[OSM-talk] Rutland boundary

[matthew-osm at newtoncomputing.co.uk]

Hi!

On Thu, Oct 19, 2006 at 06:21:48PM +0100, Etienne wrote:
> On 10/19/06, Andy Robinson <Andy_J_Robinson at blueyonder.co.uk> wrote:
> >The boundary was altered to the east side of the A1 near Stamford when the
> >last boundary change was made, so the alignment there will be adjusted to
> >reflect that but may not be accurate. We have a marker from someone for the
> >southern point where the boundary crosses the A1 so we should not be too far
> >off the plot.

> I have a few more boundary waypoints to the east of the A1 that are not
> uploaded yet.   When you are done I'll review how well they fit.

I have a several locations where the boundary crosses roads; in fact I possibly
have all of them south of the A47 (except the road to Allexton): need to check
that. For better or worse, I have split most ways that cross the border at the
border.


Which makes me remember a "think" I was having on borders a while back. Handling
of borders seems to make sense with the idea of ways and meta-ways, rather than
segments and ways. Then each shared border (i.e. Leicestershire-Rutland) is a
way, and the Rutland border is the meta-way comprising all ways around the
county.

In fact, you then end up with a way being defined as an ordered group of nodes,
and a meta-way being defined as an (ordered?) group of ways. Not sure yet how
you would assign short things to a way like you can currently do by assigning to
segments (such as is used for bridge or oneway).

Currently doing a single county is easy---lots of segments and one way---but as
soon as you try and do the county next to it you have to add some of those
segments to another way. A few segments in multiple ways isn't too bad, but lots
could get quite hard to manage, I would have thought.

(This isn't a dig at anyone who is working on any of this stuff at all! Just
sharing what I was thinking about a while ago, and what others have probably
already perfected anyway ;-) .)

Cheers,

-- 
Matthew