[Talk-GB] C roads again

[Robert Norris]

>
> Ignoring the source information for now, but I suspect it is very
> similar to rights of way information in that it is probably derived from
> OS maps.
>
> The following overpass query highlights the issue, Norfolk standing out
> as especially bad. This is just tertiary roads, there are issues with
> unclassified too.
> http://overpass-turbo.eu/s/4xS

AFAIK there are some (but very few) roads where the C number is sign posted but not that I'm aware of any explicitly.

Whether any of these have ever been captured in OSM is hard to tell.

Unfortunately a brief cross check with Google Streetview, for the very few tertiary roads with 'source:ref=survey' don't seem to bare much scrutiny. The visible signposts don't have a 'C' in them. The 'source:ref' bit is only on a short section of an otherwise longer road anyway, so possibly a road split editing leftover.

Obviously source=survey tags it too imprecise to tell whether a ref was surveyed.

However I am in favour of this edit, but I think the edit needs to *only* change 'C' Roads, as some B roads are tagged tertiary.

e.g. using something like this:
   <has-kv k="ref" regv="^C"/>

In the above query will prevent altering too many things.

Possibly only change things without source tags or with source=[nN][pP][eE], as a first iteration too.


> I am going to use the horrible word, mechanical edit, but I feel it
> needs sorting out.
>
> I propose that nothing is removed, but the ref tag for tertiary and
> unclassified is moved to official_ref. This will retain the data and
> allow OSM to be used by those who can make use of this data.

> I know we should not tag for the renderer, but I think its ok to give
> the renderer a clue as to which are displayed on signs and which aren't.
>
> Phil (trigpoint)
>
>
>
> _______________________________________________
> Talk-GB mailing list
> Talk-GB at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb