says, it is important to use common sense and to remember that
the definitive line was never surveyed so the line shown may never have
been the line (or even possible).
The definitive map as drawn onto OS maps in the 1950s by Parish
Councils, some were better and more conscientious than others. But
mistakes were made, pens slipped and at that time fewer people will
have been familiar with map reading.
Most of the time the system works well and right of way mapping is one
of the areas we can excel and produce a better map. We can map surveyed
lines and also map the 'barriers'. It is important that we map the
positions of stiles, gates and kissing gates.
We should certainly not map a path as permissive just because it
differs from the line on the definitive map, 50m is a rough guide for
it accuracy. If its the surveyed line then its the path.
Phil (trigpoint)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-gb/attachments/20170206/41308124/attachment-0001.html>